GCC 4.1 Released

Luineancaion writes "Looks like GCC 4.1 has been released. From what I know this includes the GNU Classpath merge and means that Azureus can now be used in a 100% Free-Software system. Thanks to everyone that worked on it, and keep up the good work!"

Comments

But...

[brilinux]

But I just finished compiling 4.0...

Re:But...

[ciroknight]

I just moved to an operating system with 4.0. Wonder how long this'll take to hit the repositories.

Re:But...

[dtfinch]

Very likely never. You'll have to wait for the next major release, or install "unstable" packages.

Re:But...

[Col. 2.7.0-default]

Does it slice? Does it dice? Does it even make Julienne fries?

Re:But...

[strider44]

You're making too much of this. Between 4.0 and 4.1 isn't that big-a change. The only reason why it took so long for many distros to move between 3.3 and 3.4/4.0 was the change in the binary interface between the compilers. In other words a program compiled with 4.0 couldn't link to a library compiled with 3.3. There isn't this restriction between 4.0 and 4.1 so there's no reason why it can't go into the repositories straight after testing.

Re:But...

[larry bagina]

Just to clarify, the ABI changes only apply to C++ libraries/code, not C.

Re:But...

[jZnat]

Ah, the wonders of a versioning system. If you understood it at all, you'd know that GCC 4.x would be incompatible with GCC 3.x. Going by this logic, GCC 4.1 should be an easy upgrade from GCC 4.0, but just as bad if you still had to upgrade from GCC 3.3 or 3.4.

Re:But...

[strider44]

yep sorry I should have mentioned that.

Re:But...

[strider44]

No actually that's wrong, if only slightly. GCC 3.3.* and below use a different C++ ABI to GCC 3.4, 4.0, and 4.1. Even Debian has upgraded from GCC 3.3 to GCC 4.0 so an incremental version upgrade shouldn't be too much of a deal.

Re:But...

[digidave]

"But I just finished compiling 4.0"

You should have compiled it using 4.1. It's *much* faster.

Re:But...

[hdparm]

Just installed Fedora 5 test 1 - comes with gcc4 (0.2) - I bet 4.1 will be in at least preview and included with FC5 final release.

Re:But...

[ultranova]

It would have to be infinitely faster to have helped the original parent poster to compile 4.0

No, just faster than light. If you use a 5-millimeter high font, and put each statement into its own line, that means that it should compile more than 59 958 491 600 lines of code per second.

That's ridiculous speed, but not quite ludicrous speed {;<]-=

Re:But...

[IntergalacticWalrus]

No, this is why the GCC C++ team sucks. It's not C++'s fault the GCC devs don't seem to grasp the concept that binary compatibility breakage is a total PITA that screws up everyone.

I have yet to see similar large-scale breakage issues in Microsoft's C++ library and compiler.

Re:But...

[Harik]

so when the C++ ABI was ratified, it was a BAD thing for GCC to support it? Because they had picked their own before then? Even though GCC is now compatable with all other standards-based compilers? Those terrible developers.

Re:But...

[be-fan]

The GCC devs don't break C++ binary compatibility for fun. They do it because they have to fix bugs in the ABI. The fact that the ABI is so hard to implement can be blamed on C++ and the IA-64 ABI itself. C++ is a bastard of a language to implement, with complex rules and many special cases. On top of that, the IA-64 ABI guys were obsessed with coming up with the fastest-possible C++ ABI, to save every single worthless clock cycle. So the ABI is full of little optimizations that, again, make it more complicated.

The Visual C++ ABI is much older and much simpler. Their exception-handling mechanism, for example, is an order of magnitude simpler, but also incurs more runtime cost*.

* A cost C++ programmers freak out about (3-5%), but one negligable to anybody with a brain.

Home depot

[Stevyn]

Most people who program, myself included as an engineering student, probably take this for granted, but GCC is like having a Home Depot down the street that gives their stuff away. For no cost, anyone can use these tools to create just about anything they want. It's pretty amazing, and fitting for Thanksgiving to show some appreciation, that we all have access to these incredible tools for free.

Re:Home depot

[Seumas]

EXACTLY like Thanksgiving.

Damn opensource man coming in and taking over the closedsource man's customer base.

Re:Home depot

[Lisandro]

GCC is an impressive piece of software - either considered asopen source or in general. A lot of people preffer more "serious" C/C++ compilers, like Intels', but it's a remarkably good compiler by itself, very fast, produces stable and well optimized code AND it is available for every platform you could dream of developing on. Ah, and constantly improving support for other languages (specially Java) is a nice perk aswell :)

    It's safe to say that if we have a healthy OSS community, is because of the great developing tools available on OSS platforms. GCC is a strong contender for that crown, IMHO.

Re:Home depot

[eosp]

Something that basically said just like Home Depot

Long lines and everything...it is the holiday season.

Re:Home depot

[Billly Gates]

I saw some benchmarks a few months ago that closes the gap in performance in c/c++ performance with gcc/g++ 4.0 and the Intel compilers.

Intel wrote them when gcc2.95 was still out and c++ performance was not that good nor was it truly modern ansi compliant by the iso. For example things like the STL were merely emulated and performance for non x86 cpu's was behind too.

With gcc3.x and now gcc4.x its fully caught up in almost all area's. Its nice now to have a nice c/c++ compiler for the alpha and mips processors that produce fast code.

I wonder if the rise of Linux and Free software is what made the compilers catch up?

By the way the Intel compiler is still the way to go for Fortran.

Re:Home depot

[hackstraw]

it's a remarkably good compiler by itself, very fast, produces stable and well optimized code AND it is available for every platform you could dream of developing on.

Yes, gcc is a great compiler, but it is not as good as a commercial compiler like Intel's, or PathScale's, or the Portland groups' or another compiler that is designed for a specific platorm.

I use it by default on my Itanium, Alpha, x86, and SPARC systems, but when performance matters, I go for one that is better optimized.

The most grateful aspect of gcc to the world is that Linux would be impossible without it. And that in itself is enough. I learned C from using gcc, and like I said, I use it daily, but it is not the best performing compiler on the block. It is not uncommon to get up to 100% speedup using an optimized compiler for the platform over gcc, but for portability and familiarity of the compiler across platforms (which is its purpose), it is damn good. Especially when many vendors do not provide a C/C++ compiler for their OS without extra charge.

Thank you GNU and the gcc people.

Re:Home depot

[Lisandro]

The most grateful aspect of gcc to the world is that Linux would be impossible without it. And that in itself is enough.

    Linux, *BSD, and every other single open source OS, as far as i know. Like you said, GCC paved the way for them to exist.

    As for the compiler itself, yes, it's not the best of the bunch, but that doesn't mean it's any bad either (quite the opposite!). Intels' compiler, for example, still beats it for performance (at least the last few times i tried it), but i could live happily with GCC alone.

Re:Home depot

[soulhuntre]

For no cost, anyone can use these tools to create just about anything they want. It's pretty amazing, and fitting for Thanksgiving to show some appreciation, that we all have access to these incredible tools for free.

I agree, it is good to support and be appreciative for cool free tools for developers.

Re:Home depot

[hughk]

Good point. before GCC had spread around and on our systems, the C compiler was quite expensive. Sure we had other compilers but we couldn't compile anything distributed as C source code. There was a lot of stuff around in FORTRAN at the time but the language was stuck in extension hell, i.e. it was pretty difficlt to write cross platform stuff.

When GCC first arrived for the platform VAX/VMS, it was far from perfect, but it gave us access to a lot of public-domain code as well as the early GPL stuff. Of course it got rapidly better. The early VAX C compilers were pretty broken and GCC seemed to outperform them.

The funny thing is by providing a free, platform independent compiler, this was probably the single greatest contribution to open source.

Re:Home depot

[macshit]

As for the compiler itself, yes, it's not the best of the bunch, but that doesn't mean it's any bad either (quite the opposite!). Intels' compiler, for example, still beats it for performance (at least the last few times i tried it), but i could live happily with GCC alone.

I've spent quite a bit of time hacking on gcc, and I'd say my biggest complaint is that a lot of the gcc code really sucks. It's chock full of gigantic impossible-to-understand chunks of code -- few comments, huge numbers of global variables, an "enumerate every case I could think of with 25 page if-statements" coding style, vast numbers of unwritten assumptions about the way your processor works. That it works at all, never mind as well as it does, is a testament to the dedication of gcc hackers.

I think a lot of this is historic, and the newer parts of the compiler are much better (and so the overall code quality is slowly improving as old code gets replaced), but gcc can still be a real pain to work on. If you're trying to port to an architecture that differs in some way from "typical" architectures, be prepared for misery.

Re:Home depot

[GauteL]

If anyone, like me, has been programming Fortran 95 for your research, then they understand what the world would be like without GCC. Every compiler up until the newly released gfortran have been "for pay". If you have four different platforms to cover and a very low budget, then the situation was just crap. Using several different compilers is just bound to give you issues anyway.

In the world of scientific computing, gfortran still needs to prove itself, stability and accuracy-wise, but it is a bloody good start.

Re:Home depot

[TheRaven64]

Actually, that's probably not true. Linux (and most other Free kernels) are written in C, and there are quite a few free C compilers - Fabrice Bellard wrote on that can compile the Linux kernel, and it's only a few thousand lines long.

The thing gcc gives us is a set of compilers for languages like C++, Objective-C and Java that are used by a lot of Free Software that sits atop your kernel of choice. I'm not sure how long gcc will continue to lead however - there are a couple of (BSD-licensed) compilers available now that are actually designed to support multiple front (language) and back (ISA) ends (gcc was designed as a C compiler, and gradualy hacked to fit other languages), which have much better documented code making them a lot more hacker-friendly. And since they're coming from academia they incorporate the latest in compiler research (unlike gcc, which has only recently got SSA, something which originated in IBM research in 1985).

Re:Home depot

[LWATCDR]

And these BSD compilers are?
How about some links?

Re:Home depot

[Anonymous Coward]

LLVM would be one. See http://llvm.org/

But, LLVM uses the GCC front ends, and the GCC folks are actually discussing the integration of LLVM into GCC itself.

Re:Home depot

[Anonymous Coward]

Actually, I'm pretty sure that C++ is the way to go for Fortran.

Re:Home depot

[reanjr]

Actually, PathScale or Absoft are the way to go for Fortran. Yes, even on Intel chips. This is coming from someone who, until very recently, worked at a company that makes Fortran compilers.

Re:Home depot

[andymadigan]

(1) Yes, the Visual C++ command-line compiler is free, but it's damn hard to use without a good GUI, mainly because Windows itself does not have a very good command line. There are also very few tutorials on how to use that compiler.

(2) VS Express is only free for a year, then they're going to charge $50 for it.

I develop in Java, so I just use eclipse. I do use Outlook (the rest of the Office suite isn't even installed, I use OO.org). Those MS IDEs seem to have just a little too much "we'll write the code for you" and not enough "here, go ahead and write." I hate wizards, gimme a god damn text editor.

(3) Allright, you have a point there, but that just shows that Free Software allowed a quality compiler for the OS X platform to be developed (what's the license on that compiler btw?)

The last part of that is flamebait, but I'll respond anyway. I run all sorts of Linux software on Windows e.g. Gaim, Dia, etc. If you make cross-platform software instead of locking yourself into Windows or Apple, a lot more people will be able to run it. I may be a college kid, but I know for a fact that lots of linux systems are servers. Not because it's run by "freaky zealots" but because it's better. Finally, my neighbor runs Ubuntu, I installed it after their Windows system finally died. I've come over to their house about three times since then, I have yet to receive anymore complaints. They use OO.org, Gaim, and Firefox with no problems.

Re:Home depot

[Anm]

100% speedup takes 50% of the time:

speed = progress / time

So since a 100% improvement is 2x the original:

2*speed = 2*progress / time

or

2*speed = progress / ( time/2 )

Re:Home depot

[Ben Hutchings]

GCC wasn't the first or only free and portable C compiler. While it is extremely useful, I don't think it was absolutely essential to the development of free software.

Re:Home depot

[bluGill]

Most estimates place linux/BSD as nearly as popular on the desktop as the Mac. It is hard to measure, because most people just download what they want, but there are plenty of users. It isn't just college kids - many programmers run Unix desktops (mostly KDE or GNOME, but there are others). If your programs target Microsoft Windows, than you run that. If you programs are embedded you can run whatever is handy - Linux is very popular in the embedded world, and those working on such systems use linux on the desktop.

Linux is doing very nicely in the server room. Once again the numbers are hard to count, though companies are more likely to have bought something countable for the backroom. You don't see these systems though as they are locked behind secure doors, and only authorized people are allowed to see them.

I don't care if J. random Luser runs my programs or not. He is not paying me, if he uses them great. If not it doesn't cost me anything. I am not a zealot, I just use what works for me.

Know and love GCC

[JoeShmoe950]

As a developer, I love GCC. Its great, easy, and best of all free. GCC is probably one of the most benifical open source projects around, more important even than linux.

Re:Know and love GCC

[AFairlyNormalPerson]

GCC is bigger than Jesus!

Re:Know and love GCC

[Leroy_Brown242]

I see them more as a symbiosis.

Where would GCC be without Linux?
Where would Linux be without GCC?

Re:Know and love GCC

[kjots]

I think someone might notice half the Internet disappearing overnight.

Re:Know and love GCC

[kocsonya]

> I see them more as a symbiosis.
>
> Where would GCC be without Linux?
> Where would Linux be without GCC?

Well, I was using gcc way before Linux hit the streets. Gcc and the GNU tools were the compiler and utility package that you could run on a bunch of systems without much tweaking of your makefiles. The same input created a binary with the same behaviour. So, I think gcc would be alive and well without Linux. On the other hand, Linux was possible because of the availibility of the GNU tools. So, RMS actually has a point in insisting that it should be called GNU/Linux. As soon as you log in, you probably run bash (GNU), when you type rm, ls, cc, make and so on, it is likely that you invoke a GNU tool or even if not, whatever you start will quicly suck in the C library (GNU).

Zoltan

Re:Know and love GCC

[Per Abrahamsen]

Eh, GCC was the default compiler for several platforms before Linux existed, and was the prefered compiler for cross-compilation to embedded platforms, and for software that ran on a wide range of platforms. It was also the preferred compiler for ANSI C (GCC got support for ANSI C very early).

In other words, GCC would be exactly where it is today, had it not been for Linux.

Re:Know and love GCC

[ObsessiveMathsFreak]

As a developer, I love GCC. Its great, easy, and best of all free. GCC is probably one of the most benifical open source projects around, more important even than linux.

But as a developer, you know that the intel compilers give you that extra cutting edge. Give in to the power of the dark side of the source.

Re:Know and love GCC

[chthon]

GCC existed long before Linux, but Linux made it possible to deploy it large scale on PC's (I know about DJGCC).

When I got into PC's in school in the 80's, Borland ruled.

When I started working in 1990, I wanted to learn C++. There were no inexpensive options however. I knew about GNU and FSF, but all the software they sold was targeted at commercial 32-bit platforms (I think that they did not even had a C++ compiler back then). All other options where too expensive, until the summer of 1991, when Zortech came out with their C++ compiler, which included also 286 and 386 extension software, which made it possible to run 32-bit code.

Linux a year later made it possible for people really interested to run real 32-bit code on a real 32-bit platform.

But you should really add another question to your list :

Where would both be without the Internet ?

Re:Know and love GCC

[DrXym]

In other words, GCC would be exactly where it is today, had it not been for Linux.

I doubt that. GCC was seriously stagnated way before 2.95/3.0 (hence the reason egcs appeared for a while) and was no match at all compared to various commercial compilers. Linux was about the only popular OS which *needed* a modern gcc and thus most of the development came from Linux stakeholders - Red Hat etc. Without Linux I fully expect that the compiler would be an also-ran by now, along with most commercial Unices.

Re:Know and love GCC

[yfkar]

Umm... I don't think that all Linux servers that suddenly disappeared could be replaced before anyone notices.

Re:Know and love GCC

[n1ywb]

Word. Let us not forget that Linux is just a kernel, nothing else. GNU provides all of the essential system utilities that we have come to know and love.

Masters of understatement

[ReformedExCon]

You'd never know it by the link provided that there was anything special about this release.

I am interested in how well it supports ARM5, seeing as how it was dropped as the recommended compiler for certain platforms.

Re:Masters of understatement

[diegocgteleline.es]

Hey, I like this: "GCC can now emit code for protecting applications from stack-smashing attacks. The protection is realized by buffer overflow detection and reordering of stack variables to avoid pointer corruption"

Re:Masters of understatement

[shish]

Would this be the same thing that gentoo's had for a few years? I seem to remember it somehow broke my python :-/ (I forget what it was exactly, and can't look it up; I did the sensible thing and switched to debian a year or so ago :-P)

Re:Masters of understatement

[IamTheRealMike]

No, IIRC this is called Mudflap and generates calls to special glibc APIs which check buffer sizes for calls like read(). It's another layer of security added into Linux lately: on top of exec-shield and SELinux things are feeling pretty secure round here ... too bad not every distro has these things (basically on Red Hat distros have them all).

on the java side

[petermgreen]

are they using the gcj as chacheing jit (e.g. GCJ run on demand to turn class files into shared objects which are then loaded dynamically) system that was mentioned in one of the papers i read recently or what?

Re:on the java side

[ghakko]

The caching JIT has been available since 3.4, but is disabled by default. To turn it on, you'll need to add these switches to your gij command line:

-Dgnu.gcj.jit.compiler=/usr/bin/gcj -Dgnu.gcj.jit.cachedir=/tmp -Dgnu.gcj.jit.options=-O2

In practice, this is not a great help because gij and gcj are so slow. You may be able to get much better results compiling directly from Java source to machine code, and then prelinking the resulting executables and shared objects to reduce startup time.

Re:on the java side

[petermgreen]

i think the idea was to precache most of the stuff your app would use and then let the caching jit take care of anything user supplied or otherwise not possible to precache.

Re:on the java side

[BeeRockxs]

Uh? The compiled-to-machine-code-using-GCJ Eclipse is a lot faster than the normal bytecode version.

Re:on the java side

[ultranova]

-Dgnu.gcj.jit.compiler=/usr/bin/gcj -Dgnu.gcj.jit.cachedir=/tmp -Dgnu.gcj.jit.options=-O2

I get a feeling that the use of this feature is discouraged ;).

Java status?

[harmonica]

From the story: From what I know this includes the GNU Classpath merge and means that Azureus can now be used in a 100% Free-Software system.

Sounds interesting. Is there any ChangeLog to read? I browsed the gcc and the gcj pages, but I couldn't find anything.

Re:Java status?

[Anonymous Coward]

Changelog here

      http://gcc.gnu.org/gcc-4.1/changes.html

Re:Java status?

[harmonica]

Thanks, but it doesn't say anything related to Java.

GCC is the Key to Open Source's Success

[slashfun]

For 13 years I have been a professional UNIX administrator, and if I had to pin down the single most influential software that help propel the Open Source revolution, I would name GCC.

Back in the day the first step in loading up a UNIX workstation with Open Source tools, was to go out and grab a limited precompiled version of GCC, then bootstrap compile an more suitable version, then go to town on compiling all the rest of the goodies that we couldn't live without. We did it so often that it became second nature to go through this process.

I salute you, makers and maintainers of GCC.

Re:GCC is the Key to Open Source's Success

[TorKlingberg]

Looks just like the Gentoo installation process to me.

Re:GCC is the Key to Open Source's Success

[EzInKy]

Looks just like the Gentoo installation process to me.


Not anymore. Gentoo has deprecated building the system from scratch and only supports binary installs now. Add on software is still compiled from source though.

Re:GCC is the Key to Open Source's Success

[jbn-o]

Please do take this in the supportive spirit in which it is intended. It's a letter from GCC's initial author, Richard Stallman (also founder of the free software movement) to a CNet article author who referred to GCC as an "open source" programming tool.

Quoting from that letter:

I appreciate the admiration expressed in your article about upgrading the GNU Compiler Collection, but it erred in describing the program as an "open source" programming tool. I developed GCC as part of the Free Software Movement--so that people can use computers in freedom as part of a community.

Free software means software that respects the users' freedom. The philosophy of the movement is that users of software should be free to run it, study it, change it, redistribute it and publish modified versions.

With these freedoms, you're free to engage in cooperative development; you're also free to develop it on your own or to redistribute it unchanged. Describing this as a "philosophy of cooperative development" emphasizes one beneficial consequence of freedom at the expense of freedom itself.

It was impossible in 1984 to use a computer in freedom, since all the operating systems were proprietary. So I launched the development of GNU, a free Unix-like operating system.

A Unix-like system must include a C compiler, so I wrote one: GCC. I designed it to handle other languages, also, so that GNU users could use more than one. GCC, like the GNU/Linux operating system in which GCC is a crucial part, exists because of the ideals of the Free Software Movement--the ideals that are forgotten when speaking of open source.

Of course, this is not a letter from RMS to you or directly pertaining to your article. However, I thought that it was worth mentioning in case people want to tell their friends about the new GCC release. It seems that people who frequent /. go to some length to make sure that they describe Linus Torvalds' initial authorship of the Linux kernel in a manner according to his chosen movement. I thought that the same respect should be due to RMS.

Re:GCC is the Key to Open Source's Success

[ReformedExCon]

Free Software is a subset of Open Source Software, so to say that GCC is Open Source is absolutely correct, if perhaps more general than RMS would prefer.

You would make the butt of jokes anyone who tried to demand that you call Roma tomatoes "Roma Tomatoes" every single time you referred to them instead of referring to them generally as tomatoes. It simply does not make any difference to anyone what you call them as long as you get the point across that the thing in question falls into a certain group. You can make that group as broad as you want or as narrow as you want, but there is a happy medium where the label is non-offensive and clear to any and all that hear it. Going too narrow may provide more information but at a loss of understanding to your audience.

RMS is in error here, but not because he thinks that software designated Free Software ought not be referred to as Open Source software, but because he thinks that anyone cares.

Re:GCC is the Key to Open Source's Success

[Cardoe]

This isn't necessary correct. You can still install from stage1 and stage2, it just won't be documented directly in the Install Doc because of too many users that were a bit too clueless at Linux attempting to do a stage1 install. Also it reduces the confusion and complexity for first-timers.

Also considering the age of the packages found in all the stage tarballs all of them have been replaced by newer versions so when you install Gentoo you will actually recompile everything from source.

Re:GCC is the Key to Open Source's Success

[gnarlin]

Free Software is a subset of Open Source Software, so to say that GCC is Open Source is absolutely correct, if perhaps more general than RMS would prefer.

First of all, it is true that if one only goes by the definition of open source then the GNU compiler collection does fulfill all the conditions of that definition. GCC also fulfills the Free software definition. Since this software package seems to satisfy both we must therefore look at some other facts to see which group this software appropriately belongs to.

Who started writing it and for what reason?
Richard Stallman did so that we could have more freedom. Not because he thought it simply had more technical merit to publish the code. So GCC would not exist if it were not for the philosophical base upon which it was started continues to rely on. GCC was not started because of any open source philosophy.

What about the name of the software package?
Humm, it appears to have the GNU name right at the beginning. I thus deduce that it is strongly involved with the Free software movement.

In the GCC mission statement it says that GCC is not only a Free software project, but more importanly that one of the main goals is Supporting the goals of the GNU project, as defined by the FSF. Not to mention that the copyright of GCC are kept by the Free Software Foundation.

You can make that group as broad as you want or as narrow as you want, but there is a happy medium where the label is non-offensive and clear to any and all that hear it .

First of all, when talking about open source software you are referring to all software that fulfills the open source definition. That is just as specific and narrow as Free software and the free software definition.
Regarding the non-offensive jib. There is absolutely nothing offensive about free software nor is there anything offensive about open source. Although I feel all warm inside whenever I hear about someone spending their lives work in defending and increasing my freedom in a non-violent manner.

Regarding the clarity of "Free software", that is something which is only a problem in the english speaking world. In most other languages the word for freed(dom) and the word for free(of charge) are different.
Open source is not quite as clear to a non insider of the computer idustry. Freedom is something generally a lot more meaningful then "open source(code)", since most people have no idea what source code is nor how it can be more open or closed but do have some grasp of the meaning of freedom. Many people have certainly heard of open source in the media and such but infer no real meaning from it, except perhaps that it "makes your computer go faster or something" (this is something which I have heard many people say when asked about what open source means to them). If you were to ask those same people if they cared about their freedom, I think you would get at least a little more intelligible answers. Going too narrow may provide more information but at a loss of understanding to your audience.

How can you fail to put across your meaning if you are being as specific, lengtly and clear about your explinations as possible?! This is just nonsense.

RMS is in error here, but not because he thinks that software designated Free Software ought not be referred to as Open Source software, but because he thinks that anyone cares.

Well, the government of my country seemed to care enough for the minestry of education to both endorse his visit and our local LUG (with money and other resources) and to hold a special ministry session were they listened to his recommendations and from that wrote a

Re:GCC is the Key to Open Source's Success

[EvilSmile]

There is a reason why RMS insists it is "Free software" and not "open source software". There is an ethical aspect of the philosophy of free software which is cast aside when you say "open source software". He mentioned this explicitly on many occasions.

Re:GCC is the Key to Open Source's Success

[ironman_one]

I care.

Re:GCC is the Key to Open Source's Success

[oever]

Open Source Software is a subset of Software, so to say that GCC is a computer program is absolutely correct, if perhaps more general than you would prefer.

You are error here, but not because you think that software designated Free Software may carelessly be referred to as Open Source software, but because you are ignoring the most important values that lead to the birth of Free Software. These values are very important to the further developement of our computer usage and should be emphasised to people who are ignorant of them.

Re:GCC is the Key to Open Source's Success

[top_down]

You sound like a Jehovah's Witness citing the bible.

Here is a list of contributors to GCC: http://gcc.gnu.org/onlinedocs/gcc/Contributors.htm l

I would be quite annoyed with all this talk about initial authorship if I had just worked my butt of to get the current release out of the door.

Re:GCC is the Key to Open Source's Success

[Zoxed]

> RMS is in error here, but not because he thinks that software designated Free Software ought not be referred to as Open Source software, but because he thinks that anyone cares.

I care.

Re:GCC is the Key to Open Source's Success

[oever]

Anonymous Coward,
No doubt, your rude trolling will get you modded down soon. But I'd like to point out I have am not involved in the project you're linking to. I only mention it on my homepage as a recursive acronym joke on my first name, which is Jos.

Re:GCC is the Key to Open Source's Success

[Anonymous Coward]

Wrong. There's an install process called stage 1/3 that produces the same result as a stage 1 installation using a stage 3 binary installation. The only other noteworthy source-based distro, SourceMage, does exactly the same: you install binaries and then rebuild the system.
Oh I almost forgot: don't go around spreading FUD you stupid piece of rotting bat shit

Re:GCC is the Key to Open Source's Success

[EzInKy]

I don't know what your problem is but the information I posted was to point out that Gentoo does indeed install a precompiled toolchain by default, something of which the other poster seems not to be aware.

Re:GCC is the Key to Open Source's Success

[LeninZhiv]

RMS is in error here, but not because he thinks that software designated Free Software ought not be referred to as Open Source software, but because he thinks that anyone cares.

Hate to make a "me too" post along with some of the others who have replied to this, but I care, and I care enough to vocalise about it anyway.

What I find even more horrifying in your post is that you make the claim that the term Free software is offensive. I cannot fathom how you have come to hold such an attitude, but I can simply recommend that you reflect on it, and whether you really want to go through life holding a view like that, and whatever other views presumably accompany it.

Re:GCC is the Key to Open Source's Success

[Mjlner]

"Free Software is a subset of Open Source Software, so to say that GCC is Open Source is absolutely correct, if perhaps more general than RMS would prefer."

I don't know if your American or not (I'm not), but would you say it is absolutely correct to describe the US as a country where people are not routinely put in jail without due process? Don't you think that most Americans would prefer to call it a free country? (And this goes for a lot of countries, but Americans tend to create most noise about it! ;-)

"RMS is in error here, but not because he thinks that software designated Free Software ought not be referred to as Open Source software, but because he thinks that anyone cares."

No, I don't think RMS has any illusions about people caring. Quite the opposite. But he *wants* people to care and that can not be achieved by sitting at home doing nothing. I think the biggest problem is still that people don't *know* the difference, and RMS definitely recognises that.

Re:GCC is the Key to Open Source's Success

[ookaze]

Free Software is a subset of Open Source Software, so to say that GCC is Open Source is absolutely correct, if perhaps more general than RMS would prefer

Not only RMS. It was his project, it was created with a purpose. That you lose this purpose in the discussion is not correct like you say. It's not wrong, but you lose lots of very important meaning, and thus later confusing people.

You would make the butt of jokes anyone who tried to demand that you call Roma tomatoes "Roma Tomatoes" every single time you referred to them instead of referring to them generally as tomatoes

Of course, that's because your analogy is wrong.
A correct analogy would be with "sweet potato" and "potato". Replace where approriate, and your comment becomes really stupid.
Of course, you chose a bad analogy to make RMS look stupid, but it was too obvious.

It simply does not make any difference to anyone what you call them as long as you get the point across that the thing in question falls into a certain group

Go tell that to a rogue company. Open Source does not mean anything as to how you can use the code.
It's far more important than you imply.
When people talk about a specialised subject, they are more precise.
Like for cars.

Going too narrow may provide more information but at a loss of understanding to your audience

BS ! Like for cars as I said. You think the general public give a damn about gcc 4.1 ?!!!
A GCC 4.1 release is interesting for people in the know, who sure enough know the difference between Open Source and Free Software, especially as GCC is a core Free Sofware component.

RMS is in error here, but not because he thinks that software designated Free Software ought not be referred to as Open Source software, but because he thinks that anyone cares

You are in error then. You are not in Windows warez land here. A lot of people care about Free Software despite what you claim, and a lot of company care about the implications of Free Software against Open Source.

Re:GCC is the Key to Open Source's Success

[elmartinos]

I think the most influental software is ls. Where would we be without this command?

Re:GCC is the Key to Open Source's Success

[arose]

Well, the government of my country seemed to care enough for the minestry of education to both endorse his visit and our local LUG [..]

Don't you mean GLUG? ;-)

Re:GCC is the Key to Open Source's Success

[Anonymous Coward]

Typing "echo *" probably. Ever screwed up your machine so badly that ls doesn't work? I have. :)

Re:GCC is the Key to Open Source's Success

[drew]

First of all, it is true that if one only goes by the definition of open source [opensource.org] then the GNU compiler collection does fulfill all the conditions of that definition. GCC also fulfills the Free software definition [gnu.org]. Since this software package seems to satisfy both we must therefore look at some other facts to see which group this software appropriately belongs to.

I would say that since this software satisfies both requirements, it could accurately be called by both names. In fact, all Free Software could accurately be called Open Source. Would you say that a Ford Taurus is not a car, because Ford Taurus is a more descriptive name and gives more credit to the makers?

There are times when it is most appropriate to call a Ford Taurus by name, and there are times when it is acceptable to call it a car. The same is tre of Free Software. In this casse, the original poster has said that "GCC is the Key to Open Source's Success", and he is absolutely, undeniably correct, because he is not just talking about Free Software, he is talking about all of Open Source Software. It may not give RMS or his movement the credit that he wants (or deserves, in many people's opinion) but that does not mean that it is not correct.

Re:GCC is the Key to Open Source's Success

[ultranova]

Open Source Software is a subset of Software, so to say that GCC is a computer program is absolutely correct, if perhaps more general than you would prefer.

I believe this is too narrow a definition, actually. After all, it is perfectly possible to use GCC to compile a program without a computer - simply print the instructions and follow them by hand, using pen and paper to mark down variable values.

Would it be correct to call GCC a list of instructions, then ? No. I'm quite sure that the program also contains data structure definitions somewhere, and perhaps actual data too. To be absolutely safe, GCC should be referred to as "entity".

Remember, not all of us can afford the luxury of computers; some of us have to follow the instructions by hand, and connect to Slashdot by whistling to the phone. Not me, but someone might, and it is impolite to make assumptions :).

Re:GCC is the Key to Open Source's Success

[gnarlin]

Don't you mean GLUG? ;-)

Local User Group. Thanks though, but I did have that covered :)

Azureus

[DavidLeeRoth]

Azurues can now be used in a 100% free system to download not so free software :)

Changelog?

[Theovon]

No mention of a changelog? If you're going to announce something, it sure would be nice to have a link to a page that explains some interesting stuff about what's new in it. I've tried looking at their wiki, but its 'news' section and its stuff on 4.1 hasn't been updated since like March.

Re:Changelog?

[Yotsuya]

Google to the rescue.

http://gcc.gnu.org/gcc-4.1/changes.html

That's good.

[Sheepdot]

From what I know this includes the GNU Classpath merge and means that Azureus can now be used in a 100% Free-Software system.

That's good for us, considering that the #1 use of Azureus is to pirate 100% commercial software.

Re:That's good.

[jsight]

That's good for us, considering that the #1 use of Azureus is to pirate 100% commercial software.

I doubt that. It's probably movies and tv shows.

Re:That's good.

[dr2chase]

[This has not too much to do with gcc, but...] You could do something about that. I use Azureus, and according to its accounting I have downloaded 14.5GB of stuff, and shared 371.9 of that back. All legal, some of it even the time it took to suck it down. For all the kvetching about the RIAA, you'd thing people would do more to raise the proportion of legal file sharing.

[Ah, I CAN bring this back on topic.] The reason I use Azureus, and the reason that free software purists should be happy about this, is that Java is a safe language. Look at all the buffer-overflow-based holes in software written in C and C++; Java prevents those by design. Good Free Java support means that a person can be a paranoid/prudent peer-to-peer platform provenance purist.

Re:That's good.

[jZnat]

Porn and upgrades to Azureus I'd imagine would be a better response.

Re:That's good.

[slavemowgli]

Is it? How do you know? Maybe it's just me, but I've uploaded about 150 GB in the past 7 months on BT on my measly 192 kbps uplink, and not a *single* bit of that was infringing on copyright. Why do people always think that BT doesn't have any uses besides copyright infringement? It's not true, and while it *seems* that copyright infringement accounts for a significant part of BT use, there simply is no data that would show that it's actually the #1 use (of course, there's no data that shows it's not the case, either, but without any data at all whatsoever, you shouldn't pull numbers out of your ass).

Sense and portability

[noz]

I was always angry with Sun touting Java(R)(TM)*** as portable when run-time environments were made available for only a small (albeit popular) set of architecture/operating system pairs. My Alpha running Debian at home and my Alpha running FreeBSD at work were left cold, lonely, and wanting Java; running a subset of Java applications with free software partial implementations. This is a triumph for FOSS.

Re:Sense and portability

[Ph33r th3 g(O)at]

Java, cross platform. So long as you're running exactly the right version of the JVM, under exactly the right version of the right web browser, under the right operating system.

Re:Sense and portability

[LDoggg_]

under exactly the right version of the right web browser,

Huh?
You talking about applets?
I didn't know that the new gcj/gij stuff even had a browser plugin.

Re:Sense and portability

[jZnat]

GCJ's had a plugin for a while now, but I wouldn't recommend it for untrusted applets.

Re:Sense and portability

[LDoggg_]

Neat.

Don't know why I haven't come across that before.

Re:Sense and portability

[jZnat]

If you have the time (like I do), it's fun to search through the Debian unstable/ repository and any interesting ones you can find at http://apt-get.org/. I've found many useful things in there; gotta love it.

Re:Sense and portability

[RAMMS EIN]

``I was always angry with Sun touting Java(R)(TM)*** as portable when run-time environments were made available for only a small (albeit popular) set of architecture/operating system pairs.''

You're not alone. Especially since there are so many other languages out there for which "write once, run everywhere" is more true.

``My Alpha running Debian at home and my Alpha running FreeBSD at work were left cold, lonely, and wanting Java;''

Didn't the FreeBSD folks strike some deal with Sun, so that users can get the JDK source code and compile it for their system?

``This is a triumph for FOSS.''

It would have been more of a triumph if people had adopted a language that came out of the FOSS communities, instead of buying into the Java hype. Of course, it's great that FOSS Java implementations are getting closer and closer to running applications not developed for them, but it's still playing catch up.

Re:Sense and portability

[kl76]

My Alpha running Debian at home and my Alpha running FreeBSD at work were left cold, lonely, and wanting Java; running a subset of Java applications with free software partial implementations


Speaking of Alphas, has GCC's Alpha code generation improved any since, say, 3.0? ISTR, GCC-compiled Alpha code used to sometimes run about 2x - 3x slower than code from the DEC/Compaq C compiler.

I'm thinking of contributing to GCC...

[TwoBit]

I love GCC, but I lament that its ability to do inlining is rather bad.
I'm wondering how hard it would be join the project and work on rectifying this.

Re:I'm thinking of contributing to GCC...

[larry bagina]

Depends. Did you get an A++ in your graduate-level compiler construction class? Are you willing to spend a couple months (or more) reading through the gcc source code to get an understanding of how all the pieces work. Are you willing to spend a few more months testing your optimizations (mathematically and with source code) to make sure they don't break anything?

The basics of compilers aren't difficult. 2nd year CS students can understand toy compilers. But gcc isn't a toy compiler, it's a real compiler in the real world dealing with a sometimes crazy language, and even crazier users. If you can write better optimizing code, prove it works, and it doesn't infringe on any patents, submit your code.

Re:I'm thinking of contributing to GCC...

[Microlith]

prove it works, and it doesn't infringe on any patents, submit your code.

Hypothetically, I could prove it works. I cannot prove it doesn't infringe on any patents without:

- opening myself up to willful violation
- hope to possibly find said patent
- expect it be something unique enough that I wouldn't run across it ANYWAY in the course of coding (really making the patent worthless...)

Re:I'm thinking of contributing to GCC...

[Jonny_eh]

WOW! It's amazing anything gets done!

Re:I'm thinking of contributing to GCC...

[multipartmixed]

Parent's not kidding. GCC contribution is non-trivial.

I got an A in my 4th-year undergrad compiler construction course, and build toy languages with flex and bison on a regular basis (about one every 18 months for 8 years).

And the GCC code is way hard stuff.

Re:I'm thinking of contributing to GCC...

[Ian Lance Taylor]

gcc's inlining got much better in the 3.4 release, and got better still in the 4.0 release. If you still see a penalty for inlining in the 4.0 release, please file a bug report as described at http://gcc.gnu.org/bugs.html. Thanks.

As far as joining the project goes, it is easy. Sign up for the mailing lists, check out the code, and start working. Contributing a substantial patch requires signing a copyright assignment before it can be accepted.

Re:I'm thinking of contributing to GCC...

[woolio]

In what sense do you mean that its inlining is bad? Inlining seems like a very simple optimization to implement... I believe GCC uses heuristics (that can be modified with command-line parameters) to decide whether to inline a function or not. Perhaps isn't not doing what you want because either 1) What you want isn't likely to improve performance... 2) Your not supplying the correct options to force it to do what you want... 3) what you want isn't really possible: (like to inline a virtual function call from a base pointer).

Not a programmer but

[Anonymous Coward]

what (if any) implications will this have on OO.o?

Re:Not a programmer but

[ReformedExCon]

quote
what (if any) implications will this have on OO.o?
/ quote

You probably won't be able to link it anymore. You'll have to rebuild from source.

Re:Not a programmer but

[LDoggg_]

Fedora Core 4(with updates) has Open Office 2.0 linked to gcc4.0's java stack.
Mostly works, but I've found some problems with Base connecting to databases other than the local hsqldb ones.

I'm hoping that 4.1 has more bux fixes for this.
And it would be nice if fedora could throw in the the mysql and postgres jdbc drivers into the correct classpath for gij/OpenOffice

Re:Not a programmer but

[codergeek42]

OpenOffice.org already uses the Free Java stack in Fedora. (Thank you Red Hat!) See my screenshot here: http://peter.ramshacklestudios.com/images/openoffi ce.org-free-java-stack.png

What I wonder is: will Fedora Core 5 use GCC 4.1? :-)

Re:Not a programmer but

[ds9]

Well, this means that the guyes at OO.o have even more to do. Would be nice if they would actually create platform independant code, so I could compile OO.o on amd64!

A big thanks to the GCC team.

[rice_burners_suck]

Well, it's Thanksgiving, so let's give all the good folks in the GCC team a big warm Thanks for all the years of improvement to this centrally important software package. Without GCC, it is doubtful that the free software movement could be taking place at all. And with the improvements that have been added in the last year or two, GCC is getting to the point that commercial software vendors will have to come up with some really innovative ideas to compete with it, even for production of commercial software! In fact, I think all software for the Mac is compiled with GCC 4 and onwards.

Thanks folks, and happy Thanksgiving.

Re:A big thanks to the GCC team.

[zoloto]

My big warm thanks will be to donate some substantial funds to them. They seriously kicks some major domo for this kind of work. /me looks for a way to donate!

Actually, it's not released yet

[Dan Berlin]

I'm not sure what gave the person who submitted the story, or the editor who posted the story, the idea that 4.1 was released, but it isn't. In fact, it was just branched less than a week ago. We haven't even put out an RC yet! Really, it's not out. When it is, you will see something sent to gcc-announce

Re:Actually, it's not released yet

[Matt Perry]

I'm not sure what gave the person who submitted the story, or the editor who posted the story, the idea that 4.1 was released, but it isn't.

The GCC home page very clearly states "Current release series: GCC 4.1.0" while below it all the others say "Previous release series" and one at the bottom says "Active development (mainline)." That indicates as clearly as can be which are the old, the current, and the development releases.

If you are right, and 4.1 wasn't released, then the GCC web page might need to be reworded to better communicate that fact.

Re:Actually, it's not released yet

[vidarh]

I guess you didn't see the "(released [date])" bit after the version numbers then... (and the corresponding lack of one for 4.1. Nor the lack of a release history on the 4.1 page.

Re:Actually, it's not released yet

[floamy]

It's just a lack of understanding on your part of how the GCC development process works. There is no need to dumb the site down. The site is by people who know what is going on, for people who know what is going on.

That's not how JIT works

[horacerumpole]

Just regarding your description of JIT - that's not how JIT works, at least not the implementation I know off.

What JIT does is to actually create binary code in-memory and then execute it. Nothing to do with shared objects.

Re:That's not how JIT works

[petermgreen]

its not how conventional jit works indeed

it was mentioned in some GCJ paper i read and i wondered if they were actually doing it yet or not. The advantage is it uses traditional compiler tech which is well understood and it can be preseeded so the compiler doesn't have to be used at runtime unless an unknown class is encountered without breaking the dynamics of java (java is very dynamic, you can load classes at will from anywhere and start calling thier methods immediately through reflection).

Re:That's not how JIT works

[Anonymous Coward]

So shared objects aren't made out of binary code, they're made out of... mayonaise? cream cheese? horseradish?

Re:That's not how JIT works

[LDoggg_]

"it`s so easy to make! You just get cocaine and some baking soda... and I think I tasted egg and cinnamon..."

Still not released

[lancelott]

GCC 4.1.0 is not yet out as far as I know. This story is misleading. Just because the site lists 4.1.0 on the front does not mean it is out. Notice that it doesn't have a release date on it.

Re:Still not released

[griffster]

4.1.0 is not out yet - the lastest 4.1 development snapshot was released on the 12th of November - it can be downloaded from ftp://ftp.mirrorservice.org/sites/sources.redhat.c om/pub/gcc/snapshots/4.1-20051112

Re:Still not released

[jmv]

What? Of course it's released. You can get it on ftp.sco.com right next to the 2.7 Linux kernel.

athlon64?

[RoadkillBunny]

How's the amd64 support in it? Last time I checked it still wasn't stable...

Re:athlon64?

[g-to-the-o-to-the-g]

Hmm...looks like your sources suck. I've been happily running amd64 Gentoo for more then a year with no complaints, and loads of stability. See for yourself.

Re:athlon64?

[elknco1]

i think he's talking about the 4.x.x chain.

http://packages.gentoo.org/search/?sstring=gcc

3.4.4-r1 and 3.4.3-r1 are the only stable gcc's for amd64

Re:athlon64?

[pomac]

I run gcc 4.0.2, but i mainly recompiled most of my system with gcc 4.0.1.
Some apps refuse to compile but it's more of a exception than a rule.

gcc --version
gcc (GCC) 4.0.2 (Gentoo 4.0.2-r1, pie-8.7.8)
Copyright (C) 2005 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Re:I fart in your general direction

[mvdw]

Spelling transitional? That's shite. Try strict.

What was wrong with Azureus?

[Nermal6693]

Azureus can now be used in a 100% Free-Software system.

For those of us who aren't Linux geeks :) what was wrong with Azureus before?

Re:What was wrong with Azureus?

[sr180]

Azureus is programmed in Java. Therefore it had to execute under a Java Runtime Environment (JRE) as provided by Sun or IBM. All of these JRE's are not free. Now Azureus can be compiled by GCC (which is free), Azureus can be executed in a 100% free environment.

Re:What was wrong with Azureus?

[mbaciarello]

Thanks for the info, I have one more question though - that's just me knowing pretty much nothing about Java. Say I have an existing installation of Azureus and upgrade to GCC 4.1. Does that mean I could scrap my JRE? Or would I have to recompile Azureus to do that?

Re:What was wrong with Azureus?

[sr180]

You would need to recompile Azureus using GCC.

Re:What was wrong with Azureus?

[Chuckstar]

Do you lose any functionality by compiling Java to machine code? Maybe not with Azureus, but in general don't some of Java's features preclude compiling directly to machine code (as opposed to JVM code)?

Clearly you could just avoid using those features, but I'm curious if you get "full-featured" Java when using the GCC Java compiler.

Re:What was wrong with Azureus?

[LDoggg_]

You may be able to use Azureus as is with gij(gnu intepreter for java) as opposed to natively compiling it with gcj(gnu compiler for java).
Considering that Azureus pulls down bytecode updates, gij is probably the way to go.

I've been using gij and gcj for a while in fedora and am very pleased with how its development is coming along.
That said, I would go scrapping the Sun jre just yet.

Though I haven't tried this new 4.1 version, I have had some minor issues with gij/gcj and the java apps I'm using and even some I've written :)

Re:What was wrong with Azureus?

[Nermal6693]

I'm still not following :(

You say that JREs are not free, but, as jofi points out below, Sun's JRE seems to be free.

Re:What was wrong with Azureus?

[digidave]

Not free as in beer, Free as in speech.

Re:What was wrong with Azureus?

[sr180]

And if you dont follow that:

Sun's JRE is free IF: you can agree to their license agreement, the terms and conditions etc etc. Also, the Sun JRE does not provide the source, whereas the GCC does.

Re:What was wrong with Azureus?

[Nermal6693]

OK, I understand now, but I have one more thing to ask.

"Sun's JRE is free IF: you can agree to their license agreement" - isn't GCC the same? You still need to agree to GCC's licence (the GPL), do you not?

Re:What was wrong with Azureus?

[sr180]

The GPL has no restrictions at all if you only use it. The GPL only applies if you try to distribute GCC or distribute versions of the GCC that you have modified your self. If you simply compile your source code, and distribute that, there are no restrictions at all.

However, The major point here is that GCC provides the source code. Sun does not.

Re:What was wrong with Azureus?

[chromatic]

You don't need to agree to GCC's license agreement to download it or to use it.

Re:What was wrong with Azureus?

[LarsWestergren]

Also, the Sun JRE does not provide the source, whereas the GCC does.

Sun provides the source for the JDK (the compiler), same as GCC does, though not under the same licence obviously.

Since the specs are open, there are open source implemenations of the JRE available.

Re:What was wrong with Azureus?

[Nermal6693]

"The GPL only applies if you try to distribute GCC or distribute versions of the GCC that you have modified your self."

Wow, I didn't know that. Thanks :)

Re:What was wrong with Azureus?

[throx]

However, The major point here is that GCC provides the source code. Sun does not.
You really need to check your facts before posting this stuff. Sun does provide the source code.

Re:What was wrong with Azureus?

[mlk]

> Do you lose any functionality by compiling Java to machine code?

To a degree security, as you can not verify native code.
However GCJ includes a JVM, so the "unsafe" plugin-type code could be kept in Java.

> Java's features preclude compiling directly to machine code (as opposed to JVM code)?

Dynamic loading of JAR files, no, as it include a JVM, so it can start up the JVM when ever it gets a Java class file. It might even the compile it to native (this is what .net and JET do I believe), and keep the compiled code.

> if you get "full-featured" Java when using the GCC Java compiler.

Excluding the parts of the core lib that are not implemented, yes.

The only other feature would be "hot-spot"; this is an optional extra of a JVM (Java Virtual Machine). Any good JVM will include it, potentially making you application faster than statically compiled (to native) code.

I believe the GCJ VM includes a "hot-spot" like tech, however it (and its static-compiler) are not as good as the Sun-provided versions, as such code compiled with GCJ is often slower than that running on Suns JVM.

Re:What was wrong with Azureus?

[mlk]

often slower than that running on Suns JVM.

I know this is going to upset people, so just a quick note on this.
1) Most Java code ends up compiled to native.
2) This (the VM) is not the "slow(+)" bit of Java, GC, and Swing(*) are.

+) Slow, and memory intensive are different things.
*) I love swing as an API, and when put into the hands of a GOOD developer, it can be speedy.

Re:What was wrong with Azureus?

[TheRaven64]

Neither of your links points to a full Free JRE. Kaffe provides a Java VM, and so does Blackdown. Blackdown is bundled with the Sun (non-Free) class libraries, while Kaffe does not include the class library at all. Without the library, Java is pretty much useless. The missing part of the equation is something like GNU Classpath or Apache Harmony.

Re:What was wrong with Azureus?

[Kristoffer Lunden]

Free But Shackled - The Java Trap by Richard Stallman

Agree or not, but that's the gist of the argument. :)

Re:What was wrong with Azureus?

[JonJ]

Where is the source code for their java-compiler then?

Re:What was wrong with Azureus?

[digidave]

"Sun does provide the source code."

It's the restrictions on the Java source code that prevent it from being Free. Also, do they provide the JVM sources? If they did and it was a suitably free license then I'm sure running Java apps on weird architectures and OSes would be made much easier, but I'm pretty sure that lack of Free JVM sources is what's kept Java off these platforms.

The worst part is that some of these platforms are entrenched in enterprises because they run legacy apps that can't be replaced.

Re:What was wrong with Azureus?

[Chuckstar]

Ah, so another question then comes to mind (since I know nothing about GCJ). Can the JVM be static linked into the binary, or does it require GCJ to be installed on any machine that runs the binary?

Thanks for the enlightenment. :)

Re:What was wrong with Azureus?

[mlk]

Either I believe.

> Thanks for the enlightenment. :)
NP

Java like a sieve

[Danuvius]

Java leaks like a sieve. Azureus can't run for more than 24 hours without causing noticable slowdown... java is still an unprofessional mess more than a decade after it was supposed to change the world.

Re:Java like a sieve

[dr2chase]

Your general statement isn't true for me. I run Azureus for weeks at a time on a 400Mhz G4, first running 1.4, now running 1.5. I also run speed-scheduler and keep track of its performance, so I'd notice if it had problems. I'm pretty sure gcj uses the Boehm-Weiser collector; it's proven (tested) to be less leaky than most programs written by C programmers (the tools that find leaks in C programs use the same technology (conservative GC) as BW GC, but instead of collecting the garbage, you're expected to fix your program). By-the-way, 64-bit address spaces are expected to be a win for conservative GC, since they tend to make pointers overlap a smaller set of integers.

And even if did leak -- it's vastly easier to restart a program, than it is to decontaminate a wormy box that's been jacked into by buffer overflows.

But anyhow -- if it's open source, and you have leaks, you can see about fixing them yourself. In many ways the Boehm-Weiser collector is a good deal more approachable than the ones that you find in many Java VMs (I've worked on both); ordinary programmers have a prayer of understanding its behavior (a multithreaded generational collector with finalizers and weak/soft/phantom/JNI-weak references is a truly astonishing piece of software).

Re:Java like a sieve

[sbrown123]

I've ran Azureus for quite a bit, for days as a matter of fact. Haven't experienced the problem you speak of. I was using Sun's JRE 1.5 though so maybe you are using something else. But you shouldnt have since Azureus clearly states to use Sun's JRE 1.5 on their homepage. Change it and your problem will probably go away. If not, post your problem to the Azureus bug tracker and someone can probably help you deduce where your real problem is. Slashdot is not a good forum for airing your problems since it will be unlikely that it will get fixed since you will probably only be seen as someone who has an issue with the Java programming language rather than someone will a real problem. If it is that you just have a general problem with the Java programming language: get over it. Its just a programming language!

Re:Java like a sieve

[cortana]

I always wondered about that thing. Do you just link your program with -lgc and get free garbage collection for no extra effort?

Re:Java like a sieve

[dr2chase]

I don't know the exact incantation, but yes, you can use the Boehm-Weiser GC (perhaps with entrypoints renamed) as a drop-in replacement for malloc and free. People (including me) have been doing this since the late 1980s (long ago, I added the macro definitions for 68030-based Sun workstations; their stack was located somewhere different). The one place where it makes a difference to not be completely oblivious is if you know you are allocating arrays/structures that contain no pointers at all; there's an interface for telling the GC this important information, and it speeds up GC, sometimes a lot.

For code that deals with the net, you're still better off working in a safe language because of the buffer overflows, but otherwise it is pretty much don't-worry-be-happy. I'd worry a little on a 32-bit machine that was completely stuffed with memory, but now you can take your pick of 64-bit architectures, this is not such a problem.

Nothing like research...

[OSDever]

Nothing like research BEFORE posting an article. If you look at the homepage you will see that 4.1 hasn't been released.

And by the way, XHTML 1.0 of any type just won't work in the real high-quality web designing world.. XHTML 1.1 is the way to go, if you can't get 2.0 to work. (Yes, I know it hasn't been completely defined yet.)

Re:Nothing like research...

[Jesselnz]

XHTML 1.1's name is misleading, it's really just modular XHTML 1.0. Unless you plan on taking advantage of its modular features, there's no reason to use it over 1.0 strict.

Re:Nothing like research...

[jZnat]

XHTML 1.1 is the official modularised XHTML 1.0 Strict, and it also requires usage of the application/xhtml+xml MIME type. The good (?) thing about modular XHTML is that you can include only the XHTML elements that you use, even if those are old crappy ones like font, center, etc. There's more to it than just being modular XHTML 1.0...

Re:Nothing like research...

[TheRaven64]

Actually, there is a very good reason to use XHTML over HTML. Since XHTML is XML, you can embed other sorts of XML in the document (properly namespaced, of course), such as SVG, and expose them to scripting via the DOM. If you're not into SVG, how about MathML? The standard way of including mathematical notation in a web page is to render it as a GIF using TeX and embed it as an image - with MathML you can include it in a form where someone can copy and paste it into their equation editor or modeling tool, and play with it.

Re:I fart in your general direction

[ketsugi]

I suppose "Strict" is for "tighters" then?

But what about fixes?

[DoofusOfDeath]

About a month ago, I submitted a bug report for an internal compiler error. The GCC guys jumped on it, but I don't see in that change log a mention of what particular bugs got fixed in GCC 4.1.

Is the changelog just oddly incomplete, or am I looking in the wrong place for the list of bugs that got fixed in this release?

Re:But what about fixes?

[imroy]

The "changes" document appears to just list the added features, changes in ABI, and changes in the language parsers. The GCC homepage has two links to their bugzilla system: serious regressions and all known regressions. Your bug *should* be in there somewhere.

Let me know when it stops sucking

[Mancat]

Gcc has a history of, well, sucking arse for the first few major revisions of any release tree. Is gcc 4 ready for prime-time now, or should I continue on with gcc 3.x? I've already heard of gcc 4 enforcing strict rules that breaks otherwise functional code. I remember how much of a headache that was when gcc 3 rolled around and started spitting out compiler errors that basically said "your code is ugly."

Re:Let me know when it stops sucking

[Pr0xY]

you make it sound like enforcing strict rules is a bad thing. Really the only bad thing that gcc has done is accept that _broken_ code in the past. The fact that it no longer will compile constructs which are invalid in c and c++ is an improvment. c and c++ are just like any other standard (think html/xhtml and such) and when a compiler accepts invalid constructs it destroys the portability of the code.

The true ideal is to be able to write code that if it compiles on gcc you can say "i know for certain that this is valid c++". Such a goal is difficult, if not impossible (many things are "implementation defined") but is stilla goal worth shooting for.

Re:Let me know when it stops sucking

[Mancat]

Oh there's no problem with enforcing strict standards. The only problem I have is when the compiler refuses to compile non-standard code without providing compatibility switches. I'm sure there are such switches present in gcc3/4, but I haven't figured out what they are.

Re:Let me know when it stops sucking

[thaig]

As you know, it hasn't actually been released yet to 4.1 is not ready. The most recent round of dramatic changes were from 3.4 to 4.0. 4.02 seems very good and I haven't seen a Fedora development update for it for quite a long time.

4.1's main new features relate to an improved use of the optimisation architecture that was implemented in 4.0 - so 4.1 is not radically different. It just produces faster object code. It will probably be of extremely high quality but, as ever, you can wait for 4.1.1 if you wish to play it safe.

I am looking forward to it a lot and even more to 4.2 which should reach a peak of sophistication and performance before the next big advance that is even now being discussed on the GCC dev mailing list - the possible integration of LLVM.

Regards,

Tim

Re:Let me know when it stops sucking

[ObsessiveMathsFreak]

I've already heard of gcc 4 enforcing strict rules that breaks otherwise functional code. I remember how much of a headache that was when gcc 3 rolled around and started spitting out compiler errors that basically said "your code is ugly."

Being told your code is ugly is obviously difficult. It's like your childs tutor coming up to you and saying, "You child is a deliquient". In both cases you can get offended, but the simple fact of the matter is, in both cases, the professional is correct. You might see the inner beauty of your creation, but all everyone else sees is a mess in need of repair.

It's also worth noting that in both cases, the situation is largely your fault.

Re:Let me know when it stops sucking

[sashang]

If it said 'your code is ugly' it was probably right. Chances are you wrote some bad incorrect code.

LLVM and GCC will likely merge within a year

[Anonymous Coward]

It looks like GCC will merge with LLVM. Follow the discussion thread to see key GCC developers agree to the many benefits of using LLVM. And even if LLVM is not folded into GCC, Apple plans to use LLVM as its backend anyway.

Re:LLVM and GCC will likely merge within a year

[qbwiz]

They just changed to using Tree-SSA and they gave it a new major version number, and now they're throwing that work out and changing it again? This new system must have a lot of benefits.

C and Objective-C

[SpinJaunt]

* The old Bison-based C and Objective-C parser has been replaced by a new, faster hand-written recursive-descent parser.

I wince at the thought. The sick f*ck(s) deserves a pat on the back and a six-pack at least. Oh and a pay raise.

Re:C and Objective-C

[mark-t]

Hmmm... the textbook I used in the compiler course I took way back in the olden days said that recursive descent compilers were, in general, slower and in general larger than state-based ones.

Mind you, the book is over 20 years old now.

Re:C and Objective-C

[am 2k]

According to a lecture I took last year, this has changed in recent years. Now it's much more important to work on a small set of data at a time, since the processors have large caches. The recursive approach helps to keep the data you're working on inside the cache.

20 years ago nobody had a cache, so it was an imperative to keep your stack size down.

Re:C and Objective-C

[Anonymous Coward]

Part of the problem with state based parsers is that reduction rules can cascade quite a bit for each input token, especially with a complex grammar. It's possible to optimize away a lot of the reductions by just noting that whatever rule you're evaluating can only reduce to another rule, etc. and simply apply the reduction once. I imagine this is the biggest reason that recursive decent is faster in practice. It probably also makes generating the syntax tree easier, since the structure of the parse tree can be used to implement the recursion, making everything faster. No need to keep track of both a state based parser and a syntax tree. I haven't looked at gcc's parser though, so I don't know exactly what it does. It's probably smarter than me.

RMS wrote GCC to pursue software freedom.

[jbn-o]

The philosophical difference RMS describes is quite clear and RMS points it out quite well. The benefits we get from free software are great, but they shouldn't be celebrated at the expense of celebrating the freedom free software gives us for its own sake. You can't "make that group as broad as you want or as narrow as you want" and still convey the same point. People might not know about software freedom, so it's easy to make that mistake without any malicious intent (as I think was the case here). But to set out to refer to programs like GCC—programs written to make software freedom real—in the name of a movement that was built in part to not mention software freedom is ahistorical.

Re:RMS wrote GCC to pursue software freedom.

[Anonymous Coward]

So it's wrong to sneak software Freedom in through a back door?

Open Source Software is software that is available also in source code format. Under this grouping you can have everything from Free Software (GPL), freer software (BSD), not-so-free software (MPL), and wholly un-free software (MS Shared Source). However, the key ingredient here is that the source is available. Hence "Open Source".

Now you can go ahead and say, well Free Software gives you a bunch of other rights and obligations beyond what the other OSS types give you, but that does not in any way diminish the fact that its source is available, making it "Open Source".

In fact, gcc is more than just RMS' "Free Software tool", it is actually a "Free Software multi-computer-language source code compiler for multiple CPU instruction sets." You can go about as specific as you like, but at some point the rest of us are rolling ours eyes and waiting for the tirade to end so we can go back to referring to it as an Open Source tool.

As far as your contention that history is somehow violated because the "Open Source" community ignores the Freedom aspects of the Free Software movement, you are incorrect there as well. You will find that in most cases proponents of OSS are also strong believers in FS. The primary difference is that they see the benefits of Freedom as the primary reasons to choose OSS, whereas Free Software zealots can't see the forest for the trees. In order to realize the benefits of Open Source software, Software Freedom is a necessity. Here's a quote from opensource.org:

The basic idea behind open source is very simple: When programmers can read, redistribute, and modify the source code for a piece of software, the software evolves.

None of this is possible without some degree of Software Freedom, and OSS proponents recognize this inherently. What they also recognize is that bitching about definitions and the assignment of credit is pointless and doesn't motivate anyone except the lunatic fringes.

Re:RMS wrote GCC to pursue software freedom.

[jbn-o]

So it's wrong to sneak software Freedom in through a back door?

Free software would be just as free by any other name, but the Open Source movement doesn't work to get people to recognize and cherish software freedom in its own right. RMS is asking people to recognize that his work was done in pursuit of software freedom, not the developmental goals of the Open Source movement. By the way, GCC was initially developed well before the Open Source Initiative existed.

Open Source Software is software that is available also in source code format.

Not according to the first sentence of the introduction to the definition of the term "Open Source" as defined by the Open Source Initiative, which claims credit for coining the term and starting the Open Source movement. That definition tells us that "Open source doesn't just mean access to the source code.". I take it you mean to refer to the OSI's work since you cite them as an authority on the matter.

Here's a quote from opensource.org:

The basic idea behind open source is very simple: When programmers can read, redistribute, and modify the source code for a piece of software, the software evolves.

That is a development methodology chiefly aimed at programmers, not a call for defending user's freedoms to run, inspect, share, and modify software.

If the OSI is so sympathetic to software freedom, they shouldn't continue to call freedom-talk "ideological tub-thumping" in their FAQ and they shouldn't try to phrase Peruvian Congressman Villanueva's work as pursuit of "Open Source" when Villanueva explicitly told the Microsoft representative that he wanted to call attention to free software (as the OSI does on their front page). Villanueva told Microsoft that his bill doesn't call for "Open Source". Microsoft just uses that term to try and steer the conversation to matters Microsoft's people believe they can address better—price and reliability. Furthermore, I think RMS hit the nail on the head when he talked about the philosophical differences between the two movements and when Eben Moglen talks similarly in his speeches. If the OSI and its defenders are so sympathetic to software freedom, they should be fine with people explicitly engaging in freedom talk, not just (as you say) "inherently" agree with freedom talk but never mentioning it.

It's telling that in an instance where the preeminent Free Software movement advocate, RMS, writes a program to make software freedom a reality, and doing so well before there was an Open Source Initiative, Open Source advocates are arguing that it's okay to associate RMS' work with a movement he "is not against [...] but [doesn't] want to be lumped in with" (quoting his aforementioned essay). These admonitions suggest that even in cases like this one, the objective is to get everyone to remain silent about software freedom so that we can all get on with pleasing business and framing issues in terms of increasing developmental efficiency. I won't go along with that. I will continue to find the ethical and social examinations the Free Software movement raises far more compelling. And I will continue to associate what was initially RMS' work with the movement he started.

Re:RMS wrote GCC to pursue software freedom.

[Anonymous Coward]

Do you get a prize for being a martyr or something for sucking RMS' dick?

His revolution is done. It won. It continues to win, despite his constant screwing with the definitions of "Free".

Take a look at the damage he is now trying to do to 'his' movement with the updates to the GPL. That's not software Freedom, it's just as much about lock-in as any proprietary license. If he continues down this path, touting Software Freedom when he means nothing of the kind, his movement will be eclipsed by other more balanced and sensible movements. In some ways, this has already happened with the non-religious OSI taking the limelight away from the FSF.

The philosophical stuff has its place. And luckily for it, it has already won over the majority of people. In fact, it has been so successful that many people now believe that something without a tangible physical presence has no value, and they are actively circumventing the law in order to "share".

Make no mistake, Free Software, indeed the whole idea of Information Freedom has succeeded. What remains are business issues, legal issues, and the rest of the day-to-day issues involved with any usage paradigm. RMS and the FSF have outlived their purpose, and prove that with every release of a new and more onerous "Free" license such as the restrictive GPL4 or the wholly un-Free GFDL.

But, if you gotta suck RMS' dick, Slashdot is probably the best place to do it.

Re:RMS wrote GCC to pursue software freedom.

[rvalles]

Both those are free software in compliance with the GNU's "4 freedoms". You could have chosen some better example like, say, microsoft's "shared source".

Re:RMS wrote GCC to pursue software freedom.

[greenrd]

That's incorrect, as you can easily find out by reading the FSF site.

Re:RMS wrote GCC to pursue software freedom.

[Paul Crowley]

Microsoft's "shared source" is not free software/Open Source software.

Re:I fart in your general direction

[HeroreV]

Not only that, but it's being sent as "text/html" and has presentation mixed into the markup. Quite nasty.

Thx 2 RH

[Libor Vanek]

IANAGCCD (I am NOT a GCC developer ;-))) but AFAIK we should all thanks to RedHat for putting lot of "people-time" (read - money) into GCC and making constant improvements of it!

Some minor corrections ...

[ghakko]

GCC 4.1 has not been released yet.

A modified version of Classpath has been included with GCJ since 3.2.

Azureus may start in GIJ 4.0, but won't work properly because it relies on parts of the Sun JDK which aren't completely implemented yet in GCJ.

Re:Some minor corrections ...

[bmgz]

I spent 2 days trying to get Azureus to work under my new Ubuntu Breezy installation. Azureus oddly enough showed no indication of Java being the problem ie. it would start slower than usual but would otherwise load properly and look fine, but it wouldn't upload or download a single bit? Installing Sun's Java and making it the default Java sorted things out.

Re:Division and incompatiability.

[LDoggg_]

There is going to be some growing pains with this stuff included in distros, but it should be temporary.
The gnu classpath guys are working towards compatibility with Sun's API specs, not doing their own thing.

Overall I think this is huge for java. I've been playing with the latest gij/gcj stuff on fedora 4 and it looks pretty good. Eclipse looks good for everything except its update mechanism(rpms vs. eclipse's online updater). Most swing and swt apps work fine.
My only problems right now(as of gcc 4.0) are the java sound APIs and some older java.awt repainting stuff.

Thank you! PLEASE MOD PARENT UP.

[RedBear]

GCC 4.1 has not been released yet.

A modified version of Classpath has been included with GCJ since 3.2.

Azureus may start in GIJ 4.0, but won't work properly because it relies on parts of the Sun JDK which aren't completely implemented yet in GCJ.

Thank you for posting some actually useful and accurate information, which is exactly what I was looking for after the extraordinary claims of this poster. Looks like the original poster really went off half-cocked, even beyond most /. standards. Guess I don't need to waste any further time on this article since it's TOTAL BUNK. Congrats, "editors".

If only I had a subscription, so I could request a refund...

Re:Free! Free! I'm Free!

[LDoggg_]

GO MONO!

Now that we got a pretty good and free(speech) java stack, makes you wonder why Miguel didn't just contribute his effort there.
I know he's like C#, but java is much, much closer to that style of coding than the C code the gnome guys were writing.

Nope.

[Anonymous Coward]

The 4.1 *branch* was created recently (in fact last week).

The *release* is still months away.

Toon Moene (GCC Steering Committee).

Don't argue with Daniel Berlin.

[CyricZ]

Dan Berlin is one of the best known GCC developers. His contributions have been many. Just a few weeks ago he helped transition the project from CVS to Subversion. If he says it hasn't been released, then it hasn't been released.

Where?

[Jacek Poplawski]

Where it is released? I don't see it on any ftp server.

Re:Where?

[slavemowgli]

It's not released at all, yet. What the article states is false (boy, I wish I had a dollar for every time *that* happened on Slashdot in the last year or two...)

Re:Free! Free! I'm Free!

[m50d]

Now that we got a pretty good and free(speech) java stack, makes you wonder why Miguel didn't just contribute his effort there.

Because we don't have a pretty good java stack. We've been trying to get one for years, and failing miserably. You still can't take a random java program and expect to be able to run it in a free way. There was even a fresh start with that Apache project recently, which seems to be no more successful with all the rest. Wheras mono has started, got working, and given us a properly usable C# stack already.

Re:Free! Free! I'm Free!

[killjoe]

Did you read the latest update to mono Miguel wrote? There is a new version of mono and he was explaining what is done and what still needs to be done.

I have a tremendous amount of respect for Miguel but in reality he could have saved himself lots of hours and words if he simply wrote "we are still chasing MS taillights but they are pulling away".

I felt bad for the guy, just as he gets close to .NET 1.1 spec MS has already moved on to 2.0. I agree with you, it would have been much better for the OSS community if he worked on java (classpath) or even parrot.

I get the feeling parrot will put the final nail in the coffin of mono and kick both java and .net in the balls.

GCC website

[Per Abrahamsen]

It is surprisingly hard to find out what the current release is from the GCC webpages. The front page has a misleading 4.1. Press "Releases", and you get a misleading 3.4.4. You have to go on to the "Development plan", under Future timeline, to find the actual latest and greatest 4.0.2 somewhere down the ASCII art tree.

Objective-C++

[erikdalen]

The most interesting new feature IMO is Objective-C++ support. Should make it a lot easier for the GNUstep folks to port over MacOSX software.

Re:Objective-C++

[TheRaven64]

Oddly enough, the changes document makes no mention of Objective-C++. This is confusing, since I'm fairly sure several people were using 4.1 betas to compile Objective-C++ code. The biggest benefit I can see coming from Objective-C++ will be a GNUstep WebKit-based browser - it's somewhat depressing that the OpenStep platform was where the original web browser was developed, and yet GNUstep is without one.

What were the BSDs using?

[RAMMS EIN]

Slightly offtopic, but I was wondering this the other day:

Have the *BSDs always been using GCC, or did they use a different compiler in the early days?

Re:What were the BSDs using?

[kl76]

If by *BSDs, you mean 386BSD, NetBSD, FreeBSD etc, then, yes, they've always used GCC (starting with GCC 1.39 in 386BSD, I think). Before that (ie. 4.3BSD and earlier), there was the closed-source pre-ANSI pcc (Portable C Compiler). Not sure whether 4.4BSD used pcc or GCC...

Umm... You clearly don't understand the point

[Kiaser Wilhelm II]

I dont think the grandparent asked for a biographical sketch of Mr. Berlin.

Dan Berlin isn't exactly the only person who works on GCC, so the fact that the webpage, which should be the sole ultimate authoritative source of GCC information is saying something that is contradicting what he is saying is something that needs to be brought to his attention.

No need to get all uppity and act as if he just smited your God.

Re:But.....

[maxwell demon]

No, but it compiles Linux.

Re:Free! Free! I'm Free!

[Decaff]

I get the feeling parrot will put the final nail in the coffin of mono and kick both java and .net in the balls.

I agree that a free multi-language VM is a good thing, but to suggest that it is going to have any impact at all on the huge number of Java and .NET developers is wildly exaggerated to say the least.

Topic closer

[Chealer]

This is certainly the best post in this thread...not that it's hard to figure though. I mean, you read the title and think "Ah, another time Slashdot ridiculizes itself".

More people than you imagine ...

[hummassa]

> RMS is in error here, but not because he thinks that software designated Free Software ought not be referred to as Open Source software, but because he thinks that anyone cares.
More people than you imagine do care.
I know I do, and I know lots of others do, too.

GCC is important, but what about progress in C++?

[master_p]

GCC is a very important piece of open source software, but it is time to see some real progress in C++:

1) type inference and the 'auto' keyword, as it was suggested a while back.

2) real garbage collection. It is not possible to have effective collection without support from the compiler. The Boehm's collector is not a precise collector, because it can not use type information; it is only the compiler that knows such things.

3) a standard library that goes beyond collections, algorithms and files (and is based around garbage collection, mentioned above). We need gui, threads, sockets, XML, etc. Java did it, I see no reason why C++ can not do it. The argument that it is too difficult to do it no longer holds, due to the cross platform toolkits that exist (but none of them is standard or open as GCC).

And I don't want to hear anyone say 'why don't you join GCC to do it yourself', because I have a full-time job working as a C++ programmer and I have barely the time to take a bath, plus it is extremely difficult to enter the GCC development process, due to being highly not-documented (you have to read the sources) and pretty much a closed circle.

Re:Native code compilation?

[mlk]

(unless I miss understand you) GCJ is part of the GNU Compiler Collection (GCJ).

GCC is NOT open source

[CarpetShark]

Please refer to GCC as Free Software. Open Source is something completely different, and not nearly as inspiring, imho. Moreover, when GCC pretty much started the whole Free Software movement, it deserves a bit more caution in terminology ;)

Re:GCC is NOT open source

[wootest]

All political discussion aside, isn't being Open Source a prerequisite of being Free Software (and to avoid stepping on toes: no, that does not mean that I think Open Source came before Free Software)? I think GCC is both Open Source and Free Software, simply because it needs to be Open Source if it's going to be Free Software; releasing your code and allowing others to fork it are key parts of both Open Source and Free Software. I guess the point I'm trying to make is that it's not *just* Open Source (which would imply not being Free Software), but it is by definition *also* Open Source.

Re:GCC is NOT open source

[TheRaven64]

Open Source refers to a development model. Free Software refers to a licensing model which grants certain freedoms to end users. It is possible to have Free Software that is not Open Source - a large number of commercial bespoke software projects are. They are developed by one company for another company. The customer gets all of the rights to the code, but the community is never involved. It is not, however, possible to produce Open Source non-Free software, since the Open Source development model has community participation as a pre-requisite, which requires open licensing terms.

Note, however, that a lot of people say Open Source when they mean Free Software. This is understandable, since the Open Source definition (which reads like it was written by a lawyer) talks more about licensing than about development. Read some of the things ESR has written to learn more about Open Source, and RMS's writings[1] for more about the Free Software ideology.

[1] Although not the whine about how having a secure system destroys user freedoms.

Re:GCC is NOT open source

[wootest]

Open source does not need to "involve the community" because everyone has a different opinion of what that means - from simply accepting patches or advice to letting people submit new code which you'll audit to letting some 'outsiders' help maintain and build the code. If a project whose forks "involve the community", but which itself does not, is that project Open Source?

I hold open source (spelled Open Source or not) to *only* be about releasing the code under a license that enables people to modify or fork the code under reasonable conditions, and even redistribute it - not to define a development model. If we're going to go by public definitions as the one and only, not even the Open Source Initiative's Open Source Definition defines a development model beyond barring discrimination of people, products or technologies *in the license*. (Even by these terms, it wouldn't be discrimination to just develop an open source product yourself and not allow other people in on it - they have full access to the code and can fork it at will, and I think you agree that given this, demanding control or influence of the original copy is a bit ridiculous.)

Maybe it's possible that we're simply misunderstanding each other about how we define the term Open Source. I've laid out my definition above, and I also disagree with your notion of it being impossible to create Open Source software which is non-Free - just because you're not required by the license to redistribute your modifications under it doesn't mean you can't distribute them anyway. (I release as much of my code as possible under the (revised) BSD license.) But that's just it, because I hold it to be a small set of rules and a general spirit rather than a set of licenses or a development model.

Finally, of course this whole subject is very personal, both in the sense of it being something you decide on yourself and in the sense of being something that can offend some people easily. I don't mean to offend anyone with my ideas, and I also don't mean to force my ideas on anyone else. I'm just stating what I think.

Re:GCC is NOT open source

[Bruce Perens]

Open Source refers only to software with a license that complies with the Open Source Definition. That document was created as the Debian Free Software Guidelines and was blessed by RMS as a good definition of Free Software.

The development model is important, but is not part of the definition of Open Source and software that is developed solely by one guy or one company, as Monty Wirzenius and (later) MySQL AB developed the MySQL server entirely inside of their company, is still Open Source even though it does not share the development model.

If Open Source had community participation as a prerequisite, it would never be created. It's almost always created by one person or group and then opened up for community participation.

Although there has been some divergence between Open Source and Free Software, that was driven by personalities that are now irrelevant, not by any real philosophical difference. The two names still refer to the same thing.

Bruce

Re:GCC is NOT open source

[arodland]

Open Source refers to a development model. Free Software refers to a licensing model which grants certain freedoms to end users.

This is, strictly speaking, wrong.

NOT a subset

[CarpetShark]

You can't philosophically be a subset of something you don't philosophically agree with. Free Software is Free Software. Open Source came later, and if anything, is a watered-down version of Free Software.

Re:NOT a subset

[Anonymous Coward]

Fields are just watered-down rings, I say!

Re:NOT a subset

[stonecypher]

You can't philosophically be a subset of something you don't philosophically agree with.

Never met a catholic, a post-Reagan republican, a neo-nazi or an existential realist?

Hell, creation scientists do it twice.

Re:NOT a subset

[CarpetShark]

Never met a catholic, a post-Reagan republican, a neo-nazi or an existential realist?

No. People who mix incompatible philosophies and claim to be all of them are actually eclectic without realising it. You can take different things from different philosophies, and identify with all of them, but when a project is the direct offspring of a particular philosophy, it makes no sense to describe that as part of a different philosophy that came later.

Re:Free! Free! I'm Free!

[killjoe]

Ruby is already having an impact on both Java and .NET developers. Once you can mix and match the vast perl, python and java libraries expect it to explode.

Also the preliminary results suggest that parrot will be faster then the default interpreters. This means the performance difference between Java and Ruby will decrease as well (although I have to give sun some credit 1.6 beta screams).

Re:GCC is important, but what about progress in C+

[BobaFett]

1) type inference and the 'auto' keyword, as it was suggested a while back.

Not part of the standard (yet) and may change before it becomes standard. G++ already has 'typeof' keyword, no reason to replace one extension by another extension until the standard is finished.

2) real garbage collection.

Again not part of the standard, furthermore it's pretty hard to reconsile the standard's requirements that destructors must execute deterministically with any garbage collection. Standard may be extended to include GC in the future, but for now you have to use a 3rd-party GC, there are several for C++.

3) a standard library that goes beyond collections, algorithms and files

G++ includes TR1 libraries, which is a likely next standard for C++ standard library. Beyond that, you can try Boost for a larger library, but don't expect any compiler to include a library which goes far beyond the standard.

Wrong...

[Cardinal Biggles]

You claim that Free Software is a subset of Open Source, a more narrow definition.

If that is so, can you give me an example of a piece of software that is Open Source, but not Free Software?

No? That's right. There is no real difference. Free Software and Open Source software are the same subject, explained differently.

Bogus Blathering

[ChaoticCoyote]

As many have pointed out, GCC 4.1 is actually several months away from release. Slashdot "editors" might want to learn about a concept called "fact checking." I'm disturbed by the amount of GCC bashing in this list. I've never met a perfect compiler, and GCC is far superior to many commercial tools I've used. It provides professional-quality C, C++, Objective-C, Fortran 95 (almost), Java, and Ada compilers for dozens of platforms; the code generation is imprefect, but then again so is most of the code GCC is required to compile! The vile lack of appreciation for GCC simply astounds me -- it is the foundation of Free Software. And it is a fine piece of work that is constantly growing and evolving -- though not as fast as Slashdot's headlines might suggest... ;)

Re:GCC is important, but what about progress in C+

[master_p]

Not part of the standard (yet) and may change before it becomes standard. G++ already has 'typeof' keyword, no reason to replace one extension by another extension until the standard is finished.

'Typeof' and 'auto' are completely different things: 'auto' would be used in place of a type in variable declarations. For example: auto var = 1 + 2;

furthermore it's pretty hard to reconsile the standard's requirements that destructors must execute deterministically with any garbage collection

If it is hard, than make it not hard: why should destructors be executed deterministically? manual memory management will still be there for that reason.

ou have to use a 3rd-party GC, there are several for C++

The only one I know of is the Boehm's collector, which is a very great hack. Trying a search in sourceforge reveals 0 C++ garbage-collection projects. Trying a search in Google even shows my lame attempt for a collector as the 3rd result (!). Further Google results are either bad links, academics pages with lots of theoritical work and even more bad or outdated links, or old implementations mostly for specific Unix flavors.

G++ includes TR1 libraries, which is a likely next standard for C++ standard library. Beyond that, you can try Boost for a larger library

TR1 offers many goodies, and boost even more, but a modern app needs quite a lot more.

but don't expect any compiler to include a library which goes far beyond the standard.

why? just because the C++ commitee is so stubborn? hello? 2005 called and said that a modern language needs modern libraries.

Re:GCC is important, but what about progress in C+

[Anonymous Coward]

RTTI, Garbage collection, a GC-based class library... sounds like
you'd rather be programming in Java. So why don't you?
I use C++ for high performance real time software. I don't
want GC or libraries which depend on it. If I were writing
software for a bank or something, I'd use Java - it's easier,
and already has all the tools and libraries. Why screw around
with C++ if you don't need performance? And if you do need
performance, why are you using garbage collection and other
stuff which will just turn your C++ compiler into a poor man's
java compiler? I don't get it... use the right tool for the job!

-- Tristan

Re:Free! Free! I'm Free!

[LDoggg_]

Mono isn't any better of in this respect.
Is every API available to win32 .net developers available to mono developers on linux?

Of course gcj/gij isn't the perfect 100% Sun compatible platform yet.

If you develop an app with what's available in gnu classpath and compile it to bytecode with gcj, it will run in gij, and more than likely run in Sun's JRE.

Re:Free! Free! I'm Free!

[m50d]

Is every API available to win32 .net developers available to mono developers on linux?

Of course not, but the coverage seems to be better already after what, 6 months, than what has been done with Java in four years or so.

Of course gcj/gij isn't the perfect 100% Sun compatible platform yet.

It's miles short. Mono is like wine - try it and it might work. Gcj I've never seen work on any app I've tried it with. I'm not saying it doesn't work for anything, I'm sure it does, but the probability of a random java app working with gcj seems to be very low, far lower than for mono.

If you develop an app with what's available in gnu classpath and compile it to bytecode with gcj, it will run in gij, and more than likely run in Sun's JRE.

True, but that doesn't seem to be a path anyone is advocating. Do the development tools exist to do that? And will it work with any of the other free java efforts? Is there even a working windows distribution? If you develop for what's supported in mono your program will work on any C# platform that exists at the moment, and you have monodevelop etc. that make doing this a viable option.

Re:Free! Free! I'm Free!

[LDoggg_]

It's miles short. Mono is like wine - try it and it might work.

Still no different than "gij -jar somerandomapp.jar"

Gcj I've never seen work on any app I've tried it with. I'm not saying it doesn't work for anything, I'm sure it does, but the probability of a random java app working with gcj seems to be very low, far lower than for mono.

Or are you just experienced enough with mono to know which win32 libs aren't done yet?
I've gotten plenty of java apps to run with gij, and have a good idea where the problem areas are.

Do the development tools exist to do that?

Fedora has already been bundling gcj/gij, gnu classpath, and eclipse. Works pretty well.

And will it work with any of the other free java efforts?

Java bytecode is java bytecode. The other free java efforts don't have as much of the API implemented as gcj + gnu classpath. Is there another free .net effort?

Is there even a working windows distribution?

here's a win32 build.

If you develop for what's supported in mono your program will work on any C# platform that exists at the moment, and you have monodevelop etc. that make doing this a viable option.

"For what's supported." Just like gcj and gnu classpath.

Re:GCC is important, but what about progress in C+

[zlogic]

GUI? There are QT, GTK, wxWidgets, Motif(sic!), possibly even MFC support.
If you include all these into GCC, you'll get a bloated thing.
If you create something new, people will have more trouble choosing the right toolkit.
If you want gui, threads, sockets, XML, etc., you can try QT - all these in a single relatively small package.

Re:GCC is important, but what about progress in C+

[LizardKing]

GCC is a very important piece of open source software, but it is time to see some real progress in C++

(list of features snipped)

Why not use a toolkit that gives you much more than a plain STL, something like Qt or Boost. Alternatively, why not use a language and environment that already has those features, like Java?

Re:GCC is important, but what about progress in C+

[master_p]

I can not use Qt because for the following reasons: 1) I am developing commercial products but I can not afford the Qt commercial licence, 2) Qt 4 goes in the wrong direction, relying more and more on the MOC compiler, something which I did not like from the previous version.

Boost does not offer database, gui and xml libraries.

Java is too slow for the applications I am doing.

Re:GCC is important, but what about progress in C+

[master_p]

Wrong...it wouldn't be traditional C++. It would be C++ for the 3rd millenium.

There are no tools that support D...IDEs for example.

Re:GCC is important, but what about progress in C+

[master_p]

Why should I sacrifize a) templates (Java generics are not like C++'s templates), b) enums, c) RAII for having garbage collection? I want all of the above, plus the real-time performance of C++. Furthermore, Java IDEs are very slow.

Re:Free! Free! I'm Free!

[m50d]

Still no different than "gij -jar somerandomapp.jar"

md401@md401 ~ $ gij -jar OT42.jar

Exception in thread "main" java.lang.Error: not implemented

at javax.swing.UIDefaults.put(java.lang.Object, java.lang.Object) (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.plaf.basic.BasicDefaults.BasicDefaults () (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.plaf.metal.MetalLookAndFeel.getDefault s() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.UIManager.getDefaults() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.UIManager.getUI(javax.swing.JComponent ) (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JLabel.updateUI() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JLabel.JLabel(java.lang.String, javax.swing.Icon, int) (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JLabel.JLabel(java.lang.String) (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at itunes.client.swing.One2OhMyGod.One2OhMyGod() (Unknown Source)

at itunes.client.swing.One2OhMyGod.main(java.lang.Str ing[]) (Unknown Source)

md401@md401 ~ $ gij -jar jedit42install.jar

Exception in thread "main" java.lang.Error: not implemented

at javax.swing.UIDefaults.put(java.lang.Object, java.lang.Object) (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.plaf.basic.BasicDefaults.BasicDefaults () (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.plaf.metal.MetalLookAndFeel.getDefault s() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.UIManager.getDefaults() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.UIManager.getUI(javax.swing.JComponent ) (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JPanel.updateUI() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JPanel.JPanel(java.awt.LayoutManager, boolean) (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JPanel.JPanel() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JRootPane.createGlassPane() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JRootPane.getGlassPane() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JRootPane.JRootPane() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JFrame.createRootPane() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JFrame.getRootPane() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JFrame.frameInit() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at javax.swing.JFrame.JFrame() (/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/libgcj.s o.4.0.0)

at installer.SwingInstall.SwingInstall() (Unknown Source)

at installer.Install.main(java.lang.String[]) (Unknown Source)

md401@md401 ~/packages/yaggui-0.9.2 $ gij -jar Yaggui.jar

Exception in thread "main" java.lang.NoClassDefFoundError: while resolving class: de.hampelratte.yaggui.Main

at java.lang.ClassLoader.resolveClass0(java.lang.Clas s) (/usr

Notice: gcc doesn't suck. Period.

[Spicerun]

The only problem I have is when the compiler refuses to compile non-standard code without providing compatibility switches.

I disagree. Non-standard code either needs to be rewritten to be standard code, needs to be replaced, or needs to disappear.

Re:Native code compilation?

[moro_666]

it can do both, native and bytecode ... take your pick :)

Every field is a ring.

[weierstrass]

Shouldn't that be the other way around?

Re:GCC is important, but what about progress in C+

[Paralizer]

2) real garbage collection. It is not possible to have effective collection without support from the compiler. The Boehm's collector is not a precise collector, because it can not use type information; it is only the compiler that knows such things.

I'm betting most C++ programmers will disagree with integrated garbage collection. If you can't deallocate memory yourself, you're probably writing bad code anyway. It is not difficult to execte a "delete" or "free()" call where appropriate, an entire abstract layer to do this for you is utterly absurd for such a language.

If you are having problems finding where you need to deallocate memory, or are using memory in an invalid manner, I suggest looking into Valgrind. You can run your program through valgrind and it will give you real-time analysis of your program and any invalid memory read/writes you are performing.

The point is that the release hasn't been made.

[CyricZ]

The point is, regardless of what the webpage says, that the release has not been made. That fact has been confirmed by one of the most regal GCC developers.

Open Source can be closed; Free Software cannot

[Clith]

releasing your code and allowing others to fork it are key parts of both Open Source and Free Software

Well, no, that's the whole point. If it's Open Source, the code does not have to be released, whereas if it is Free Software (e.g. GPL), it does. The code is free from being hidden away.

Open Source can become closed source. Free Software cannot. This is why companies have a hard time with Free Software -- they can't own their changes.

RMS thinks that anyone cares

[Artemis3]

RMS is in error here, but not because he thinks that software designated Free Software ought not be referred to as Open Source software, but because he thinks that anyone cares.

Actually it is you who are in error here. In this sentence, RMS is not "thinking" anything, he is just promoting freedom. Others like you "think" people don't care about freedom and even equate open source with free software because both are "open". But "open" means nothing if you don't have the freedom to use it, study it, share it and modify it with your peers and ensure these conditions remain in all derivated works.

Openness is good, but not as good as freedom, which includes openness, not the other way around as some people think.

BT doesn't have any uses besides copyright infring

[Artemis3]

As long as you have some means of giving back, some people will always think about misuse, and some people will want you to go back being a mere consumer of whatever they choose to feed you with.

Their real worry is people bypassing the system and doing independent content production and distribution (which destroys their controlled market); and just have the perfect excuse...

You don't get it, STILL

[Kiaser Wilhelm II]

That is not the point at all.

The point is that the webpage is making conflicting information. This is what the poster is trying to point out, yet you keep reverting to your God-like defense of someone who dared to point out a glaring contradiction. It was very noble (see, I can use your sort of terminology too!) of the poster to point this out so that King Berlin might take a chance to clarify or fix the problem.

How dare anyone question Mr. Berlin! He is even more royaller and Godlier than the entire GCC project itself!

Jesus christ.

Re:You don't get it, STILL

[FlaSheridn]

The main point is, again, that the release hasn't been made. (If you don't like arguments from authority, try to find a 4.1 release on one of the mirrors.) I agree that the Slashdot headline is misleading (that's why I know that the mirrors don't have it...)

Re:GCC is important, but what about progress in C+

[mclaincausey]

b) enums,

Java 1.5 introduced enumerated types

Re:You give gcc too much credit.

[jandrese]

I know you love your m88k boxes, but it's time to let go.

Re:GCC is important, but what about progress in C+

[Anonymous Coward]

"And I don't want to hear anyone say 'why don't you join GCC to do it yourself', because I have a full-time job working as a C++ programmer and I have barely the time to take a bath, plus it is extremely difficult to enter the GCC development process, due to being highly not-documented (you have to read the sources) and pretty much a closed circle."

And yet you have time to sift through slashdot posts and rant? Well... priorities...

[OT] your sig

[Mr Z]

Fear: When you see B8 00 4C CD 21 and know what it means

Doesn't that terminate a DOS program, returning 0 as the exit code? It's been awhile.

Different obscure machine language: What does this do? 2C 30 C0

--Joe

Read the legalize.

[bluGill]

Read the legalize of both licenses. I cannot use java for my projects at work (an embedded system), because the license will not allow me to distribute it. For GNU software there are requirements that management doesn't like (but not a big deal as we are careful not to 'taint' our code), but we can distribute it with the product.

Buffer overflows a non-issue.

[bluGill]

Proper C++ does not get buffer overflows. The STL does not allow them. Even with C code, buffer overflows are rare in new code, as programmers have learned the hard way to be careful to prevent them.

Historically buffer overflows have accounted for about half of all successful attacks. Secure programming is much more than buffers, your attitude that it is the most common problem makes it more likely that programmers will ignore the very real security holes that are not buffer related.

Java does not allow RAII, which solves a large class of errors that C++ cannot have.

The bottom line is stupid programmers can write insecure software in any language. Smart programmers look for a language that helps, but they know that even perfect tools can be subverted if they are not careful.

Re:Buffer overflows a non-issue.

[dr2chase]

Your claims assume that we are infinitely smart and work infinitely fast. Lacking infinite resources, working in a language where half my vulnerability is taken care of lets me double the attention I direct towards the others. If there are three sorts of vulnerabilitity, and buffer overflow accounts for about half, then it is also the most common problem, though not the majority problem. Any well-informed security-conscious programmer would be aware of these numbers, and my "attitude" will not change their focus in the least (which is to say, your claim is crap, please back it up with numbers). We have known about the danger posed by buffer overflow since 1988 when the Morris Worm hit; all buffer overflows that have occurred in software written since then demonstrate the inherent stupidity of sufficiently many programmers to make networked life over-interesting.

Or, put another way, if half the security holes are a non-issue, then the other half cannot be more of an issue, hence we have no security problem at all (check my math, please).

Perhaps you can see why I prefer to use programs written in safe languages.

Re:GCC is important, but what about progress in C+

[stonecypher]

1) type inference and the 'auto' keyword, as it was suggested a while back.

This has been expected of c++0x for some years.

2) real garbage collection. It is not possible to have effective collection without support from the compiler.

Baloney. Several perfectly good portable snap-in garbage collectors are for sale on the open market. Don't confuse "not possible" with "nobody's done it for free yet." Besides, garbage collection behavior by needs must vary between devices with different architectural considerations; what would be right for a cellular phone would be a nightmare on a vector supercomputer, and vice versa. Pervasive garbage collection is anathema to a language whose focus is extreme portability. C and C++ have instead provided language-level modularization so that you can replace new, malloc and so on with your own mechanisms. If that's not good enough, then you need to hit the books.

3) a standard library that goes beyond collections, algorithms and files (and is based around garbage collection, mentioned above). We need gui, threads, sockets, XML, etc.

We must not have GUI, because user interface varies so wildly between machines that C and C++ are intended to target, including machines without user interface. Threads, sockets and XML are all already in Boost, the set of libraries developed specifically for consideration for inclusion into the C++SL and with a large membership reflecting the ISO/IEC standards committee.

The argument that it is too difficult to do it no longer holds

That was never the argument. The argument was and remains that there is no portable way to do such a thing without imposing a rigid ideal for user interface. C and C++ programmers who do embedded programming are frequently appalled by Java. If you've ever used J2ME in a situation where performance is important, you know exactly why this should never be done, and why things like BREW and Symbian fail so hard.

And I don't want to hear anyone say 'why don't you join GCC to do it yourself', because I have a full-time job working as a C++ programmer

So do the rest of us. Hell, I have two, and I make time for several OSS projects, a gaming community and my significant other. Besides, it's not as if implementing something in GCC makes it part of C/C++.

egcs

[Per Abrahamsen]

The one thing that is correct is that the official gcc was growing stagnant. But that was due to the official maintainer, who was (and is, he still contribute to gcc) a great compiler engineer, but a poor free software project leader. The majority of the work was done by Cygnus Support, whose customers were mostly in the embedded arena. Cygnus then decided to open up development based on their own branch, under the name egcs (and with an understanding from FSF), in order to involve more people in the development. It became a huge sucess, and the egcs branch became the official FSF branch.

The biggest contribution from Linux may be that Linux (together with the favorite /. hate-object, ESR's, Cathedral and Bazaar paper), served as an inspiration to move away from the traditional relatively closed FSF style of maintainership.

Today, SUSE makes good contributions. So does Red Hat, although it is hard to see which part of those contributions come from the old Cygnus part of the company (Red Hat bought Cygnus during the .com boom, and showed their first profit right after, most of their listed "wins" were in the old Cygnus business area). But the maintainer is from CodeSourcery, who does contract compiler work in a rather wide area. And the main contributor may, somewhat ironically, be Apple, it is certainly their email adress I see most on the developer list. Other than that, HP, IBM and Intel contribute a lot.

Re:GCC is important, but what about progress in C+

[BobaFett]

'Typeof' and 'auto' are completely different things: 'auto' would be used in place of a type in variable declarations. For example: auto var = 1 + 2;

G++ allows this: typeof(1 + 2) var = 1 + 2; Not as convenient as auto, but until auto becomes part of the standard, juggling non-standard extensions hardly seems justified.

If it is hard, than make it not hard: why should destructors be executed deterministically? manual memory management will still be there for that reason.

Because a language in which destructors are not executed deterministically is not C++, and G++ is a C++ compiler. Does it mean that such language is "bad"? Of course not. C++ has several strengths, and the standard committee watches above all that any new language features to not destroy those strengths. If you're willing to trade some of these strengths for other features you want, you're looking for a different language. C++ should not try to be the one language to end all languages (we already tried that, remember PL/1?)

The only one I know of is the Boehm's collector, which is a very great hack

There are others, but they are compiler-specific: while not part of the compiler itself, they use the internal knowledge of how a particular compiler, and usually a particular C library, handle memory allocations. Which is exactly what you asked for, a compiler-specific GC.

TR1 offers many goodies, and boost even more, but a modern app needs quite a lot more.

Perhaps modern app, at least the one you're writing, needs a different language then?

why? just because the C++ commitee is so stubborn? hello? 2005 called and said that a modern language needs modern libraries.

And what do you expect G++ developers to do about that? Putting together a library which is even larger than Boost is enormous amount of work, and can only end up one of three ways: nobody uses it, it becomes an standalone project like Boost (because at this point it really has very little to do with the compiler, thus no reason to keep the two projects tied together), or it supplants the standard. The first is a waste of effort, the second is a very worthwhile project but G++ developers are already busy working on one, and the third ... let's just say that if someone knew how to do that, we'd have a larger standard library by now.

Re:Free! Free! I'm Free!

[Decaff]

Ruby is already having an impact on both Java and .NET developers.

There has been a lot of talk about Ruby in recent months, but there is little evidence that in reality is having any impact. But anyway, being influential does not mean success. For example, Smalltalk is an influential language but is rarely used. I'm sure Ruby will be far more used (I use it!), but I really can't see it taking away much jobshare from Java for a long time.

Once you can mix and match the vast perl, python and java libraries expect it to explode.

You already can, and have been able to for years. You can run Python, Ruby, perl and hundreds of other languages on the JVM. So where is the explosion?

Also the preliminary results suggest that parrot will be faster then the default interpreters. This means the performance difference between Java and Ruby will decrease as well

I'm sure it will, but getting a really fast cross-platform VM is incredibly difficult - Java has only been able to match C/C++ speed in the last year or so, and the work put into the JVM by many companies was vast.

Let me make this clear:

[bluGill]

By your definition C++ is more safe than Java. C++ when used correctly uses STD:string for all buffer operations, and this makes C++ just as safe as Java. C++ also allows RAII (Java cannot - at least not in a clean way), which makes it safer.

There are several good C string libraries that do not allow buffer overflows. Thus C is now just as safe as Java.

Of course if you are stupid it is slightly easier to create a buffer overflow in C/C++ than Java, but that is not a factor for you. You like safe programs, and stupid programmers will never write safe programs, no matter what the language.

In my experience, the culture of C/C++ is more attuned to security problems, and that helps prevent the rest of the problems. The Java culture in my experience has not learned to pay attention.

BTW, if you like programs done in a secure language, does that mean you insist that everything on you system be written in e?

Re:Free! Free! I'm Free!

[LDoggg_]

the swing UI defaults thing must have been fixed with gij (GNU libgcj) version 4.0.1 20050727 (Red Hat 4.0.1-5.

I just did a "gij -jar jedit43pre2install.jar" and the installer ran fine. Unfortunately, when running the app it dies at the splash screen.

I'm sure this will get better now that gnu classpath is the default standard API for gcj 4.1.