I'm saying that when the FSF was confronted with KDE licensing trouble, they launched two projects - one intended to embrace and make liveable (the free qt project, which died I believe), and one intended to entirely replace (gnome).
Such a dual approach might make sense with.net: start one project which is a free implementation of.net, and another with similar goals that doesn't owe a single thing to microsoft.
BTW, what about the PKI for OpenPGP? Could it be augmented to do this somehow?
Ok, I really like mozilla 0.9.1. mozilla has been my main browser for quite a while, and recently it became able to talk to a slightly unusual https site I need access to, so I almost never run netscape 4.x anymore.
However, that damn installer is keeping me from downloading mozilla 0.9.2, much as it used to keep me from downloading some of the netscape 6.x releases.
Is there some trick to getting it not to segfault? It seems phase of the moon for me.
Is there some way to download mozilla without the silly installer?
I have a palm IIIxe and an Agenda VR3. I actually use the IIIxe and not the VR3, but if I write any software, it'll be for a linux-based pda. I may end up getting a yopy though, as I prefer GTK+ (yopy) to FLTK (vr3).
BTW, the VR3 costs not that much more than a IIIxe these days, I believe. If you don't mind fltk, and plan to do some programming, it'd be a good way to go.
iPAQ is a poor choice, because Compaq refuses to sell them without wince preinstalled. So if buying is voting with your dollars, every time you buy an iPAQ and put a linux variant on it, you're voting for Microsoft world dominance.
If you need something today, I'd take a serious look at the Agenda VR3. It's gotten twice as fast recently, since the EIP feature, and there was some talk of optimizing gcc for the machine better too at one time... It appears to have a pretty active developer community. I'm not that enthused with the FLTK gui toolkit it's using, but it could be worse.
I'm hanging some high hopes on Yopy though. GMate isn't funding Microsoft or contributing to Microsoft's marketshare numbers on every purchase, and Yopy appears to be targetting just the right technologies now: X, GTK+ and python.
I'd never consider teaching students C++ as a first language. C either.
Python yes. Java yes. Perl no. ML would be reasonable, as would Eiffel.
C and C++ need to be taught, but they should be in a Junior level course.
If I were in a "yeah, Java's not the best, but it does have a good combination of reasonable power and critical mass" mood, I'd teach java first. If I didn't care about who's using what at that moment, I'd teach python first.
Actually, such a protocol would need an option like:
Complete transmission only if all hops are encrypted? y/n
...because some mail has to get through, but it'd be nice if it were encrypted, while other mail shouldn't be transmitted the rest of the way if it can't go the whole way encrypted.
Er, actually, I avoid learning about Microsoft software about as much as I can, but I believe that in 95 (not sure about 98) and I believe even some early SP's of NT, passwords aren't sent cleartext, but replayable hashes are sent in cleartext, which is about as bad.
So you can't just log in by typing someone's password after stealing it, by you can write a program that'll send the replayable hash at the right time when asked.
Ok, I'm expecting some rabid disagreement on this, but the time period of my life when I had the least back trouble, was when I was doing deadlifts regularly. Since quitting, I think I have back trouble less often than I had it before my stint of deadlifting, but I am getting it a little bit again.
A lot of folks say you should never deadlift. I tried it for a while, and got to really like it. It thickened my "love handles", which isn't the most aesthetic thing in the world, but folks, that muscle is stabilizing, and it taught me proper form for picking up heavy objects.
It's good your soliciting ideas. You'll get a mess of conflicting opinions, but as long as you do what you think is best in the end (giving a coherent whole), taking suggestions can only make your language better.
I want to comment on one of the least important bits semantically, that has a big payoff: avoid those damn space-wasting, error inducing curly braces or begin-end pairs. If you do it where all if's require and end, ala modula-2, that'd be great. Or even better: do it in the python style, where the nesting is taken directly from the indention, eliminating a really pesky class of errors as well as all the wasted vertical space (or alternatively, eliminating those obnoxious curly braces that don't line up).
More importantly, study what's already out there. Don't do a language because you want to do a language, do a language because you think you have just the right combination of features taken from other languages, and/or have a novel idea that you feel the world would benefit from.
Oh, and -don't- go the perl route, for Pete's sake! What a mess, and getting worse. An ideal language is simple enough for an idiot, but still exceptionally powerful in the hands of an expert - complexity should come from algorithms, not a morass of alternative ways of doing things In The Language. Operating overloading is about as bizarre as the language should allow.
Also, if you do have a novel idea, make sure it wouldn't be better added to an existing language that already has some critical mass than to a new language. Some ideas could simply be implemented as a class library, while others could be added to a language proper by an open-minded project coordinator. I'd guess python or ML (Meta language, _not_ machine language) folks (two languages I respect quite a bit) would be very open to hearing of a novel idea, and I'm pretty sure there are others too.
I work in a centralized support organization at a university.
Some of our clients are happy to upgrade whenever. Ok, that's stretching the truth, but it isn't that hard to convince some folks.
I'm not writing about them. I'm writing about the people who hate upgrading.
Some of our clients just refuse to upgrade. "We must have a stable system".
This wouldn't be so bad if it weren't for the facts that:
1) Redhat releases often - about every 6 months it seems.
2) Redhat stops releasing patches for an OS release fairly early.
This combined means that anything but a x.y release, where y is the last minor revision of the x major revision of a redhat os, doesn't work so well for our clients (because redhat continues to release patches for a decent amount of time for the last y of an x, but not for earlier y's).
So for a while, we decided we were only going to support the last y of an x. However, naturally, clients started saying that was just too old. "Why do we want to install something 2 releases back? What good does that do?" We can explain all we want, clients just don't understand.
So I'm asking, can redhat consider either releasing less often, or continuing to support releases with patches longer? If neither of these work for redhat, do you have any suggestions for us that might help our situation?
I should add that upgrade procedures don't work well with the heavy sysadmin automation we use here. We always reinstall from scratch, and our many scripts (in one unified framework) put back the tweaks.
It seems to me the whole idea of a rebate is a scam.
What good are they?
The only thing I can think of is that some people will think "Oh good, $1000 with $200 rebate, that's really $800", but when push comes to shove, they neglect to do the _extra_work_ to get the rebate.
Sure, most people will probably do the rebate, but if 8% don't, then that's a lot like the company got to jack their price, without nearly the same reduction in sales.
Now, if they do rebates, and go one step further to make doing rebates an even bigger PITA... Perhaps that's even better for the company.
That is, until people start to wise up and avoid the extra unnecessary work associated with rebates. Personally, if I see a car after rebate for $n, and a car without rebate for $n+100, I'm probably going to do the car without the stupid rebate. What a lot of arbitrary busy-work. If they really want my dollars, they should include the price reduction in the sales price. I mean, really, why wouldn't they?
It is an irony of the computing industry, that when someone sees multiple factions in software for providing a function, that people decide to solve the problem by introducing another faction.
Consider the tar and cpio factions, and "pax", the faction to end all (archiver) factions.
So what's wrong with calling your coworker with a message (EG "I like green tea") and encrypting that message with your private key (using gpg or pgp), and putting your public key up on the web (giving the URL to the public key over the phone) or putting your public key on the keyservers?
Then if your overseas coworker can decrypt the message and get the correct text they learned of on the phone... it would seem this would go a good ways toward establishing trust.
Sure, someone could snoop the overseas call and send the same message encrypted with their own private key, but they'd have a hard time taking down your website (which has the public key) and replacing it with their own (which would then hold the attacker's public key) without you noticing and calling back your coworker.
I tend to like high tech solutions like verisign, but if I have to pay for a key, I'll look more closely at the free solutions first.
Vote with your money folks. If they don't support the OS you want to run officially, they don't deserve your money.
I personally am going to stay with a 56Kbps modem until ADSL comes to town (later this year supposedly), for exactly this reason. I won't do cable modem, because my local cable co. wants to pretend linux doesn't exist. They aren't working hard enough to get my support.
If those screenshots are all running in an X server, that means someone somehow got all those little microsoft icons into their X server. I'm not saying that's impossible (it's easy), but why bother making this look like something it isn't?
That is, my guess is you don't need an X server for this.
Not sure this is what you're looking for, but I have
a bunch of intro-unix stuff (which includes linux material) at http://nis.acs.uci.edu/~strombrg/intro.html .
It'd be nice to be able to overlay the events of one calendar user over those of another. Not just copy: overlay.
Then you can have a virtual user who takes care of a particular project, and you can hand it to whatever real user needs it, much like having a mailing address alias for the person who's doing the contract changes this month.
There's a new trend in the free software world, and it isn't a healthy one.
Suddenly, free software people seem to think they're Doing the Right Thing, when they complain about businesses making money using free software. Redhat, Netscape, and I'm betting there are others, are examples.
Folks, if you don't want someone else to make money from your hard work, don't give away your code. For Pete's sake, there's nothing in a typical free software license that prevents it, so why publish under a free license then?
Me, on the other hand, I want anyone, individuals, corporations, whatever, to be able to use my code any way they want, as long as they don't try to prevent others from using it, or embrace-extend-extinguish it. So for me, the GPL works out great.
So in short, if you want to contribute to free software, do. If you don't want to contribute to free software, don't. Don't confuse the two.
I taught a chapter out of the 2nd edition, and thumbed through the rest of the book.
It was so rife with factual errors about Solaris that I had to recommend against reading the book.
So is the 3rd edition still this bad?
I've been told Casper Dik (Solaris deity) met Evi Nemeth (coauthor of this book) once. I was told that approx every other sentence out of her mouth was about how much she hated Solaris.
Granted, Sun and Solaris aren't perfect, but that's no excuse to make up things about how bad they are. That's the sort of thing that creates backlash.
This is of course payback for all the times a unix person told a mac person that MacOS wasn't a real OS for various reasons.
On an almost related note, would you believe one of the best Mac people I know considered "protected memory" of mere "buzzword" value?
I used to hang a lot of hopes on the Hurd. I even ran NetBSD and FreeBSD for a while just because they were the OSes of choice for bootstrapping the Hurd.
But the Hurd just doesn't have nearly the critical mass Linux does, and likely never will. It's still enshrowded, to some extent, in the early days of Hurd's totally closed development model, for one thing.
Like it or not, Linux is the premier free kernel, particularly in terms of mindshare, and mindshare means a LOT to a company like IBM.
A linux emulation layer sounds great, but it's a little risky; you never know when something will be added to the real linux, that would be a total PITA to emulate with your sorta-kinda-almost-similar kernel. If I were running a big business, I'd seriously frown on that kind of unnecessary risk.
I say if Linus doesn't want big iron patches in the mainline kernel, that's a shame, and oh well, let's fork it.
I'm guessing it really hurts a company like IBM to commit to linux like it has, and then be told, "sorry, you're s second class citizen in the linux world". I'd think twice being saying that to IBM, if we want to continue getting big-company support.
And yes, big-company support is very valuable for linux.
First: I like redhat. That could change, but right now I like them. They've been really great about subsidizing linux development, they've been choosing the right directions for development (gtk, gnome, python) and the distribution's not at all bad. (Maybe they release too often for the sysadmin in me, but that's another issue entirely. The developer in me likes the frequent releases)
Using gcc "2.96" was questionable, as was their choice of c library version. I can see how internationalization might have been important, so there wasn't a 100% good-for-everybody choice.
To me, the big question is, for example, "What are they going to do when gcc 3.0 is released"?
1) Will they pull a sun (not a terrible thing), and just expect that everyone will install 3.0 themself if they want 3.0?
2) Will they issue a patch that updates the compiler to 3.0, and hope it doesn't cause tons of problems that c++ libraries have to be recompiled? Will they then also patch all the libraries that come with redhat that were compiled from c++?
3) Will they release a "redhat 8" in a short timeframe to get back to binary compatibility with what the gcc maintainers intended?
4) Or will they try to pressure the gcc maintainers into preserving binary compatibility with "2.96"?
Basically the same questions are there for glibc too.
I used to put off redhat upgrades for quite a
while.
I had a lot of good reasons for it.
One of the reasons I've discovered not to put
off the upgrades: a lot of good software gets
released that just won't compile without the
newer build environments.
I stampeded into redhat 7. As soon as it was released, I ordered a CD. The day after
it arrived in the mail (I needed the system to be usable that day), I upgraded.
I haven't regretted it yet. While some binaries
are segfaulting, some sources that wouldn't compile on 6.2, do compile on 7.
Slashdot Mirror
No, I'm not saying KDE is bad because
I'm saying that when the FSF was confronted with KDE licensing trouble, they launched two projects - one intended to embrace and make liveable (the free qt project, which died I believe), and one intended to entirely replace (gnome).
Such a dual approach might make sense with
BTW, what about the PKI for OpenPGP? Could it be augmented to do this somehow?
Ok, I really like mozilla 0.9.1. mozilla has been my main browser for quite a while, and recently it became able to talk to a slightly unusual https site I need access to, so I almost never run netscape 4.x anymore.
However, that damn installer is keeping me from downloading mozilla 0.9.2, much as it used to keep me from downloading some of the netscape 6.x releases.
Is there some trick to getting it not to segfault? It seems phase of the moon for me.
Is there some way to download mozilla without the silly installer?
I have a palm IIIxe and an Agenda VR3. I actually use the IIIxe and not the VR3, but if I write any software, it'll be for a linux-based pda. I may end up getting a yopy though, as I prefer GTK+ (yopy) to FLTK (vr3).
BTW, the VR3 costs not that much more than a IIIxe these days, I believe. If you don't mind fltk, and plan to do some programming, it'd be a good way to go.
If you need something today, I'd take a serious look at the Agenda VR3. It's gotten twice as fast recently, since the EIP feature, and there was some talk of optimizing gcc for the machine better too at one time... It appears to have a pretty active developer community. I'm not that enthused with the FLTK gui toolkit it's using, but it could be worse.
I'm hanging some high hopes on Yopy though. GMate isn't funding Microsoft or contributing to Microsoft's marketshare numbers on every purchase, and Yopy appears to be targetting just the right technologies now: X, GTK+ and python.
I'd never consider teaching students C++ as a first language. C either.
Python yes. Java yes. Perl no. ML would be reasonable, as would Eiffel.
C and C++ need to be taught, but they should be in a Junior level course.
If I were in a "yeah, Java's not the best, but it does have a good combination of reasonable power and critical mass" mood, I'd teach java first. If I didn't care about who's using what at that moment, I'd teach python first.
Actually, such a protocol would need an option like:
Complete transmission only if all hops are encrypted? y/n
...because some mail has to get through, but it'd be nice if it were encrypted, while other mail shouldn't be transmitted the rest of the way if it can't go the whole way encrypted.
Er, actually, I avoid learning about Microsoft software about as much as I can, but I believe that in 95 (not sure about 98) and I believe even some early SP's of NT, passwords aren't sent cleartext, but replayable hashes are sent in cleartext, which is about as bad.
So you can't just log in by typing someone's password after stealing it, by you can write a program that'll send the replayable hash at the right time when asked.
Ok, I'm expecting some rabid disagreement on this, but the time period of my life when I had the least back trouble, was when I was doing deadlifts regularly. Since quitting, I think I have back trouble less often than I had it before my stint of deadlifting, but I am getting it a little bit again.
A lot of folks say you should never deadlift. I tried it for a while, and got to really like it. It thickened my "love handles", which isn't the most aesthetic thing in the world, but folks, that muscle is stabilizing, and it taught me proper form for picking up heavy objects.
It's good your soliciting ideas. You'll get a mess of conflicting opinions, but as long as you do what you think is best in the end (giving a coherent whole), taking suggestions can only make your language better.
I want to comment on one of the least important bits semantically, that has a big payoff: avoid those damn space-wasting, error inducing curly braces or begin-end pairs. If you do it where all if's require and end, ala modula-2, that'd be great. Or even better: do it in the python style, where the nesting is taken directly from the indention, eliminating a really pesky class of errors as well as all the wasted vertical space (or alternatively, eliminating those obnoxious curly braces that don't line up).
More importantly, study what's already out there. Don't do a language because you want to do a language, do a language because you think you have just the right combination of features taken from other languages, and/or have a novel idea that you feel the world would benefit from.
Oh, and -don't- go the perl route, for Pete's sake! What a mess, and getting worse. An ideal language is simple enough for an idiot, but still exceptionally powerful in the hands of an expert - complexity should come from algorithms, not a morass of alternative ways of doing things In The Language. Operating overloading is about as bizarre as the language should allow.
Also, if you do have a novel idea, make sure it wouldn't be better added to an existing language that already has some critical mass than to a new language. Some ideas could simply be implemented as a class library, while others could be added to a language proper by an open-minded project coordinator. I'd guess python or ML (Meta language, _not_ machine language) folks (two languages I respect quite a bit) would be very open to hearing of a novel idea, and I'm pretty sure there are others too.
I work in a centralized support organization at a university.
Some of our clients are happy to upgrade whenever. Ok, that's stretching the truth, but it isn't that hard to convince some folks.
I'm not writing about them. I'm writing about the people who hate upgrading.
Some of our clients just refuse to upgrade. "We must have a stable system".
This wouldn't be so bad if it weren't for the facts that:
1) Redhat releases often - about every 6 months it seems.
2) Redhat stops releasing patches for an OS release fairly early.
This combined means that anything but a x.y release, where y is the last minor revision of the x major revision of a redhat os, doesn't work so well for our clients (because redhat continues to release patches for a decent amount of time for the last y of an x, but not for earlier y's).
So for a while, we decided we were only going to support the last y of an x. However, naturally, clients started saying that was just too old. "Why do we want to install something 2 releases back? What good does that do?" We can explain all we want, clients just don't understand.
So I'm asking, can redhat consider either releasing less often, or continuing to support releases with patches longer? If neither of these work for redhat, do you have any suggestions for us that might help our situation?
I should add that upgrade procedures don't work well with the heavy sysadmin automation we use here. We always reinstall from scratch, and our many scripts (in one unified framework) put back the tweaks.
It seems to me the whole idea of a rebate is a scam.
What good are they?
The only thing I can think of is that some people will think "Oh good, $1000 with $200 rebate, that's really $800", but when push comes to shove, they neglect to do the _extra_work_ to get the rebate.
Sure, most people will probably do the rebate, but if 8% don't, then that's a lot like the company got to jack their price, without nearly the same reduction in sales.
Now, if they do rebates, and go one step further to make doing rebates an even bigger PITA... Perhaps that's even better for the company.
That is, until people start to wise up and avoid the extra unnecessary work associated with rebates. Personally, if I see a car after rebate for $n, and a car without rebate for $n+100, I'm probably going to do the car without the stupid rebate. What a lot of arbitrary busy-work. If they really want my dollars, they should include the price reduction in the sales price. I mean, really, why wouldn't they?
It is an irony of the computing industry, that when someone sees multiple factions in software for providing a function, that people decide to solve the problem by introducing another faction.
Consider the tar and cpio factions, and "pax", the faction to end all (archiver) factions.
So what's wrong with calling your coworker with a message (EG "I like green tea") and encrypting that message with your private key (using gpg or pgp), and putting your public key up on the web (giving the URL to the public key over the phone) or putting your public key on the keyservers?
Then if your overseas coworker can decrypt the message and get the correct text they learned of on the phone... it would seem this would go a good ways toward establishing trust.
Sure, someone could snoop the overseas call and send the same message encrypted with their own private key, but they'd have a hard time taking down your website (which has the public key) and replacing it with their own (which would then hold the attacker's public key) without you noticing and calling back your coworker.
I tend to like high tech solutions like verisign, but if I have to pay for a key, I'll look more closely at the free solutions first.
OMIGOSH.
That explains so much.
No wonder *ix is beneath contempt to microsoft. If they can't play, no one should.
Vote with your money folks. If they don't support the OS you want to run officially, they don't deserve your money.
I personally am going to stay with a 56Kbps modem until ADSL comes to town (later this year supposedly), for exactly this reason. I won't do cable modem, because my local cable co. wants to pretend linux doesn't exist. They aren't working hard enough to get my support.
Er, wait.
Yes, this isn't a native port.
But does it really require an X server?
Did you look at the screenshots?
If those screenshots are all running in an X server, that means someone somehow got all those little microsoft icons into their X server. I'm not saying that's impossible (it's easy), but why bother making this look like something it isn't?
That is, my guess is you don't need an X server for this.
Not sure this is what you're looking for, but I have
a bunch of intro-unix stuff (which includes linux material) at http://nis.acs.uci.edu/~strombrg/intro.html .
Don't forget
It'd be nice to be able to overlay the events of one calendar user over those of another. Not just copy: overlay.
Then you can have a virtual user who takes care of a particular project, and you can hand it to whatever real user needs it, much like having a mailing address alias for the person who's doing the contract changes this month.
There's a new trend in the free software world, and it isn't a healthy one.
Suddenly, free software people seem to think they're Doing the Right Thing, when they complain about businesses making money using free software. Redhat, Netscape, and I'm betting there are others, are examples.
Folks, if you don't want someone else to make money from your hard work, don't give away your code. For Pete's sake, there's nothing in a typical free software license that prevents it, so why publish under a free license then?
Me, on the other hand, I want anyone, individuals, corporations, whatever, to be able to use my code any way they want, as long as they don't try to prevent others from using it, or embrace-extend-extinguish it. So for me, the GPL works out great.
So in short, if you want to contribute to free software, do. If you don't want to contribute to free software, don't. Don't confuse the two.
seki-strombrg> /home/Amaya/LINUX-ELF/bin/amaya
/home/Amaya/LINUX-ELF/bin/amaya 2>&1 | tail
...}) = 0
/etc/redhat-release
*** Thot: Irrecoverable error ***
seki-strombrg> strace -f
fstat(4, {st_mode=S_IFREG|0644, st_size=67,
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x4001d000
read(4, "$&\'*-./0123456789abcdefghijklmno"..., 4096) = 67
--- SIGSEGV (Segmentation fault) ---
rt_sigaction(SIGBUS, {SIG_DFL}, {0x810847c, [], SA_RESTART|0x4000000}, 8) = 0
rt_sigaction(SIGPIPE, {SIG_IGN}, {SIG_DFL}, 8) = 0
rt_sigaction(SIGSEGV, {SIG_DFL}, {0x810847c, [], SA_RESTART|0x4000000}, 8) = 0
rt_sigaction(SIGABRT, {SIG_DFL}, {0x810847c, [], SA_RESTART|0x4000000}, 8) = 0
write(2, "*** Thot: Irrecoverable error **"..., 33*** Thot: Irrecoverable error ***) = 33
_exit(1) = ?
seki-strombrg> cat
Red Hat Linux release 6.2 (Zoot)
seki-strombrg> uname -m
i686
I taught a chapter out of the 2nd edition, and thumbed through the rest of the book.
It was so rife with factual errors about Solaris that I had to recommend against reading the book.
So is the 3rd edition still this bad?
I've been told Casper Dik (Solaris deity) met Evi Nemeth (coauthor of this book) once. I was told that approx every other sentence out of her mouth was about how much she hated Solaris.
Granted, Sun and Solaris aren't perfect, but that's no excuse to make up things about how bad they are. That's the sort of thing that creates backlash.
This is of course payback for all the times a unix person told a mac person that MacOS wasn't a real OS for various reasons. On an almost related note, would you believe one of the best Mac people I know considered "protected memory" of mere "buzzword" value?
I used to hang a lot of hopes on the Hurd. I even ran NetBSD and FreeBSD for a while just because they were the OSes of choice for bootstrapping the Hurd.
But the Hurd just doesn't have nearly the critical mass Linux does, and likely never will. It's still enshrowded, to some extent, in the early days of Hurd's totally closed development model, for one thing.
Like it or not, Linux is the premier free kernel, particularly in terms of mindshare, and mindshare means a LOT to a company like IBM.
A linux emulation layer sounds great, but it's a little risky; you never know when something will be added to the real linux, that would be a total PITA to emulate with your sorta-kinda-almost-similar kernel. If I were running a big business, I'd seriously frown on that kind of unnecessary risk.
I say if Linus doesn't want big iron patches in the mainline kernel, that's a shame, and oh well, let's fork it.
I'm guessing it really hurts a company like IBM to commit to linux like it has, and then be told, "sorry, you're s second class citizen in the linux world". I'd think twice being saying that to IBM, if we want to continue getting big-company support.
And yes, big-company support is very valuable for linux.
First: I like redhat. That could change, but right now I like them. They've been really great about subsidizing linux development, they've been choosing the right directions for development (gtk, gnome, python) and the distribution's not at all bad. (Maybe they release too often for the sysadmin in me, but that's another issue entirely. The developer in me likes the frequent releases)
Using gcc "2.96" was questionable, as was their choice of c library version. I can see how internationalization might have been important, so there wasn't a 100% good-for-everybody choice.
To me, the big question is, for example, "What are they going to do when gcc 3.0 is released"?
1) Will they pull a sun (not a terrible thing), and just expect that everyone will install 3.0 themself if they want 3.0?
2) Will they issue a patch that updates the compiler to 3.0, and hope it doesn't cause tons of problems that c++ libraries have to be recompiled? Will they then also patch all the libraries that come with redhat that were compiled from c++?
3) Will they release a "redhat 8" in a short timeframe to get back to binary compatibility with what the gcc maintainers intended?
4) Or will they try to pressure the gcc maintainers into preserving binary compatibility with "2.96"?
Basically the same questions are there for glibc too.
I used to put off redhat upgrades for quite a
while.
I had a lot of good reasons for it.
One of the reasons I've discovered not to put
off the upgrades: a lot of good software gets
released that just won't compile without the
newer build environments.
I stampeded into redhat 7. As soon as it was released, I ordered a CD. The day after
it arrived in the mail (I needed the system to be usable that day), I upgraded.
I haven't regretted it yet. While some binaries
are segfaulting, some sources that wouldn't compile on 6.2, do compile on 7.