Compiling Under Wine

now3djp writes "Interesting article over on CodingStyle that demonstrates how I successfully eliminated wasted time maintaining an MS-Windows computer when I could build natively from my GNU computer! /. has followed other cross compilers in the past. This article is different because I used MS's own compiler! This allowed me to get on with real games porting; with only a proportional increase in compile time. Wine has really come a long way in supporting simple apps, let us hope it reaches a 1.0 soon."

Comments

Awesome

[Anonymous Coward]

I can't wait til we have a fully functioning windows emulator. Even if it will kill the need for native apps (read games).

Re:Awesome

[EvilSporkMan]

Well, it won't be WINE, since WINE Is Not an Emulator, as we all know. =P

Re:Awesome

[SCHecklerX]

Be careful what you wish for. It's excellent handling of Windoze apps is part of what killed OS/2. Developers: "Why port it when it can run the windoze version?"

Not exactly the same, but it would be much better to have native apps, as opposed to having to emulate/VM that other OS.

Re:Awesome

[fredrik70]

for games maybe something like codeweavers is a better approach, focus on converting DirectX and let the rest (non-graphics part of the game) be up for porting. yeah, we all could use OpenGL, but it seems that more and more games companies starts to use DX instead, unfortunately.

Re:Awesome

[harlows_monkeys]

Be careful what you wish for. It's excellent handling of Windoze apps is part of what killed OS/2. Developers: "Why port it when it can run the windoze version?"

I don't believe that this is actually true. Win95 had excellent Win3.x compatibility, but developers nevertheless rushed to develop Win95 software. Why would Win3.x compatibility in OS/2 cause developers to forego native development, but not have the same effect on Win95?

I believe that what really killed OS/2 was IBM's attitude toward developers. Microsoft made it trivial to get started in Win95 development. Hell, you could go into Egghead and buy an MSDN subscription and all the tools you needed.

Compare to OS/2, where you had to apply to IBM for permission to develop, and buy an expensive development kit (at least, to officially develop).

I believe it was Jerry Pournelle who wrote of his experiences at a trade show, where he went to the MS booth, and asked what he had to do to develop for the upcoming Win95, and they handed him, on the spot, a development kit. At the IBM booth, he asked what he had to do to develop for OS/2 (a system that was already for sale, unlike Win95, which was still in beta). Did they hand him a development kit? Nope. They handed an application. If they decided he was worthy, he'd be allowed to buy a development kit.

I think that is the reason OS/2 development never took off.

Note! I'm not saying for-pay developer programs necessarily kill a platform. Apple used to have a for-pay program, but it was a joy, because of the astounding support. You sent any question off to DTS, and they would quickly have a good engineer, with full access to the source and the developers, answer it. I was having trouble with interrupt handling in a device driver, and they send me the detailed comments from the ROM source code for the interrupt handler, which explained exactly what was going on.

With Apple DTS, the feeling I had as a developer was that I was dealing with my peers at Apple, who wanted to cooperate with me to make something great. With IBM, I always felt like an insignificant pawn in whatever they were doing.

Re:Awesome

[flex0]

a) windows is, and will always be further along than wine. chasing an API puts you in second place, from the moment you start.

Sure. But developers don't use always the newest APIs with their untested buggy implementations.
Otherwise their products wont run on old machines. There are still many win98 boxes out there.
And MS still offers IE6 for win98.

And the fact that their software runs fine in wine could aswell attract them to not use the newest apis in near future.

b) vmware can run windows fine. it's not expensive, and it'll run most things better than wine ever will, if you want to actually get some work done (i used it for almost a year to do PCB design and circuit design using protel for uni, and this was on a celeron 366!)

Yes but you actually need a windows copy. This might scare vendors from shipping desktop boxes with linux preinstalled.
"Oh sure, you can run Windows apps, just buy it for $300"

c) killing games is stupid. you're just going to make linux less attractive to developers (who won't give too hoots that we need to bend over backwards to use wine, let alone attempt to get a proper port), and what with added cheat detection, wine will probably get marked as a "cheat" more and more often.

-> WineX?

d) i doubt wine will ever be fully useful for newer games. you're better off just supporting openGL in the first place, simply because wine will need to provide an directx wrapper around openGL anyway.

They already have one and it works. (WineX has got a better one tho)

i just don't see what's right about it. wine won't save you anything, except from RSI, since you won't be using it to do anything but watch games crash and run slowly.

Oh well, right now I'm down and uploading some stuff with eMule without any native dlls ... works flawless for me.

And this "slowly" is a myth.
Right now it might be a lil bit slower due to not optimized implementations but after all - Wine Is not aN Emulator.

Re:Awesome

[Losat]

I think you are refering to IBM's Developer Connection, which was a bit like MSDN. This was fee-based but may have been free if you met certain criteria. It seems like they also had a developer partner program, though I can't remember for sure.
However, there were certainly compilers and development kits just anyone could buy and use (no application to fill out, just buy the box).
Exmamples: IBM's own excellent C-Set/2 (C/C++ compiler) (later Visual Age C++); Watcom's excellent C/C++ compiler; Borland's C++ for OS/2; Two (yes two!) distributions of gcc (gcc2 and emx). There were also two "Turboish" Pascal compilers and three "Visual" Rexx packages (somewhat Visual Basic like but using the Rexx language).
Still, I do agree that IBM could have been more friendly to developers, and IBM certainly did enough things wrong with the marketing.

Re:Awesome

[Anonymous Coward]

Speaking from experience, there was a definite Kiss-The-Ring culture surrounding OS/2.

Back then, most IBM stuff could not be purchased retail and had to be procured through the sales channel (who was only interested in you if you were buying mainframes). Meanwhile, Microsoft was spamming CDs like crazy, figuring they would get paid eventually.

Re:Awesome

[Feztaa]

they handed him, on the spot, a development kit

You mean his development kit didn't come with the operating system itself?

I think your argument is flawed...

IBM restricts development kits of OS/2, OS/2 dies. MS readily gives anybody a copy of the development kit for Win95, and Win95 takes off. But the development kits for linux come with linux, nevermind "being given the development kit upon request".

So why hasn't Linux taken off yet?

Re:Awesome

[Qzukk]

Oddly enough, BSD still lives despite the Linux binary compatibility.

Re:Awesome

[Anonymous Coward]

SFU? Why don't YOU STFU!

Re:Awesome

[AvitarX]

Ummm, it has.

Linux has far surpassed where OS/2 was, and is growing in use. Linux's total openness is a part of it's success (another part being it's freeness).

Linux has a large share of web servers. A large share of new super computer instellations. A large share new renderfarm instellations. A large share of scientific workstation instalations. A growing share in educational desktop installations. And a growing share in governmental type settings.

Linux is seriously taking off in a big way. It is HUGE and has far surpassed OS/2.

Re:Awesome

[vistas]

Back in '84, I was the first person at the University of Georgia to get a Mac (I was one of the two campus-wide "microcomputer resources" at the time). For the two weeks that I had it (before a big-wig commandeered it) I BEGGED Apple for a development kit, compiler or *anything*.

"3000 smackers please" was the only comment I got out of Cuppertino.

So I kept playing on the PC with my $49 copy of Turbo Pascal, and my ~$200 copy of Lattice C. When MS dropped their Windows SDK from $900 to $300 in '88, I bit and basically never looked back.

Re:Awesome

[marcovje]

> I don't believe that this is actually true. Win95 had excellent Win3.x compatibility, but developers nevertheless rushed to develop Win95 software. Why would Win3.x compatibility in OS/2 cause developers to forego native development, but not have the same effect on Win95?

Euuh, Microsoft had declared Win3.1x dead?

I don't think Microsoft will declare Windows dead to boost application development for Linux.

Re:Awesome

[rseuhs]

It's excellent handling of Windoze apps is part of what killed OS/2.

Nonsense, OS/2 was killed because PC-makers didn't want to use an OS from a competitor.

"Why port it when it can run the windoze version?"

Great, so you would rather have them ask "Why port it at all when 100% of our customers use Windows?"

Not exactly the same, but it would be much better to have native apps,

Sure, but we have to get a significant amount of users first, then can we expect native apps.

Wine helps building that userbase.

Instead of developers asking "Why port it when it can run the windoze version?" we have to have users asking "Why use Windows when Linux can run Windows and Linux apps?".

Re:Awesome

[rseuhs]

a) windows is, and will always be further along than wine. chasing an API puts you in second place, from the moment you start.

Wrong. An API is by definition a fixed implementation that does not change at a later time. Of course there are additions, but those are not used until most of the userbase supports that extensions (which gives Wine a lot of time to incorporate those extensions).
Almost all Win32 software runs on Windows 98 (5 years old), most runs even on Windows95 (8 years).

b) vmware can run windows fine. it's not expensive, and it'll run most things better than wine ever will, if you want to actually get some work done (i used it for almost a year to do PCB design and circuit design using protel for uni, and this was on a celeron 366!)

Well, it *is* expensive. More expensive than Windows. And you still need Windows, too. And I did not even start to talk about the speed problems.

Also, it can't be incorporated with distributions, therefore distributions can't run Win32-apps "out of the box", but with a good Wine, they can.

c) killing games is stupid. you're just going to make linux less attractive to developers (who won't give too hoots that we need to bend over backwards to use wine, let alone attempt to get a proper port), and what with added cheat detection, wine will probably get marked as a "cheat" more and more often.

We have to build a userbase before expecting any native games. Without Wine, there will not be a gaming userbase, period. So Wine is the only way to make Linux a gaming platform.

Re:Awesome

[SN74S181]

At a National Press Club function that I heard over the radio (NPR) back in that era (I think it was probably in 1985 or 6) Steve Jobs boasted of how the Macintosh was 'hacker proof.' A sealed box 'for the rest of us' whoever that means...

I'm sure there's no record of that speech anywhere that we'll ever find it, of course.

Re:Awesome

[rseuhs]

Even if it will kill the need for native apps (read games).

This is not insightful, that's nonsense.

Currently, there is no real need for native Linux games, simply because there are not many Linux gamers.

Wine is the ONLY realistic way to attract gamers (and create - not kill - the need for native apps).

Re:Awesome

[rseuhs]

So why hasn't Linux taken off yet?

Stop the FUD.

Linux is already the most popular webserver, second most popular general purpose server, most popular cluster, most popular embedded system and soon most popular 3d-modelling desktop.

Re:Awesome

[andrewski]

I would assert that there are more programs for Linux , or more accurately that can be compiled for it and that the api / libs /etc that it requires are supported, than there are for Windows. I would also assert that more innovation and research occurs for Linux than any other platform. I personally think it has in most ways that count surpassed Windows.

You can roll and smoke that.

Re:Awesome

[afidel]

Wrong. An API is by definition a fixed implementation that does not change at a later time.

Try to tell that to the folks in Redmond. MS has done a great diservice to the idea of object re-use, they often fail to follow one of the most basic ideas of OO, that is you can extend but never change an interface. Almost every single service pack changes the interfaces to dozens and dozens of parts of the windows api. This is why we have "dll hell". If MS followed basic principals it wouldn't be a problem because rather than reworking the interface they would simply add new additions to the interface and depricate the old ones (leaving the origional interface and associated code so that programs would still get the behaviour they expect).

Re:Awesome

[Afrosheen]

I find it really ironic that Jobs took this stance eventually, considering the hackerish roots of Jobs and Wozniak.

Woz was a badass hardware wizard (probably still is) and had some good coders around to build decent software. It didn't take much to compete in those days since the microcomputer had basically just berthed itself from Woz's hands (and others globally). I guess it was a nodal point in history when individuals thought it'd be cool to have something like that to play with and eventually work with.

3k for a dev kit. Ridiculous.

Re:Awesome

[fredrik70]

I figure something out with setjmp/longjmp then

}

Re:Awesome

[mwood]

"It's excellent handling of Windoze apps is part of what killed OS/2."

Huh? IBM is *still selling* OS/2 Warp 4.

Re:Awesome

[scrytch]

So why hasn't Linux taken off yet?

Because good sized chunks of the development kit does not come with linux. Go browse the MSDN library CD, or browse it online. Check out the support for an object model that guarantees that any language the major compiler vendors put out, as well as a plethora of scripting languages, will be able to interface with it. I can script my mail app with everything from VB to python to C++. I am buried in more voluminous documentation than I will ever have time to read.

To say for a Linux dev environment "some assembly required" is to grievously understate the problem. The only thing a Linux environment has going for it in terms of documentation is a much bigger base of sample code.

It costs real money to put together a good development kit. But just ask Carmack why he prefers to develop on NT.

Re:Awesome

[fymidos]

if you need to atract gamers, a playstation emulator sounds much better.

But linux does not need to attract gamers. They can stick to windows, nobody cares. I have played so many games i can't count them, i never gave a shit what OS i was using. But i need linux to get my WORK done.

Wine is good for nothing propably but it's a really nice project from an educational point of view.

Re:Awesome

[rseuhs]

if you need to atract gamers, a playstation emulator sounds much better.

Playstation emulator = illegal = not included in distro = awkard to use = not suitable for masses.

But linux does not need to attract gamers. They can stick to windows, nobody cares. I have played so many games i can't count them, i never gave a shit what OS i was using. But i need linux to get my WORK done.

I can't get PDF files on my treo (PalmOS) because the PDF-preparation program from Adobe only runs on Windows.

ONLY WIDESPREAD DESKTOP USE can make those vendors make Linux versions. And games are a prerequesite.

So, you are completely wrong that nobody cares. I do care and everybody else who uses Linux seriously cares.

Wine is good for nothing propably but it's a really nice project from an educational point of view.

Yeah and Samba is also "good for nothing probably".

Heck, Samba is already beating Windows in speed and stability. Why shouldn't Wine be able to do the same?

Re:Awesome

[tunah]

Oh, you don't browse at -1?

Re:Awesome

[Feztaa]

Ummm, it has.

Then why do we keep hearing statistics like "Linux has 3% of the desktop market"? I'd hardly call 3% a great amount of success.

Sure, we've surpassed OS/2. That's because OS/2 is dead and gone. When will we surpass Windows?

Re:Awesome

[bitfoam]

Almost every single service pack changes the interfaces to dozens and dozens of parts of the windows api.

It is intentional part of the strategy: to make the Win32 API a "moving target", precisely to foil efforts to make a Windows clone or workalike.

Digital Research's DR-DOS was gaining a reputation ("a better DOS than DOS") and starting to make some inroads when MS decided to get serious about Windows... and they weren't about to make the same mistake twice.

Re:Awesome

[fymidos]

My point is that there are enough native apps for linux (except games) now so I don't need it. Maybe others do though, thats why i said propably.

Samba on the other hand is good by itself, you can use it for a linux network, in several occasions it has helped me more than NFS. I NEVER used a win desktop app instead of a linux app through wine though.

Games a prerequésite ? well, that's my point: it isn't.

ads

[ldspartan]

I find it entertaining that when I went to read the comments, I got to see an article for Visual Studio .NET :).

--
lds

Re:ads

[ldspartan]

and by "article" I mean "advertisement"...

way to go, brain.

Re:ads

[pizzarobot]

Well all games are programmed for windows...

It is not recommended

[YellowSnow]

that you compile under the influence of any type of alcahol

Re:It is not recommended

[Anonymous Coward]

it's still better than compiling under any type of framed glass.

Re:It is not recommended

[eclectro]

that you compile under the influence of any type of alcahol

but it's ok to post on slashdot???

Re:It is not recommended

[Thalias]

but its okay to write code under the influence, right?

Re:It is not recommended

[Pseudonym]

Don't drink and derive?

Re:It is not recommended

[powerlinekid]

I don't know... I've written some good code while piss drunk. I think the key is that while sober you sometimes get bogged down in doing it The Right Way (tm) where a simple ugly block of code would be better. My code may not be readable the next day but it works :).

Re:It is not recommended

[JohnFluxx]

Agreed! Otherwise you might write a game that looks really good and you think is working... until you sober up the next morning..not that I've ever done that.

Re:It is not recommended

[Negatyfus]

Well, it's not as if anybody would see the difference.

Re:It is not recommended

[XNormal]

I am not under the alcafluence of incahol, some thinkle peop I am. It's just the drunker I sit, the longer I get.

Re:It is not recommended

[scrawny]

Just one night of drunken interfacing can lead you to years and years of support.

Re:It is not recommended

[powerlinekid]

Haha. I could understand this if that coding session accidentally spawned a child process.

which begs the question

[b17bmbr]

does microsoft compile windows under wine?

Re:which begs the question

[bstadil]

Begging the question does not mean what you think it means.

It means :

"to assume something that hasn't been proven as a basis of one's argument,"

"Wine is not good because it is open source", or "Microsoft's compilers are bad because it is closed source", are examples of Begging the question.

Re:which begs the question

[Anonymous Coward]

You know what? It's got a new meaning now, and it doesn't matter what the older correct version is. The meaning has changed, and this is the common usage. That's the evolution of language.

Re:which begs the question

[Anonymous Coward]

Linux now has a new meaning. Linux is now the name of the entire operating system, not just the kernel. This is common usage. It doesn't matter what the older correct version is.

Re:which begs the question

[fredrik70]

VC++ isn't a bad compiler at all, really. They got a quite shitty implementation of the STL lib though(might have changed now - I use VC++ 6). Also isn't properly following the C++ ANSI standard. for example the scope whenyou declare vars in for loops is broken. MS is aware but they can't really fix it easily now, since *lots* of old MFC code would break if they fix it. Yuo can set a flag though to enforce ANSI but it not on by default. Compiler makes quite good code though. If you want a more 'proper' closed source compiler go for Borland's - the command line version is even free on their website! (after a rather hefty registration proc though)

Re:which begs the question

[Pseudonym]

I agree. Back to pre-Chaucer English for everyone!

(Someone with more knowledge of Old English than I is probably going to make a terribly funny joke about Beowulf clusters now. I just want to get in first.)

Re:which begs the question

[critter_hunter]

Don't like their STL? Use STLPort! Bonus points if you can compile it under Windows using their incredibly sucky installation instructions, which sorta ... stop in the middle of the process.

Re:which begs the question

[jpmorgan]

If memory serves me correctly, VC++ 6.0 came out before the C++ Standard was released. 7.0 (aka, VC++ .NET) is far more standards compliant. Of course, it's not perfect... most of the issues are in the STL (which you can replace with a more standards compliant version, stlport).

If you need total standards compliant though, Comeau C++ is the tool for you.

Re:which begs the question

[GauteL]

"Begging the question does not mean what you think it means."

Actually I think you are wrong. I do not for one second doubt that your interpretation was the original meaning, but when almost everyone uses it to mean what the original poster did, then "begging the question" has taken a new meaning. It really DOES mean what the original poster used it to mean.

Besides, this new meaning is much more obvious than the old one.

Re:which begs the question

[Textbook Error]

but when almost everyone uses it to mean what the original poster did, then "begging the question" has taken a new meaning

Except that we're thankfully not at that stage yet - this usage is common amongst North Americans (as are errors such as using "loose" when they mean "lose"), but is not widespread elsewhere.

Re:which begs the question

[Daengbo]

I'm not going to argue with you about whether it really means what you say or not, because those above have done that already, but I will say that he CAN NOT be using it incorrectly, for the following reason: your usage is idiomatic, while his is not. Translation -- your meaning is not the sum of the parts, while beg (ask for) a question, certainly holds to the literal meaning. Whatever the idiom means or has come to mean, it can not displace the literal meaning. A good example would be "give a hoot," (my girlfriend recommends it), as in "I don't give a hoot what you think." It can certainly mean to literally shout, if I care to make it that. Fantastically, unbelievably, and amazingly, while not idioms, can be used in the literal meanings of those words, despite the fact that common usage dictates otherwise.
Regarding the use of loose and lose referred to above, they are not a common American problem, and mispellings such as these almost certainly spring from a lack of daily reading.

Re:which begs the question

[awol]

VC7 is just broken. Wait for the first service pack. We had to rearrange header files _and_ install the unreleased patch from MS just to get some code to compile (code that compiled find under, g++ (HP-UX, Linux, SunOS), aCC (HP-UX), CC (SunOS) and VC6).

Re:which begs the question

[tkg]

...and mispellings such as these almost certainly spring from a lack of daily reading.

Unless, of course, your daily reading is slashdot. ;o)

Re:which begs the question

[scrytch]

"Wine is not good because it is open source", or "Microsoft's compilers are bad because it is closed source", are examples of Begging the question.

Nonsense. Let's compare:

Proposition 1:
Open Source is not good
Wine is Open Source
therefore, Wine is not good

Closed source is bad
MSVC is closed source
therefore, MSVC is bad

These are perfectly valid arguments. They may not necessarily be sound, since the truth of the propositions is questionable, but their validity is not in question.

Here's begging the question:

Steve Ballmer is a big bald mean strongman.
Steve Ballmer stands intimidatingly over people.
therefore, Steve Ballmer is a brute.

The conclusion is just the first proposition in different words. Petito Principi is a formal fallacy, i.e. a fallacy of form. Often you can substitute a different truthful proposition in place of the first and get a sound argument, so arguing against form alone is a pretty weak argument (it puts forth no rebuttals) and a good way to get rope-a-doped in a debate.

BTW, to "beg the question" is a way of saying "demand the question", and to beg a question is to force it to be asked. The fallacy is when you beg your original question, i.e. one of your propositions (and satisfy the demand). The phrase "beg the question" to mean "an obvious question must be asked" quite probably predates the naming of the fallacy -- where do you think the colloquialism for the fallacy came from in the first place?

Re:which begs the question

[hawkstone]

You can set a flag though to enforce ANSI

Out of curiosity, does that flag fix the for loop scope? If so, what the f*** is it?

(The reason I ask is that their KB says to define "for" as "if (0) ; else for" to fix the problem, and that is reeeaaally annoying.)

Re:which begs the question

[fredrik70]

check out the /Ze flag There are some more info regarding VC++ ans ansi there. Also bear in mind that many of MS headers only work with MS extensions (/Ze) on, so if you're using them you're in a bit of trouble.

Re:which begs the question

[fredrik70]

Ah sorry, typo from me, what you're looking for is the /Za flag, same link though...

Re:which begs the question

[Daengbo]

You're right. I should have said daily reading of quality material. Comic books obviously don't qualify, either. :)

Re:which begs the question

[bstadil]

Whatever the idiom means or has come to mean, it can not displace the literal meaning.

Sure it can.

His initial offer was a red herring!

If he were indeed offering a colored fish of the genus Clupea it would have been stated differently.

Interesing enough the literal meaning of Begging the question you are talking about came after the idiom.

Re:which begs the question

[alexpage]

I thought that mixing "loose" and "lose" and related typographical no-nos were just common amongst /. editors?

Re:which begs the question

[hawkstone]

Ah, yes, I now remember.... and I know why I forgot about it. It's very hard to avoid using MS headers, so I've never been able to use the flag without breaking stuff.

Thanks again for the info!

Big deal!

[Anonymous Coward]

Let me tell you something, I've compiled not only under wine, but on vodka, whiskey, shit even ACID. And when I was on acid, my monitor tried to eat me...

compiling?

[Anonymous Coward]

It's not so interesting to me that he managed to compile using VC++ under WINE. VC++ doesn't call any of the APIs you code, it just puts machine code into the file saying you can call them if you want. It's all well and good to have VC++ compile DX9_CreateSurface() (or whatever) into a bunch of PUSHs, POPs and a JMP instruction, but that doesn't help if WINE can't actually call that function when you're testing. It makes more sense to me to use Bochs or VMWare to test your application if you're developing on multiple platforms. Anything less would be short-changing your Windows clients.

GNU/translation

[phoebe]

GNU/Build GNU/natively GNU/on GNU/a GNU/computer GNU/using GNU/a MS-Windows/compiler GNU/without GNU/a MS-Windows/computer!

Re:GNU/translation

[Pharmboy]

GNU/Build GNU/natively GNU/on GNU/a GNU/computer GNU/using GNU/a MS-Windows/compiler GNU/without GNU/a MS-Windows/computer!

Oh, should have read the article I guess, but I figured since no one else does....

Re:GNU/translation

[kfg]

I know my cat likes to sleep on my keyboard. I'm just trying to figure out how she managed to post to Slashdot.

And better than some of mine too.

Funny thing is that I don't recall giving her a user account, so she's either stolen my password or rooted my box. And she looks so innocent sleeping over the sofa.

Guess my momma was right when she warned me felines were devious little fuzzballs.

(No, she's not named after the Friends character. She's named after Phoebe Snow. No, not the singer, the "train babe." Sheesh.)

KFG

Actually....

[miketang16]

I prefer to use Beer. I find that it compiles much nicer, and it's cheaper too. =)

Re:Actually....

[frankmanowar]

It's too true. Beer IS the best compiler out there. This will never change.

such a deal

[frankmanowar]

You know (shaking the finger of logic and wisdom), if they (MS) sold such a product for the GNU environment for a couple hundred bucks...I would buy it. And I know a few other developers who would to. Providing, of course, it didn't totally suck!

Re:such a deal

[chunkwhite86]

You know (shaking the finger of logic and wisdom), if they (MS) sold such a product for the GNU environment for a couple hundred bucks...I would buy it.

Too bad it'll never happen. I'm sure if such a project began within MS, Bill (or some exec) would scream "Heresy!". Remember, a win for GNU = a loss for MS in most cirsumstances.

Providing, of course, it didn't totally suck!

MS Sales Rep: "Well Sir, "doesn't suck" wasn't listed as one of your product requirements... Sorry, no refunds!"

Compiling under wine eh?

[trotski]

Last time I tryed to compile something under the influence of (way too much) wine, I ended up on the floor infront of my computer. All I remember was waking up to a formatted hard drive; compiling under the influence of wine sucks!

Re:Compiling under wine eh?

[delta407]

All I remember was waking up to a formatted hard drive

Would you prefer to wake up to an unformatted hard drive?

Re:Compiling under wine eh?

[jsse]

I ended up on the floor infront of my computer. All I remember was waking up to a formatted hard drive;

You fsck'd it! You bastard!

So now can i expect..

[OmniVector]

wine compiling cygwin compiling wine compiling cygwin compiling....

Re:So now can i expect..

[pizzarobot]

The article is here.

WRONG!

[www.sorehands.com]

It may eliminate the need for a reasonably fast machine to develop on, but you always need a target machine for testing! But, the test machine should be slow so that one can find performance bottlenecks and see the program operate under non-optimal conditions.

If the people are forced to test applications on slow machines, we may not have word processors that need 40MB of ram and a 933MHZ pentium III to run.

Re:WRONG!

[chunkwhite86]

but you always need a target machine for testing!

Either that or VMware...

Re:WRONG!

[burns210]

"...but you always need a target machine for testing!"

If i understand you right, you are saying that even with visual studio under WINE, the developer still needs a windows box to make sure the darn thing works? Well, if Visual Studio runs under WINE(as does office) then most of the average programs will to. But in an extreme case, there is always Bochs.

As for the part about testing on a slow computer, I agree. Having rediculously high minnimum requirements for some programs just seems like sloppy programming. Again, if you are using Bochs, then your slow machine is covered, since it is emulating the x86(thought version 2.0 is much faster than previous version).

Re:WRONG!

[spinkham]

He was testing on a seperate machine, just wanted to avoid the hassle of transfering all the sources every time he wanted to compile.
Quote from article:

The transfer of source to an MS-Windows machine and the correction of filenames and text format issues have been avoided. Therefore, I can build the game from one machine, then I only have to copy the final chaos.exe to my MS-Windows machine to test.

Re:WRONG!

[Jah-Wren Ryel]

Duh! If it compiles it must be fine! What kind of newbie developer are you?

Re:WRONG!

[Ryan Amos]

Do you work for Microsoft? :)

Re:WRONG!

[Tony-A]

I don't think even Microsoft is that stupid.
(But I could be wrong;-)

Re:WRONG!

[radish]

Yeah, it's not as if Samba does all that for you or anything. I move files back and forth via a samba share all the time and there are zero issues with names, line termination or anything else. There maybe other valid reasons he wants to compile under vmware, but that's a crappy one.

Re:WRONG!

[AnyoneEB]

You can do that under Windows too, just nest enough copies of cygwin and wine until your computer is sufficently slowed down.

Re:WRONG!

[crath]

He was testing on a seperate machine, just wanted to avoid the hassle of transfering all the sources every time he wanted to compile.

Too bad he's never heard of SAMBA or NFS. Mounting file systems and maintaining one source tree is a far better method than the hack the article describes; although maintaining a CVS tree and checking the code in and out over the network would have been even better.

Re:WRONG!

[Jorrit]

SAMBA and NFS only work if the two computers are on the same network. That's not always the case.

Greetings,

Re:WRONG!

[bluGill]

It does not require a slow machine to find performance bottlenecks, it requires a fast machine! Anyone who wants to improve the preformance of code needs a profiler to find those places, but profilers slow down your code themselves, so you need a faster machine to get the same speed. In theory you can find the slow spots on a slow machine, but you are wasting a lot of time, a fast machine can find the slow areas in the code faster.

That isn't to say you never run on a slow machine, but if you determin it isn't fast enough on a slow machine you need the fast machine to find those slow areas.

If you try to improve performance without a profiler you waste your time optimising parts of code that are plenty fast, while often ignoring the part of code that is the problem.

One man I know optimised every line in his program that was too slow, and got a 1% speed improvement. After that he used a profiler and discovered his program was spending 60% of its time doing arctangents, which is done in the library and was optimised by smarter people than him. Excpet they assumed you would build a bridge with their function so they used newton's mythod to get about 15 decimal places of accuracy. His program didn't need that much acuracy, in fact the first thing it did after getting the arctangent was chop off everything after the decimal place. By implimenting his own arctanget he was about to get his program to spend 2% of the time doing arctangent, and the program ran twice as fast. However without a profiler he could never have discovered the problem.

Re:WRONG!

[GlassHeart]

Anyone who wants to improve the preformance of code needs a profiler to find those places, but profilers slow down your code themselves, so you need a faster machine to get the same speed.

No, that is almost completely mistaken.

A Pentium 4 is not simply a faster version of the 80386! The fast CPU may have extra execution units, different pipelining, more (visible or hidden) registers, more cache, better memory or I/O bandwidth, etc. Any of these improvements can completely change the location of the bottleneck, or even show no obvious bottleneck at all.

Just to take an example, a common cause of slowness is that a tight loop doesn't fit in the cache. The correct fix is to make that code small enough to fit. If profiled on a machine with a bigger L1 cache, you may not even notice the problem, and will waste your time fixing something else.

In theory you can find the slow spots on a slow machine, but you are wasting a lot of time

Which is why you can either analytically predict some suspects, or use a fast machine to help predict the suspects, and then profile only those suspects and not the entire executable. In any case, it can't be stressed enough that architectural differences on the fast machine cannot be overlooked.

Your point about the absolute necessity of a profiler is well placed, however.

Re:WRONG!

[jelle]

But that test machine should be a plain windows install without the compiler environment installed on it, because your project might just work only with those special msvc++ dlls...

Re:WRONG!

[jelle]

Yes you are right, testing has to be done on slow AND fast machines. In a multithreaded application, an example of an annoying performance bottleneck is a deadlock where all threads end up waiting for another, which may only occur in particular environment (CPU, memory, network speed/latency).

Testing must be done on as many configuration variations as possible, and of course, experience helps too.

Visual C++ under WINE

[kruetz]

Wine has really come a long way to facilitate running major applications such as Visual C++. Features that "just work" often do not get mentioned because there is nothing to say. Wine has many excellent features like this. However, I have expressed the problems with Wine currently and I expect that in a potential follow up article many of these will be resolved. Wine has been in development for over a decade now. As it is finally nearing a 1.0 release, I see how much better it was than the 1.0 release of MS Windows.

Using Visual C++ on GNU/Wine gives me all the benefits of being able to develop a 100% compatible MS-Windows version of the game, while saving me the time of maintaining another Win2k machine version of the source and moving to that machine to compile. It has been a great time saver for me and I strongly expect this information will be very useful to myself and others in the future.

Okay, so you can use Visual C++ compiler under WINE. Is that terribly surprising when WINE can run MS-Office for the most part? The compiler takes the source files and libraries and produces an executable or library. I don't know for sure, but I wouldn't think that too much of this would involve heavy usage of the Win32API, much less the lesser-used and less-tested-under-WINE parts. For the most part, the compiler would be doing tokenising, parsing, translation and optimisation, which would in all likelihood use no external libraries or anything.

I don't mean to rubbish this article, I'm just saying that I don't see it as being terribly surprising. On the other hand, I think this is a great use of WINE and is definitely more innovative that anythin I would use WINE for. And as he says in the article, there was a lot of fiddling around with command line arguments and environment variables. But if you're compiling from the command-line under Windows, it's just as bad (no, really).

A much greater "victory" for WINE would be to have the whole VisualStudio ensemble running. But I'm not sure if this is feasible, especially in the short-term. By "victory" I don't mean something along the lines of "Linux now allows you to run a quality IDE", because KDevelop and Eclipse are great IDEs. Instead, VisualStudio and Office are probably the most complicated pieces of software written by MS (excluding Windows itself) and for WINE to be able to run them both as if they were running under Windows would be truly a fantastic achievement.

Re:Visual C++ under WINE

[SN74S181]

I was happy back in 1996 when I got DOSEMU under Linux to run my EPROM Programmer (a Needham PB-10, the one with an ISA card).

I agree that it seems pretty trivial that a compiler works, as compilers mostly just push text and binaries around, they don't call much of the Win32 API.

Re:Visual C++ under WINE

[civilizedINTENSITY]

Well I downloaded the latest GVIM for windows this evening and finally installed the ole for visual studio. Not exactly VC++ under linux, but a little vi in the VC sure is sweet! I might even actually play with it now.

Re:Visual C++ under WINE

[HopeOS]

Okay, so you can use Visual C++ compiler under WINE. Is that terribly surprising when WINE can run MS-Office for the most part?

At first glance, you might think so, but it's more complicated than that. The Microsoft compiler CL.EXE has many peculiar dependencies that were not addressed immediately simply because the functionality was not in MS-Office. If they've gotten this running, it's an improvement over the last couple months. Compiling MS apps on Linux will be my primary use of Wine, so if this is working now, it's news, good news in fact, to me.

-Hope

Re:ads - I dont see em :)

[jgaynor]

I don't see me - just set ads.doubleclick.net to 127.0.0.1 in your hosts file. Granted thats not the only host you have to add, but there are regularly maintained lists out there of good people to block :).

Windows Compliant / Posix Compliant Drivers

[PetoskeyGuy]

I'm a programmer for a Supplier of Software to the US Government and we are actively looking to port our products to Linux do to the strong interest from our clients. Specific Offices decide what products they want to implement and how to do it. We are investigating Emulating Windows as an one solution.

I look forward to the days when any program that is Windows Compliant runs on any platform that supports the Windows API.

That being said, the real problem I have is DRIVERS for my hardware working under Linux. If someone could emulate Windows enough to use standard Windows Drivers, there would be no more reason to use Windows at all. I truly commend those braves souls and companies who write drivers for Linux.

Now if there was a Proxy that accepted Microsoft SQL requests and sent them to a PostgreSQL backend transparently we would be free of the beast and save the Tax Payers lots of money paying my paycheck, plus blow our competiters out of the water. :o)

Re:Windows Compliant / Posix Compliant Drivers

[hazman]

Hey, you know, if there is anybody who could pull off this Hardware Abstraction Emulation stuff, it'd be those sharp critters over at Connectix.....what's that, they were acquired by Microsoft?. Never mind.

Re:Windows Compliant / Posix Compliant Drivers

[MateIn4]

If you are reponsible for the drivers, then port them over to Linux. Try http://www.jungo.com/linux.html that might help you make drivers for both environments.

Re:Windows Compliant / Posix Compliant Drivers

[deviator]

Novell had something similar to this (SQL proxy) a few years back called SQL Integrator - however, it looks like it has faded into oblivion due to poor sales. An open-source GNU solution would be pretty nifty. It would have to handle different feature sets & emulate certain commands... perhaps something with plugins to add support for different database types. (Novell does something similar with DirXML - a product that syncs different directory services databases with plugins that support each type of directory, such as ADS, NDS, LDAP, PeopleSoft, etc.)

Re:Windows Compliant / Posix Compliant Drivers

[gregfortune]

Doesn't Postgres have an ODBC driver? Any reason why you aren't using it?

Re:Windows Compliant / Posix Compliant Drivers

[fizbin]

Hrm. Is there really a sufficient market for this? I'd think that some of the work that the freetds folks have done would be relevant. (Of course, that's getting free clients to talk to MS SQL/Sybase servers, not the other way around)

Re:Windows Compliant / Posix Compliant Drivers

[PetoskeyGuy]

I should be more specific. Our application has thousands of Stored Procedures - tablename_update, tablename_insert, tablename_select, etc, etc.

ODBC would work to look at a different database, but the bigest thing we need is to use the same Stored Procedure Language.

SQL specifies Data Manipulation, and Data Definition. If there was a standard Data Automation(?) section then we wouldn't be tied to the SQL server.

We've looked at converting things over to Oracle due to client request, but with the time required it would be cheaper to buy the client an MSSQL server.

Re:Windows Compliant / Posix Compliant Drivers

[gregfortune]

To my knowledge, nothing like this exists.. You might find it useful to place the "stored procedures" in a process resident on the server and in effect, invent you own cross database solution. It would almost certainly be slower than native stored procedures though...

GNU!

[Dunkalis]

Wow! So RMS wants us to simply call it GNU now? Dropping the Linux part altogether? Sorta hypocritical to me ;).

Or do you run Hurd? Which is technically the GNU OS.

The above is meant to be funny.

Re:GNU!

[absurdhero]

For all we know, he is running a BSD kernel with GNU system software. The point is, it doesn't much matter what kernel he is using. We can't even tell. So why write it down?

Re:GNU!

[sfe_software]

Wow! So RMS wants us to simply call it GNU now? Dropping the Linux part altogether? Sorta hypocritical to me ;).

Or do you run Hurd? Which is technically the GNU OS.

He does mention Linux -- Mandrake specifically. But I agree, calling it GNU is a bit confusing (he also calls XFree86 "X-Windows", and refers to Unix-style text files as being in "GNU format").

As others have mentioned, running the command-line compiler isn't all that much of a feat. I was a bit disappointed that he "hand edited" all the #include directives to correct the case -- something that could be done with a Perl in-place edit...

Compiling under beer is better...

[Dave21212]

...a no nasty headaches from the sulfides.

Try Maryland Wines, esp Bordy
and Beer, esp Oliver Ale !

Re:Compiling under beer is better...

[jmb-d]

That would be BOORDY.

Hope this helps.

Off-topic, yes, but still informative.

GNU computer?

[arvindn]

Does this guy really use the HURD, or is it just that he's an RMS mega-fan?

Re:GNU computer?

[Arandir]

I'm just wondering if I can use WINE to compile Windows apps under non-GNU systems. Like for instance FreeBSD, NetBSD or Linux...

CVS

[YoJ]

The author describes the problem he originally solves as being the pain of moving code between Linux and Windows, losing attributes, case problems, etc. The approach I take is to keep all code in CVS on my file server. I do compiling and editing on my personal computer; both Linux and Windows can handle CVS. This way you have to reboot into Windows for the Windows compile, but never have to worry about copying files or case changes.

Re:CVS

[Anonymous Coward]

Yea, that was my thought. I never copy any source files between any computers, just 'cvs up'. From his article, it sounds like he doesn't use any source control, so my guess he's never worked on a project with other people or shipped any heavily used software. Both require the ability to co-ordinate source code changes and manage releases from a baseline...not to mention a single place to do backups.

I've been coding Java for a while now, but back in my C/C++ days, I only used the M$ command line tools. They integrate pretty well with a unix-style command line build system. The reverse isn't true.

Re:CVS

[Minna Kirai]

That can cause problems: you'd often want to know that a change compiles+runs on both platforms before checking it into CVS.

Using CVS as the main communication link between the two platforms means that linux developers can easily break the microsoft build, or vice versa.

Re:CVS

[ragnar]

I suspect the cycle of check-in, check-out, compile is no worse that transfer-files, compile.

I'm no cvs expert, but I quasi-admin the CVS setup at my workplace (small team of developers) and we operate by the rule of only checking in code that compiles and for the most part works right. In this way, all check-ins should be bug fixes or enhancements. This makes the log handy to see the evolution of the code.

If the log was riddled with "testing on the windoze box" entries I think it would pollute the log. If I were this guy I would solve this by using Samba to mount a drive on the windows machine and operate on one location for the source. That wouldn't preclude CVS for its proper use.

moving source?

[Daengbo]

Can someone help me as to just why he needs to move his source back and forth between computers. I mean, I like that he's consolidated to one machine and all, but couldn't he just use a samba share or something to save the (apparently complicated) step of moving the source code from one machine to another. What's the real advantage here? (Real question, not a flame)

Re:moving source?

[anonymous cupboard]

I wondered about that too, especially if a Windows system is running with rsh or ssh then it is very easy to have a single Windows system as a compile engine rceiving requests from other systems.

Cygwin is great for this and even allows you to run your own makes and build scripts from Linux. The GPL restrictions only apply if you link to GPL code, which you won't if you are running Visual C.

You lose the IDE though, but generally you just want to spend most of your time developing on one platform and just checking that you can compile on the second, leaving the debugging on that platform until later.

I use Wine to compile but not for porting

[Captain Rotundo]

I use Microchip's pic assemler through wine, for a small piece of code I maintain that runs on a PIC that wasn't supported by any GNU/Linux assembler when I started. I also maintain a legacy version of a very specific proprietary MSDOS (actually we run DRDOS) program that was written with Borland C, hopefully I will be replacing the last running bit of that with a DJGPP compiled version soon, which of course can be cross compiled on GNU/Linux without the need for Wine and bcw.
I know what your thinking, but when a piece of software has worked flawlessly (well almost!) for 15 or so years, and is 'mission critical' it is very hard to drop a platform and move on. I am hoping to try out a move to Linux some day in the near future so that I can take advantage of new features and things that just arent available for DOS. But unless I can convince everyone else of the benefits I may be supporting dos for quite some time (I am the only software person at this company).

Re:I use Wine to compile but not for porting

[SN74S181]

Shouldn't you be using DOSEMU instead of Wine? Wine seems like overkill for a PIC assembler.

Re:I use Wine to compile but not for porting

[Captain Rotundo]

Well geting WINE to work was a simple "apt-get install wine" - I have not had as much luck with DOSEMU, and havent tried since the wine system works fine.

1.0 ?

[IanBevan]

I'm curious, what exactly will the "milestone" of version 1.0 of Wine actually mean ? I would doubt that it'll ever be 100% compatible with Windows 98/XP. So what feature list is the development team trying to complete before calling it 1.0 ?

I think that pretty much any other product would have been deemed a failure if it had endured a 10 year development life and not reached version 1.0. Unless of course we're talking about Duke Nukem Forever...

My experience of Wine is common to most people's I think; it looks like a great idea, but as soon as you try to run any non-trivial program, it simply locks up/doesn't work. I've looked at their website and looked at all the "passed" indicators on their test cases. That doesn't help me run my apps much though... do they need more test cases ? Are they simply too abstract ??

Just my $0.02 worth

Re:1.0 ?

[Papineau]

Before 1.0, let it reach 0.9 (0.8 was released circa 1994 or 1995). You can check the unofficial Wine 0.9 TODO for a list of features needed for the 0.9 milestone.

About the long development life... don't forget it started with Windows 3.1 as it's first target. Then, Win95, Win NT 4.0, Win98, WinME, Win 2K, Win XP came out. We're talking 2 different architectures (for some kind of operations at least), and some new features to implement at each version.

Also, the list of authors currently lists 557 different people (contributions vary from a one-liner to a complete architecture overhaul). The number of currently active developpers is of course way smaller, more along the lines of 30-50. Of those, the vast majority do it in their spare time. So a long development period is not an indication of a failure, since if it was nobody would work on it anymore.

The test cases (called conformance tests) try to verify that what Wine implements reacts the same way in Windows. Depending on the purpose of a test, it can be trivial or not, implemented in Wine or not yet. A whole lot of dlls don't have any test written for them yet, so yes, we need more test cases.

Re:1.0 ?

[Anonymous Coward]

Wine's not done until Lotus won't run?

Re:1.0 ?

[wolverine1999]

I managed to get Virtualdub to run but it did not work correctly. Hope they fix this. I hope they get some applications like Tmpgenc to work too.

Re:1.0 ?

[IamTheRealMike]

I'm curious, what exactly will the "milestone" of version 1.0 of Wine actually mean ?

It's mostly purely technical targets, for instance the wineserver protocol will be frozen, which means that (I think) you can build WineLib apps and upgrade Wine and they won't break. There are a few other targets, mostly esotoric.

Wine 0.9 is far more interesting, and will hopefully come out in the last part of 2003. It's focussed on being easy to use and setup.

A lot of problems people have with Wine are usually related to it being badly setup. Partly that's our fault, Wine isn't especially easy to setup right, and because it releases every month, distro packages are often out of date (and now redhat doesn't even ship it anymore). After being increasingly frustrated with Wine, I tried CrossOver (a slightly hacked up version of wine designed to be easy to use), and it worked so well I was blown away. Installing IE6 in Wine is a distinctly wierd experience :)

Anyway, I've been using WineHQ wine for quite some time now, but hopefully along with proper setups (DON'T use your pre-existing windows installation etc) Wine will get a much better reputation.

This is non-news.

[Doomrat]

Visual C++ doesn't do anything weird regarding Windows API. The IDE is a normal affair, and the compiler could be run without a user interface. It's really not testing Wine to it's limitations and the irony of situation is barely worth commenting on. This is non-news, the only thing this article achieves is to make Slashdot look like the anti-MS geeks with limited social awareness. Some things just aren't worth giggling at.

I don't even need to look at the poster to know that this is the work of micheal...

Re:This is non-news.

[IamTheRealMike]

Visual C++ doesn't do anything weird regarding Windows API.

I'm afraid a lot of programs, even seemingly trivial shareware apps, do wierd things regarding the Win32 API. It's not uncommon to find apps passing corrupted or bad data into the APIs, missing out vital calls, abusing COM/ActiveX etc.

For instance, look at RhymBox, a simple free chat app for Windows. Looks fairly innocent doesn't it? But this program is an IE DHTML application embedded using the WebBrowser control into an ATL C++ container. Hmmm.

In particular Microsoft apps have a greater tendancy than most to take shortcuts, or use badly defined behaviours of the API. Just because it looks simple, doesn't mean it actually is. When you've done some Wine debugging, maybe you'll rethink those assertions.

Re:This is non-news.

[JonK]

Out of interest, how does component reuse come under the heading of "wierd things regarding the API". It looks to me like the developer of RhymBox (your example above) is doing precisely what he or she should: using pre-provided functionality rather than rolling his or her own DHTML renderer.

Not that I'm denying that many developers write (or at least used to write) apps which play fast and loose with the Win32 API - hell, the amount of code which plain wouldn't run on NT is indicative enough of that - but I don't think your example's particularly relevant.

(Disclaimer: I am in no way associated with anyone having anything to do with RhymBox, although I do use both COM and messaging applications :-))

Re:This is non-news.

[IamTheRealMike]

Well, it looks just like a native app, so the fact that it uses IE to render its GUI is a tad unusual. OK, so "wierd" was the wrong word, perhaps "complex".

As an aside, IE is the king of undocumented usages of the API, so it kind of filters through.

Re:No! No! No!

[njchick]

There are bugs in Windows that VMWare may not duplicate.

Care to give an example? I think if you run Windows under VMWare, you get all Windows bugs plus maybe some bugs from VMWare.

That's quite different from Winelib, which is indeed a separate implementation of Win32 that may "lack" some Windows bugs.

Bullshit

[0x0d0a]

Use the Intel compilers, your code will be 50-300% faster.

50-300%? You're nuts. I've used both, and performance definitely varies...and if I had to choose one or the other as "generally producing faster code", I'd probably point at gcc.

Take a look at these benchmarks.

Gcc produces fastest code on 26 of the tests, icc on 9.

Furthermore, not all the optimization flags for gcc were being used (no idea why -fexpensive-optimizations wasn't used).

Re:Bullshit

[Ramadog]

Furthermore, not all the optimization flags for gcc were being used (no idea why -fexpensive-optimizations wasn't used).

With gcc when using -O2 and up -fexpensive-optimzations is automatically turned on

Try using gcc -v -Q -O2 a.c to see what options are turned on. The file a.c has to exist.

Re:Bullshit

[bm_luethke]

Not neccessarily the most fair test. First the test were ran on an athlon system. Intel compilers optimized mostly based on the Intel architecture.

For example, we ran several benchmarks at work on three computers with gcc 3.1 and the intel compilers. Basically gcc and intel were fairly equal on the pIII xeons (intel had the edge). Gcc was somewhat faster on the athlon 700, and the intel compilers blew gcc away on our p4 2.4 ghz.

So what conclusions can you make? neither intel or gcc are better than the other. As we expected it depends on several factors - one of the main is hardware (wow, who woulda thunk hardware affects optimization :) ). In fact on the P4's with multi-threaded floating point operations we saw well over 300%, none of the tests were worse than 100% faster. On the athlon Gcc was slightly better except in one case were it was signifigantly faster. Eventually we found that gcc 2.x or 3.x does not have good p4 optimization yet (we asked on the devel lists trying to get better numbers as we didn't want to pay for the intel compilers). Of course this was about 6-10 months ago so they could have gotten optimizations in by now.

If I had to choose one or the other as "generally producing faster code" I would ask "what hardware are we talking about". And that GREATLY influences the answer.

Re:Bullshit

[0x0d0a]

Huh, you're right. Thanks. I'm pretty sure didn't at one point, since I remember writing test code and going through the assembly to see how smart gcc was. The following contruct needed -fexpensive-optimizations:

if (a < 2)
{
if (a < 3) foo
else bar;
}
baz;
}

Re:Bullshit

[abdulla]

The fact is that the safest bet would be to choose GCC, and this will probably be even more true as time goes on, the best thing for probably both Intel and AMD would be to dedicate engineers to GCC to target there platforms specifically, this could be said for many other CPU developers as well.

Re:Bullshit

[bm_luethke]

no, the safest bet is to optimize around intel. A large portion of machines are intel. The numbers show that a programed optimised for an intel run pretty good in athlon, athlon optimized run crappy on intel. Gcc will only optimize for PIII's and below well. P4 optimizations also run pretty good (that is close to the optimized veriosn for them) on pIII's and below and athlons. Thus the safest "bet" would be to optimize around the P4 (as pIII vs p4 optimizations may make only 5-10% on a pIII/athlon but 100-300% on a p4 this is even more true).

Should the gcc people optimize well for the p4 (and they will) then I will say the bet goes back with them. Of course this is assuming that the only thing to worry about is speed of generated code (which in many cases is not the most important).

Re:Bullshit

[Anonymous Coward]

The GCC developers are on it already.

Now, is it really surprising that a third party compiler is not as upto date as a compiler from the CPU manufacturer?

Re:Bullshit

[bm_luethke]

hmm? no - it's not surprising. But OTOH I will not use an inferior product becuase "they are working on it" (and I also thought I was relativly clear I did not blame the developers of gcc, I said they will get it done in time and they will - they have a very good product that beats many of the commercial products out there). Nor will I pretend they have better performance than what they do. Nor will I allow a misinformed post to go unanswered. That is one of the great strengths of Open source - community feedback is used. Once they get it done - i realise it is not easy - I will be EXTREMELY happy to use it (and I have donated money to projects such as this from time to time). I think I made that clear with my posts - and if not, I do now.

Re:No! No! No!

[nyquil]

dont you have to install windows under vmware? ive always been under the impression that thats how it works. vritual machine, not os emululator

Even a man who's pure of heart

[YellowSnow]

and uses the preview right
can turn to a troll
when the typos roll
and the bottle is empty and light

This sounds really lame.

[NynexNinja]

Next you will have some moron compiling under vmware or bochs or plex86. If you don't care about the abstraction or the 60% degradation in performance, and the fact that you have to rewrite all your make files to use vc++ nmake, this is for you. Its a convoluted mess IMHO. Using cross gcc for win32 on linux is a much cleaner way of doing this.

Or if your using Gecko

[bogie]

usercontent.css

*[src*='ads.'],
*[src*='/ad/'] ,
*[src*='/ads/'],
*[src*='/Ads/'],
*[src*='dou bleclick'],
*[src*='us.a1.yimg.com'],
*[src*='ad vertis'],
img[src^='http://images.slashdot.org/ba nner/'] {
display: none !important;
}

You can add whatever else you want there as well. Things like

/* this hides the usual 468x60 Flash banner ads */
embed[type="application/x-shockwave-flash"][wi dth="468"][height="60"] {
display: none !important;
visibility: hidden !important;
}
/* this hides the not so usual but very annoying 728x90 Flash banner ads */
embed[type="application/x-shockwave-flash"][wi dth="728"][height="90"] {
display: none !important;
visibility: hidden !important;
}

Re:Or if your using Gecko

[wolverine1999]

How do you use this with Phoenix (stripped down version of Mozilla)?

Re:Or if your using Gecko

[trezor]

Or just a browser supporting CSS and user-css in genereal.

No need for a gecko, when a Viking will do :)

Re:Or if your using Gecko

[Nate B.]

Good info.

Now how about some info on blocking those annoying Java ads that start all the Java VM processes and leaves them running? Few things are more annoying. I even tried to file a bug against Mozilla in regards to this, but was basically told it wasn't an important issue.

Obviously you skipped the Wine

[kfg]

and went straight for the LSD.

And they *warned* you the brown acid was bad.

KFG

I mean for others

[www.sorehands.com]

Of course all my code runs right the first time. I was thinking of everybody elses code.

GNU/WINE?!?!

[Arandir]

What the fsck is "GNU/WINE"! Aaargh! It's one thing to give RMS some credit, but this reeks of puerile toadying in an effort to look sophisticated in front of other juvenile sycophants.

Repeat after me, "WINE is not a GNU project!"

Re:GNU/WINE?!?!

[cant_get_a_good_nick]

Yes, but now RMS is a Windows project (Google Cache). Will it now be Windows/RMS/GNU/Linux?

So you're saying my vi clone. . .

[kfg]

that needs 128MB of RAM and a 3ghz Itanium just to load in under 5 minutes is Bad Thing?

Damn, back to my ed clone I guess. I think I can get that to run under your specs if I leave out the SETI@home easter egg.

KFG

Re:So you're saying my vi clone. . .

[bofkentucky]

what is your project's name...emacs *ducks for asbestos-lined safe room*

Re:So you're saying my vi clone. . .

[kfg]

Ooooooo, cheap shot. :)

KFG

Re:So you're saying my vi clone. . .

[videodriverguy]

Somehow the words 'safe' and 'asbestos-lined' in the same sentence don't seem to fit together.

Re:So you're saying my vi clone. . .

[bofkentucky]

As long as it doesn't get to your lungs, you are pretty safe, which is why I can't figure out why people can work in offices while removal is ongoing, or in my case, attend school during it.

Re:License Violation?

[Jimithing DMB]

Oh please. Yeah, let's all cower in fear of Microsoft. Let's stop what we're doing because Microsoft might not like it. Let's always worry about being sued.

It might be a very real possibility, but it's one you should never worry about, just be aware of. It's as bad as terrorism, it only works if the target becomes terrified.

I do this with a DSP compiler

[statusbar]

The Texas Instruments Code Composer Studio for the TMS 6701 DSP is only available for windows. For this project everything else is compiled for Linux-386 as well as embedded Linux-PPC.

Running the command line tools under wine works fine, and now I am able to have one intel linux box compile all the DSP and PPC firmware as well as the Linux GUI code in a cron job from the daily cvs snapshots.

Doing this has made things SO much nicer. I would prefer if TI had a linux version of their DSP compiler, and will continue to pester them for one, but in the meantime WINE saves the day and allows me to NOT run windows for weeks on end!

My Win2000 anti-uptime is around 3 weeks now. I am weaning myself off windows!

Running command line tools like compilers is one of the EASIEST things for WINE to do - All they have to do is read and write files and allocate memory.

Biggest hassle for me: having to make special sed scripts to deal with mixed case file names in the auto-generated makefile dependency list that are incorrect since the compiler assumes case insensitive filesystems.

--jeff++

Re:A double dumbass to you.

[mabinogi]

If My application is not talking directly to the hardware, then coming across VMware specific behaviour is no different to coming across VIA specific issues, or Penitum IV specific issues, etc...

Hic

[motox]

I did it shometaimes and man it worksh great !

weird...

[dolson]

You don't need Wine if you know what you're doing...

dana@digory:battlepong$ make
g++ `sdl-config --cflags` -c sound.cpp
g++ `sdl-config --cflags` -c collide.cpp
g++ `sdl-config --cflags` -c ball.cpp
g++ `sdl-config --cflags` -c game.cpp
g++ `sdl-config --cflags` -c menu.cpp
g++ `sdl-config --cflags` -c player.cpp
g++ `sdl-config --cflags` -c randgen.cpp
g++ `sdl-config --cflags` -c init.cpp
g++ `sdl-config --cflags` -c main.cpp
g++ `sdl-config --cflags` sound.o collide.o ball.o game.o menu.o player.o randgen.o init.o main.o -o bpong -lm `sdl-config --libs` -lSDL_image -lSDL_ttf -lSDL_mixer
dana@digory:battlepong$ file bpong
bpong: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.2.0, dynamically linked (uses shared libs), not stripped
dana@digory:battlepong$ make clean
rm -rf *.o bpong
dana@digory:battlepong$ sh cross-make.sh
g++ `sdl-config --cflags` -c sound.cpp
g++ `sdl-config --cflags` -c collide.cpp
g++ `sdl-config --cflags` -c ball.cpp
g++ `sdl-config --cflags` -c game.cpp
g++ `sdl-config --cflags` -c menu.cpp
g++ `sdl-config --cflags` -c player.cpp
g++ `sdl-config --cflags` -c randgen.cpp
g++ `sdl-config --cflags` -c init.cpp
g++ `sdl-config --cflags` -c main.cpp
g++ `sdl-config --cflags` sound.o collide.o ball.o game.o menu.o player.o randgen.o init.o main.o -o bpong -lm `sdl-config --libs` -lSDL_image -lSDL_ttf -lSDL_mixer
dana@digory:battlepong$ file bpong
bpong: MS Windows PE Intel 80386 GUI executable not relocatable

It's MAGIC.

Re:weird...

[aCC]

Read the introduction: "[...] This article is different because I used MS's own compiler! [...]"

Re:weird...

[dolson]

Yeah, now let me ask you, aCC, WHY would you WANT to use Microsoft's compiler in Linux if you don't need to? What does it cost? Is it worth it to pay all that money to do something you can easily do already? Seems like a stupid and pointless thing to me.

Re:weird...

[aCC]

Using MS' compiler has - under very special circumstances! - some advantages, e.g. if you already have code that was written for it and you also have paid for the compiler. You can maintain that code under Linux then without needing a Windows machine.

I wouldn't do it and I especially didn't think it was worthy news, but your comment missed the point in this case.

Re:weird...

[dolson]

My comment didn't miss the point... The point was how great Wine is... I've run Visual C++ under Wine before. Whoopdedo. I guess I'll submit a news story about how I can now continue work on text files I wrote in Notepad all thanks to Wine. I mean, I could convert those files over to a format for Gedit or nano, but what's the point when you can use Notepad now under Wine!

Hurd?

[marcovje]

He keeps talking about this GNU OS thingy. Does Wine run on the Hurd already?

heh... wine...

[doooras]

i prefer to compile under rum

*hic*

Re:ads - I dont see em :)

[DancingSword]

or, kill 'em ALL off, with the addendum to HOSTS at...

http://www.smartin-designs.com/

( ads, tracker-sleazers, etc, all of 'em, and it speeds-up one's browsing, too. )

Re:I hope you PAID for the VC++ compiler

[cdn-programmer]

I paid for lots of operating systems too. I have 2 versions of OS/2, 2 versions of DOS, windows 95 and 98, 3 copies of NT 4.0 and a copy of 3.51 and a copy of w2k that I got stuck with because it didn't work and I had to refund the client's money. Microsoft would not refund what I paid them. Microsoft would not allow me to either return or exchange the unopened copies of Windows NT 4.0 either. I was a Microsoft dealer at the time and bought a 3 pack - which was about a month before w2k was announced. So I got stuck... then stuck again with w2k.

I have only one computer that runs windows NT so only one of these operating systems is in use.

I have red hat linux 6.1 server edition, OpenBSD 3.0 and Mandrake 8.1. I paid for all these too even though they are free.

Now, my desktop runs Debian and it is just great and I did a net install. Guess what, I don't know how to upgrade that RH 6.1 machine. Sure I can reformat the hard drive - but it contains a LOT of my work and I don't want to lose it. With Debian, an upgrade is easy.

In addition to this I bought 2 copies of Brief and I can't use it even though it is my favorite editor. I did a trial d/l of CRiSP and it is a wonderful editor, fully Brief compatible. I contacted them about licencing. I run 8 machines in my home and am a member of the local Unix users group and do consulting work. The CRiSP people told me I would need a "Licence Manager"! I decided to learn Emacs. It has a CRiSP mode and is looking better every day.

I bought 3 versions of M$ FTN and copies of M$ C as well. The last version of M$ FTN was such a botch that I could not use it. I have the Borland OS/2 C/C++ compiler. It was so broken I didn't use it.... but I paid for it. I have Borland C++ professional builder 4. I want to do C/C++ cross platform development.

I phoned Borland. I think they told me that I can buy Delphi for Linux. I think they told me that the language is supported on both platforms but that the API is different. The person on the other end of the phone didn't seem too knowledgable. So I am going into wxWIndows or gtk because these people actually SAY they are cross platform. Thus, I do not have any C++ builder code and didn't get my money's worth.

I have 4 versions of Oracle - paid for by the taxpayer (because I am one of their developers). The DOS versions were so bad we couldn't use them on that platform. The OS/2 versions were better but still not good enouf. It is not possible to install 8i for red hat 6.1 (even though that is the version oracle says 8I is for) unless you CAN STAND ON YOUR HEAD and read the paper What you need to knwo before you even THINK of installing oracle 8i . Imagine having to backdate the glibC and install an old version of Java from blackstrap just so you can get the installer to work. Why couldn't Oracle have put their developers into a room with 2 cd sets. One set - an off the shelf shrink wrapped copy of Red Hat 6.1, and the other set - a copy of the 8i source tree. And then make sure they don't have INTERNET access so they can't cheat and create a version that you can't install. Leave them there until they solve the problem or rot to death trying!!! Damit. I have lost MONTHS of my life solving other people's problems.

Guess what, I am porting the client's apps over to PostgreSQL. See ya Oracle!!! Goodbye!

So, of all the software I bought most of it did the job for a precious short time or didn't do the job at all - with the exception of Brief - which was just excellent. And what happened to Brief? Borland bought the rights, put it on the shelf and AFAIK I can't buy it any more.

So frankly I really don't care if that copy of VB ++ was paid for. Even if it wasn't, I do not feel M$ was ripped off . I feel I was ripped off.

I have been ripped off Over and Over and Over because I bought these products in good faith and paid for them with my hard earned money. Now I can't use them. Some never did what they were advertised to do. Others were discontinued and I lost out.

Let me tell you about the 2nd last copy of M$ FTN that I purchased. I found a bug where the conpiler just eliminated an "IF" statement. No errors, no warnings, and no machine instructions either. I filed a bug report with M$ with sample code. Months went by. I re-filed the bug. Months went by. Finally I phoned them and sat on wait for like an hour - and paid for this too. The rep told me the bug was fixed in the new version. So I bought it.

When the new version arrived I tried out the sample code from the bug report. The problem was still there. Oh, and the last version I bought couldn't be used it was so bad. The client dropped PC development ideas and switched to SUN.

What the Open Source movment has done is a refreshing breath of fresh air and IMHO it is the ONLY way that the programmers of the future are going to stand any chance of avoiding the endentured servant trap that closed source software creates. Anyone questioning this should realise that in order to use closed source software they have to agree to the licence terms and these are arbitrary and non-negotiable. If you find the terms unacceptable then you can't be a programmer. That is unless you can find the tools and the infrastructure you need elsewhere.

Well, we programmers are being attacked another way now.... patent law. If this gets too well established we won't be allowed to be independant. We'll have to work for a big company that has managed to negotiate cross licensing. Either that or pay the piper each time we want to go to the can.

Water flows down hill and the infrasructure required to build some of the closed source tools, like Windows NT/2K/XP, or Oracle or JAVA or say a compiler suite like Visual C++ or Borland Professional Builder is so great that it looks like an ocean. Oceans are hard to move. Oceans are also hard to re-create. There is no way that any company could recreate from scratch an operating system like XP. The environment that XP grew out of no longer exists. IBM tried and failed.

Over the years all computer companies with the exception of perhaps IBM,HP,M$ have failed. Who here thinks Microsoft will survive into the year 2050? Gates won't be in charge then. Who here thinks that another company is going to loom up that can invest the BILLIONS in infrastructure required to create an alternative?

The future of software has to be opensource because that is the only way for us programmers to ensure that we will have the tools we need and the right to use them.

If I add up the THOUSANDS I have spent on software over the years it illustrates to me that I should have been looking for better solutions a long time ago.

BZZZZT!

[Anonymous Coward]

...plus blow our competiters out of the water...

The U.S. Government already blows its competitors out of the water.

Re:No! No! No!

[Tony-A]

Seems like there have been some reports that Microsoft Windows is more stable under VMWare than on the bare metal. It does make sense that the virtualized system as presented by VMWare is more regular and less messy than the bare metal.


Yes, but. The bugs taken individually won't show up. It's when they get together (and breed or something) that you get troubles.

SCORE 5 : WORDY.

[Anonymous Coward]

Thanks for letting us know that you can't install Oracle 8i on Redhat 6.1.

Re:SCORE 5 : WORDY.

[cdn-programmer]

Well, even if its wordy, you did read it. For your information I CAN install oracle 8i in Red Hat 6.1. I have done it and dropped it because there were other problems. But it was not easy and you need a paid for tech support contract to do it - which I have.

On the other hand postgreSQL installed perfectly. Futhermore I haven't encountered any problems yet that require tech support. Go figure.

You don't need tech support for stuff that works you know. So how big is Oracle's tech support group?

To be frank, I think you missed the point. You can spend huge amounts of money on closed source software, and when you are done you will be owned by it. You will be owned because you need to walk away from your investment and your experiance and start all over unless you pay for the next upgrade.

So when you get to be say 45 years old how will you feel about having walked away from most of the work you did in your career and most of your experiance because this work is locked into closed source that a) no longer exists or b) has new licencing provisions that you can't live with or c) fell into disuse because the company that owns it went bankrupt or d) became obsolete because it wasn't maintained... just milked..? Should I go on?

How do you feel about paying a tech support contract for a product where the vendor has not allocated a single person for maintenance? And your company is investing over a million bux for development?

Debugging

[EnglishTim]

But how do you debug? You're going to require a Windows box to test/debug the excecutable on anyway...

Re:Debugging

[now3djp]

The video game is developed/tested/debugged on my GNU/Linux computer. When complete, the PS2Linux and MS-Windows versions are compiled and also tested for hardware deps/performance etc

The video game has a debug mode to log errors, if this did not detect a platform specific problem on MS-Windows we would use MS VC++ debugger, as we installed it with the compiler. However, as of yet we have not had to do this, as the game is SDL safe; all game specific testing is done on GNU/Linux.

Cheers

JG

mod this up

[cdn-programmer]

This is funny

Linux?

[Hamsterdam Vallon]

This article got me thinking about the application of emulating a GNU/Linux compiler, like cc, under the windows XP OS. The cigwin project is advanced enough for that I think. It would open a whole world to windows developers who thought about porting to linux. This way you don't need to maintain a complicated GNU/Linux computer.

This is not a troll or flamebait, just suggesting the best tool for the best job.

BSD, Linux, OS X, & Solaris folks should read and blog at *nix.org

Actually

[Sycraft-fu]

I think if you are serious about cross platform development, you need to have a native install of all the platforms you intent to port to. Either multi-boot or seperate boxes. Why? Well amy emulator, WINE espically but even VMWare, has certian quirks and differences that native hardware does not. You need to test it as it will be run to be sure. Also, if you are using any advanced 3d function, the emulators are going to fall flat on their faces.

If I were really serious about a commercial cross platform release, say Windows and Linux I think I would want a ton of testing. I'd probably want at least 3 different systems, all tht had a copy of Windows 98, ME, 2k, and XP plus at least 3 big Linux distros. While in theory something complied for any hardware under any verison of Windows should work on any other, it's just not the case. Same for Linux. You want to test for the little "gotchas" if you want to make sure you have a really stable software.

Re:Actually

[Sneftel]

It would certainly reveal that the repaint code was taking a large amount of time to execute, if in fact the repaint code took a large amount of time to execute. And if it didn't, then why bother optimizing it in the first place?

Re:Actually

[www.sorehands.com]

It was not the repaint code. The repaint code was being called twice.

I am not saying that profiling is bad. Profiling is only one method of many to use.

The best way is to use the force.

that statement no less valid for OS/2

[DABANSHEE]

Proving that OS/2's Win16 capability was no trojan horse, because W95 was exactly the same in that regard.

Re:No! No! No!

[jonathanclark]

I'm trying to reproduce a bug right now that does not occur in VMWare. This bug relates to thread creation/destruction, and it's very timing dependent. Because of VMWare's slow task switching mecanism, the bug never occurs when running on VMWare - but always occurs on a normal machine.

Dosemu 1.0

[tbspit]

When Dosemu reached 1.0, usage of DOS had become very low. So let us hope for a Wine 1.0 soon:)

Re:wow

[crath]

Now this is impressive. Things like this are what WINE should be all about. Amazing.

I disagree, it's depressing not impressive:

• The author first asserts that the process of moving files between the systems causes the upper/lower caseness of the filenames to be munged; however, the text later talks about how his personal preference for file casenames is all lower and so he changes them. There is nothing in his "transfer between Linux & Windows" scenario that causes the case names to be wrong. The real issue is simply that his personal preference for filenames differs from that of the original author of the source code.
• The author displays no knowledge of the network mounting of filesystems using SAMBA (CIFS) or NFS. This would have been a far better method and while it too would have slowed down compiles on the machine mounting the filesystem he states that he is not unhappy with such a slowdown.
• Why isn't the source code checked into a configuration management tool, like CVS? What the article should have been describing was the process he followed checking the application into CVS, making his personal preference changes, checking in those updates, and then checking it back out onto both platforms and performing the compiles there---with no compile slowdowns.
• As others have already noted elsewhere, he will still have to test on the target platform. Those who argue that VMWare is as good as a native boot-up for testing are simply displaying their ignorance and inexperience: display and network problems are often induced by a change in hardware platform, and VMWare is a distinct hardware platform from a testing perspective. While testing on VMWare is useful, it is not sufficient: a Windows application should be tested on several different hardware platforms as part of its formal QA.

In summary, while the article probably accurately describes the author's actions, there is nothing in his account that others should be emulating. More experienced developers should be consulted in the search for best practices.

I write this as someone who has done software development for almost 20 years; more than 20 if you count my high school years as a computer hobbiest.

Re:ads - I dont see em :)

[0x0d0a]

I use squid on my machine with the sleezeball filter system, so that *all* my browsers are taken care of when I add a filter.

Re:Awesome [Mod Parent Up}

[waterbear]

That looks like a pretty fundamental point, deserves higher score and visibility .....

Re:Warning: crack-smoking moderator

[njchick]

Maybe it's not offtopic, but still wrong because this thread is about VMWare, not Wine.

asshole...

[Ender Ryan]

Are you sure you know what you're talking about? Just how the hell do you know what vc++ does regarding the Windows API? That's right, you probably don't. Do you?

Furthermore, the point of the article is that this guy is now using Wine to do some real work, not that Wine has acheived some level of technical excellence that's meant to be impressive to YOU. There is one tiny blurb about how Wine has progressed.

Does anyone know for how long Wine has been able to run vc++, and what was required to get it working?

How bout you reimplement windows from the ground up before you start trolling about it.

Re:asshole...

[Doomrat]

Hello retard.

I'm not saying it's not an achievement for the Wine team, I'm saying that using an emulator designed for emulating windows programs with the fairly non-spectacular Visaul C++ isn't news worthy. Maybe I'm wrong - in which case, I just used Vi to edit some text! I'd better get to submitting that...

Re:No! No! No!

[njchick]

Next time try to read the comment you are replying to.

I don't get it...

[sayerofno]

What's the point of compiling win apps and linux apps on the same machine?

I develop windows apps for a living (hey...it's a living). Even in our 99% windows environment, I don't build _everything_ on my machine. We have an automated build environment. It's very simple to launch a little client app to trigger a build on a different machine. All of the source is checked out of source control, and then built. I wrote the "remote build" app myself so it was tailored to our companies use, but I would guess there are commercial/free utilities out there. If not, it ain't that hard to write a windows service that reads a file and executes command lines written in that file (which is what our remote build app does).

If you're going to have to have native windows machines to test on anyway, why not just set up an environment where you can build your windows apps on a windows machine while still sitting at your linux machine?

On the one hand, I think what was done was neat (sometimes just showing it can be done is worth while), but on the other hand, it doesn't sound like the most efficient way of doing things.

Nintendo

[pommiekiwifruit]

As slight counter-examples Nintendo, Sega, Sony ran that scheme and it works for them. Whereas anyone can develop for Windows but the software sells an order of magnitude less (I think by the time the PC had racked up its first million seller, Nintendo had had 70 titles break that mark). Cybiko and Symbian try to make things really easy for developers, but they still have difficulty fighting the appeal of a huge user-base.

OTOH some people actually reverse engineered the NES (using an oscilloscope and disassembler) and wrote working games on it, but then they still had to show their game to Nintendo to get permission to continue developing.

Safe performance

[fizbin]

The safest bet is to not have your job depend on tweaking the last little bit of performance out of a customer's machine when you have only a vague idea of which platform they might be running on. Either improve your algorithms in the whole, or hide the slow performing code in a place where the user won't notice. (if possible)

That said, another option is to isolate the performance-critical sections into chunks of code which get optimized separately and then have cpu-detection code choose the appropriate chunk at runtime. The debian atlas packages do something like this, though (AFAIR) they use hand-optimized assembly instead of just using compiler flags and/or different compilers. (and yes, you can have both the sse- and 3dnow-optimized packages installed on the same machine, and the code will load the appropriate shared library at runtime)

Then, instead of standardizing on a compiler, you'll need to standardize on a build system that will let you compile the code as you need it compiled.

Re:Safe performance

[bm_luethke]

well, I work in high performance computing (cluster computing more specifically). So squezing the last little bit out of the machines IS my job :)

Of course, I know exactly what the hardware for my clusters is (if you can't guess, I ran an athlon cluster, a pIII xeon cluster, and a p4 - the afore mentioned machines the benchmarks were ran on). When one of a users job takes 4 days to run even a 5% speedup is signifigant (much more signifigant than a recompile). They also tend to try and optimize thier algorithm based on architecture and how the specific compiler optimizes code. For the ones that are good at it it's really neat stuff.

I perfectly agree with what you are saying when it is about general purpose code. Unfortunatly (or fortunatly if, like me, you enjoy doing it) there are a handfull of jobs where this stuff becomes really important.

Re:Safe performance

[fizbin]

Well, that seems like another argument for the strategy of splitting out the performance-critical portion of the code into modules that are then compiled several different times, targeted at different cpus, with cpu-detection code loading the appropriate version at runtime. Then your job could run across all three clusters at once if necessary, yet still enjoy the optimized goodness as though you knew what cluster it was running on ahead of time. Of course, the development time overhead of optimizing for all the different cluster types at once would probably only be worth it for components that were heavily re-used. (by that I mean the programmer time necessary to sort out the build system and cpu detection code to have the right thing happen automatically, not the recompile time)

In fact, the reason that the Debian atlas (Automatically Tuned Linear Algebra System) packages do just this is that the packager was faced with the problem of high-performance, cpu-intensive code running across a heterogeneous cluster. (the basic atlas install assumes that you will install from source and choose the proper flags for your machine) Once he'd worked out this system for his employer, (well, once he and several of his coworkers had ...) he got permission to package the solution up for Debian.

Of course, if it's just as easy to compile the whole darn source tree for each target machine and then have a shell script run the correct program based on e.g. `uname -a` or /proc/cpu, then this approach and yours merge back into essentially the same set of actions.

Actually

[www.sorehands.com]

Actually, once running under Softice with virtualization turned on, I saw that the screen (DOS app) was being being repainted twice on every screen refresh.

I doubt that profiling would reveal that.

Re:wow

[crath]

Or,

Compile under VMWare, or
Still compile under Wine, or
Dual boot the computer.

What!?

[rbullo]

Win95 had excellent Win3.x compatibility...

Perhaps you should replace "Win3.x" with "MS-DOS". As I recall, Win3.x apps worked horribly under Windows '95, due to programmer's shortcuts that Windows '95 won't allow. DOS programs, however, often work better under Windows 95 then they do under Windows 3.x.

You are ignorant.

[Tord]

Your dissection of the authors knowledge is strongly flawed.

The author first asserts that the process of moving files between the systems causes the upper/lower caseness of the filenames to be munged

This is the case, I've run into this problem myself many times and it has little to do with how the files are saved and much to do with how fs-drivers are implemented and configured. Compare to how you often end up with read-only files if you copy them from a CD in Linux. Sure you can change some settings, but then often get other side-effects in other situations (goes for both). I believe he is the original author of his project.

The author displays no knowledge of the network mounting of filesystems using SAMBA (CIFS) or NFS.

The fact that he is dual-booting strongly suggests that he (like most of us) only has one development machine, not a complete network of machines. Don't get ignorant just because you are better equipped.

Why isn't the source code checked into a configuration management tool, like CVS?

Once again, he only has one workstation, thus is CVS out of the question since it needs to be hosted under one of the OSes and he can only run one at once. Besides, CVS is in most cases just unnecessary and complicates and slows stuff down if you are the single developer. You can make backups through normal file copying.

As others have already noted elsewhere, he will still have to test on the target platform.

Which I'm sure he does now and then too (by dual booting). However, if you just need to do some minor change in platform independent code it's really a bliss to not need to dual-boot.

Besides, your criticism of using VMWare for testing is quite irrelevant. It's true you need to test on multiple environments to know that it works on them all, but as you said, VMWare with Windows X *IS* one of these environments. If he's dual booting he only has access to one environment anyway, using VMWare as well will add one or more extra testing environments.

I would have LOVED to have a similar setup as his when I was developing BladeEnc. Like him I only had one development machine (couldn't afford more) and constantly needed to dual-boot in order to recompile and package a new version. The platform dependent parts of BladeEnc were very limited and untouched for 95% of the development, thus this setup + testing under wine (for possible quality degradation due to compiler excentricities) would have been more than enough during most of my days. Only performance tweaking would have to be done in windows environment.

With your 20+ years as software developer you obviously have found your way of working in your projects and with your budget. Don't knock others creative solutions for solving their problems with their resources in a totally different situation.

Re:You are ignorant.

[GlassHeart]

CVS is in most cases just unnecessary and complicates and slows stuff down if you are the single developer. You can make backups through normal file copying.

By "backup", I presume you mean the kind to recover the code you wish you didn't overwrite while drunk last night, not the kind due to a disk crash.

It's a lot easier to get lazy about backups. With proper source control software, your source files aren't writable unless you check them out, so you're forced to remember.

Secondly, and I admit it's probably rare among free software authors, source control makes it a lot easier to support bug fixes to older versions of released software.

You Are Wrong

[HopeOS]

This is very big news. Microsoft's compiler CL.EXE makes many non-obvious Win32 API calls before it gets down to business. These calls were not implemented immediately because they are used by very few programs. Even after the calls were implemented, CL.EXE would routinely crash (as of maybe October or November 2002). If the Microsoft compiler now runs on Wine, we have achieved a major step forward.

-Hope

Using UNIX command line to compile with MSVC

[dgk]

You should be able to use the cc front end that comes with UWIN as a front end for MSVC that you are running with WINE. This will allow you to use UNIX Makefiles to compile using MSCV. It also
will allow you do use UNIX style file names
and compile options.

I haven't tried compiling the cc front end to MSVC on UNIX/Linux, but you should be able to compile it using MSVC and run it under WINE.

The source code is part of the UWIN development package which you can download from http://www.research.att.com/sw/tools/uwin/.

why?

[Ender Ryan]

Why are you so intent on being an asshole? You don't even have a fuckin' clue what you're talking about? You don't even know what parts of the windows API vc++ touches. The article wasn't even about that anyway. The point of the article was that wine enabled this dude to do cross platform work on linux, easier. Maybe that's not interesting to YOU, or to me TBA, but then why the fuck were you reading it.

Then you go on to say, "only thing this article achieves is to make Slashdot look like the anti-MS geeks with limited social awareness".

What the fuck? This article had nothing to do with MS bashing, it was about doing windows development on Linux! Again, what the fuck? Then you go on to insult Michael... for no apparent reason.

You're just trolling, stop it.

5% desktop share?

[fymidos]

I am sorry when was that? i never even saw OS/2 on any desktop computer!!!

OS/2 was never a desktop OS and never had a marketshare there. It came bundled with ps/2s' and that's all it ever got.

What really killed OS/2 is that you couldn't get it for free.

i've been around computers for 1 1/2 decades, and i NEVER saw anything open loosing against anything closed. I could get a windows 95 cd anytime but i couldn't find an OS/2 cd before warp.

Re:5% desktop share?

[dryeo]

I am sorry when was that? i never even saw OS/2 on any desktop computer!!!

OS/2 was never a desktop OS and never had a marketshare there. It came bundled with ps/2s' and that's all it ever got.

What really killed OS/2 is that you couldn't get it for free.

i've been around computers for 1 1/2 decades, and i NEVER saw anything open loosing against anything closed. I could get a windows 95 cd anytime but i couldn't find an OS/2 cd before warp.

Huh? IIRC I got a copy of OS/2 2.11 for the price of a stamp (maybe a couple of bucks for shipping) just by filling out a form in Byte or somesuch and I'm still using OS/2 (now ver 4.5). Windows seem like crap everytime I use it and Linux I've given up on as it becomes slower and more Win like. Its easy enough to recompile *nix programs I want or run them over X

Dave

Re:5% desktop share?

[rjch]

I am sorry when was that? i never even saw OS/2 on any desktop computer!!! OS/2 was never a desktop OS and never had a marketshare there. It came bundled with ps/2s' and that's all it ever got.

You've obviously never paid any attention to what corporate users are doing. My current job involves auditing hardware for a large bank and OS/2 has only recently been phased out of the branches and is still widely used in the datacentres.

Moreover, OS/2 was fairly popular with the larger BBS sysops before Win95 showed it was stable enough for the job. My own desktop was OS/2 for years and I wouldn't have had it any other way. It was (even back then) more stable than Windows is today.

No, it never got widespread acceptance as a desktop operating system. More is the pity. I have no doubt we'd be bitching about IBM being the monopoly today, if it had, but at least we would have had an operating system worth using.

Re:wow

[msebast]

Tord covered most of the things wrong with the parent but missed something about the testing.

If you read the article (I know who would do that...)

> Therefore, I can build the game from one
> machine, then I only have to copy the final
> chaos.exe to my MS-Windows machine to test.

He is not doing his testing with Wine!

Re:Don't use the main repository

[nobodyknowsimageek]

My solution is much simpler. I use rdesktop to log into a terminal server, then mount my workarea from my Linux box (running Samba) and do the compile that way. In fact the other day I was forced to mount the CYGWIN tools from another machine, so that the only thing coming from the Windows server was VC++ itself. It was a little slow, but I usually test the Windows version last, so that cuts down on the iterations.

Re:wow

[now3djp]

crath,

You seem to have missed the point of my article. Tord also picked up on this in his reply to your msg.

Basically I set out to save my time, the novel solution that I came up with used tools freely available and did not distract me too much from developing the video game port. Please re-read my motivations section, this is exactly what I set out to do.

You mention SAMBA, NFS and CVS. Well I use these there great technologies every day. I my use of CVS was clearly visible from the end of the article where I mention the new Wine CVS version I tested (did you get that far?). Also, this was never intended to be a "how to copy files over the network" article. You mention I will still have to test, yes that is what I chose to do. Saves giving all those games testers compilers they will never need to use now! I never mentioned VMWare, and I do not intend to test with it. For my purpose it is not relevant, I achieved all I set out to. (build from one machine, avoid wasting time etc)

Some people mentioned the possibility of using the entire msdev.exe IDE. Unfortunately this does not run with current CVS Wine, a few more dlls are required to even get it just to load. However, it crashes when build is selected.

I have received lots of positive feedback, and if you are reading this considering this as an option for your self or company then I would definitely recommend you try it. Most of the Wine problems have been corrected since this article was released. They should be in the CVS version over the next few days I expect. Credit to the Wine team.

Cheers

J. Grant

Re:Warning: crack-smoking moderator

[Jimithing DMB]

Yeah, except for the fact that the post I was replying to specifically was talking about the differences between Winelib and real Windows (under something such as VMware).

So actually, it had direct relevance to the parent post as well as direct relevance to the topic at hand.

This message on the other hand, has absolutely nothing to do with the topic, but I wanted to thank the anonymous coward for coming to my defense, and point out that you and whoever moderated the message offtopic were the ones who apparently can't read.

I am posting this with karma bonus turned off so that it doesn't make it's way to the top of the discusson needlessly. Maybe I should just stop posting to Slashdot altogether. I held out for years believing, but apparently there's just as many morons who read Slashdot as there are in the real world. Quite a sad picture for a site that should be for the geekiest of geeks.

Re:License Violation?

[Jimithing DMB]

Actually. I agree with your statement about free software. You are the one who didn't apply my statement to free software as well as proprietary. I, on the other hand, am fair and applied it to everybody.

No one should get sued because they included a one liner from a a GPL program. And so far, the cases have been ones where there was a clear usage of a serious portion of code. I don't think the FSF intends to go after every little person, nor do I think it would be beneficial to their cause.

I do however think that gross violators of the GPL should be prosecuted, and likewise I feel that gross violators of any real (upholdable by law) license terms should be prosecuted. It rarely does any good to sweat the small stuff.

Whether you realize it or not, your argument that we should all be afraid of including even a snippet of GPL code is a classic FUD tactic because it attempts to make people afraid of working with GPL software.

Re:License Violation?

[Jimithing DMB]

:-)

In all seriousness, every group uses scare tactics. Some are more prevalent than others. For instance, terrorism is by definition all about scare tactics. The mob also uses scare tactics. And in this case, we're talking about large corporations using scare tactics.

Scare tactics are nothing new. It's helpful to recognize where scare tactics are used so that you do not fall victim to them. For instance, we live in this supposedly "sue happy" country. In reality the "sue happy USA" is a scare tactic used by those with money and power to make those without submit to their wishes. By modifying your behavior to avoid doing things which are not illegal, but which somebody might not like, you are inviting that somebody to run your life.

Fortunately, it is still possible in this country that once you understand the above, you rise above it and then you become the person with money and power. What you choose to do with it is then entirely up to you.

Re:why?

[Doomrat]

Calm down, nutjob.

I didn't say it's a non-achievement, I just said that I thought it wasn't newsworthy. Maybe I was wrong to say that MSVCPP doesn't use strange API, but still I believe that there is nothing particularly spectacular about the program. I really feel that this sort of thing is only posted on Slashdot because of the Microsoft connection. It's not like the editors don't have a reputation for being sad zealots.

Clearly there are two schools of thought on the matter, shown by the people who have replied to my post in anger, and the people who have modded it up. If you can't handle an opinion without having to swear (and indicating your accent by dropping the G off the end of "fucking" makes you seem oh-so-more dangerous on the Internet!) and call names like a pathetic moron, then just don't bother. I appreciate the other replies I had, but yours was just laughable.

Re: say no to non standard

[Zeio]

Just to know, ICC has been known to optimize code for AMD processors quite nicely.

An application should compile nicely anywhere, for example, Apache. You name it, gcc, icc, forte - it compiles.

Also, applications should be developed against a standard c library/headers. None of this winsock or strange SYSVisms or /usr/include/linux stuff.

GCC (with some interesting non-standard extensions) and non standard platforms like RedHat makes for difficult porting. Bad discipline and non-standardness is what created a place in the world for the giant, titanic Java. It had gotten to the point where it was easier to port a VM to here or there than to simply compile the application somewhere else. We all know speed suffers for the abstraction which is of questionable value.

So, which compiler should be targeted? Which CPU? Well, unless the compiler is broken or the CPU has severe errata, all of them.

Premature optimization is the root of all evil, so making the first order of business deciding which compiler, platform and CPU the program is for is just that. Why do you think Knuth gives examples in a mythical assembly language for a mythical processor? (Partly due to the fact that it was mythical in a superior sense at the time he came up with it, but also to point out that proper algorithmic implementation and proper programming are architecture independent.)

YAY! Saves me a step, maybe?

[$$$$$exyGal]

Very good show ;-). I still maintain 3 boxes at home (windows, linux, and mac), but maybe now I can get that down to 2! I rarely want to develop a windows program, but when I do, I always have to boot up Windows on my second Linux box, and create and compile my program, and then boot back into Linux, test my program in WINE, etc. This will hopefully save me a step ;-).