Re: [Cooker] The gcc-2.96 stupidity!
Graham Percival [EMAIL PROTECTED] writes: [...] Mandrake doesn't use *any* official releases. $ uname -r 2.2.17-21mdksecure $ rpm -qa | grep glibc glibc-2.1.3-16mdk glibc-devel-2.1.3-16mdk compat-glibc-5.3-2.0.7.9mdk glibc-profile-2.1.3-16mdk See those "-21mdksecure", "-16mdk", and the like? It means that the Official Release (tm) has been modified by Mandrake. Life goes on. ? You should do a little bit of packaging before saying such beepbeep. -- Guillaume Cottenceau -- Distribution Developer for MandrakeSoft http://us.mandrakesoft.com/~gc/
Re: [Cooker] The gcc-2.96 stupidity!
On Fri, Oct 13, 2000 at 02:37:10PM +0200, Guillaume Cottenceau wrote: ? You should do a little bit of packaging before saying such beepbeep. But he's right! Aren't there patches made by Mandrake in these packages? So it is not the original source anymore. Alexander Skwar -- Homepage: http://www.digitalprojects.com | http://www.dp.ath.cx Sichere Mail? Mail an [EMAIL PROTECTED] fuer GnuPG Keys ICQ:7328191
Re: [Cooker] The gcc-2.96 stupidity!
Alexander Skwar [EMAIL PROTECTED] writes: On Fri, Oct 13, 2000 at 02:37:10PM +0200, Guillaume Cottenceau wrote: ? You should do a little bit of packaging before saying such beepbeep. But he's right! Aren't there patches made by Mandrake in these packages? So it is not the original source anymore. Two points: 1. 1mdk (or 2mdk or anything) DOESN'T mean that there are patches. This depends on the brokenness of the code (roughly), not on the number of the release. 2. Do you know any distribution where they provide original source unpatched? There are simply not, the aim of a distribution is to provide some fixed (e.g. working) versions of the sources!!! -- Guillaume Cottenceau -- Distribution Developer for MandrakeSoft http://us.mandrakesoft.com/~gc/
Re: [Cooker] The gcc-2.96 stupidity!
On 13 Oct 2000, Guillaume Cottenceau wrote: 2. Do you know any distribution where they provide original source unpatched? There are simply not, the aim of a distribution is to provide some fixed (e.g. working) versions of the sources!!! Probably Slackware which is why I don't like that distribution much :) -- Sincerely, David Walluck [EMAIL PROTECTED]
Re: [Cooker] The gcc-2.96 stupidity!
Millions of electrons died to bring me this message. Was it worth it, Guillaume Cottenceau? Alexander Skwar [EMAIL PROTECTED] writes: On Fri, Oct 13, 2000 at 02:37:10PM +0200, Guillaume Cottenceau wrote: ? You should do a little bit of packaging before saying such beepbeep. Sorry, I think there was a misunderstanding... 2. Do you know any distribution where they provide original source unpatched? There are simply not, the aim of a distribution is to provide some fixed (e.g. working) versions of the sources!!! Absolutely. I approve of this policy, and I was attempting to point out a flaw in the anti-2.96 movement. They claim that 2.96 is bad because it isn't an official release, but they don't have a problem with non-official kernel releases and the like. There's nothing wrong with not using an official release -- in some ways, that's what open source is all about! -- Graham Percival
Re: [Cooker] The gcc-2.96 stupidity!
On Fri, Oct 13, 2000 at 04:13:49PM +0200, Guillaume Cottenceau wrote: Alexander Skwar [EMAIL PROTECTED] writes: On Fri, Oct 13, 2000 at 02:37:10PM +0200, Guillaume Cottenceau wrote: ? You should do a little bit of packaging before saying such beepbeep. But he's right! Aren't there patches made by Mandrake in these packages? So it is not the original source anymore. Two points: 1. 1mdk (or 2mdk or anything) DOESN'T mean that there are patches. This depends on the brokenness of the code (roughly), not on the number of the release. Well, sure, 1mdk doesn't mean this, but 17mdk quite often means, or have a look at the kernel. The kernel is far more enhanced than the plain "linus" kernel is. But that's fine, because: 2. Do you know any distribution where they provide original source unpatched? There are simply not, the aim of a distribution is to provide some fixed (e.g. working) versions of the sources!!! Of course. I didn't say that it was bad or anything. I plainly said (or thought I did:]) that there are patches in the packages and that it is no longer the "original" package. But I don't have a problem with that. As long as it runs, and much more so when it runs even better than the plain source, I'm totally happy. Alexander Skwar -- Homepage: http://www.digitalprojects.com | http://www.dp.ath.cx Sichere Mail? Mail an [EMAIL PROTECTED] fuer GnuPG Keys ICQ:7328191
Re: [Cooker] The gcc-2.96 stupidity!
Mattias Eriksson [EMAIL PROTECTED] writes: Ok, I have asked several times about when they plan to release gcc-3.0 and haven't got one single reply that might even look like an answer to that. From that I assume that people simply dont know. Still, they are claiming that we must get prepaired for it! as for _every_ software, nobody know the release date but everybody kwow that it'll be released when it'll be ready. The fact that they don't have a clue about when it's comming makes this argument just plain BULL SHIT! This means that the real reason to switch to gcc 3.0 is that "big brother" RedHat did so, and Mandrake doesn't dare to do anything different from "big brother". AFAIC, the main argument to switch is the better C++ support. And don't come cry 'they broke c++ ABI'. in past, for each new libstdc++ release, we've had to recompile all C++ rpms else they broke after a few libstdc++ releases. Yes we don't care of being diffenret from big brother: coherent menu tree, same menu tree among all wm, alternatives, rpmdrake, drakxtools, MandrakeUpdate, DrakX, DiskDrake, Linux4zindoz, harddrake, msec, ... And as for us, we _never_ release a distro with gcc-2.96. I think that as long as the gcc developers are against using gcc 2.96 I dont think Mandrake should. If the gcc developers claim that you dont gain anything from using gcc and that they dont recomend it. I belive them, they should know. the fact is that we don't just use a CVS snapshot of gcc but a snapshot + 90 patches which fix nearly all known bugs. please, stop flaming us but for providing a piece of code that doesn't compile with cooker'gcc but obviously should. Then we'll fix gcc.
Re: [Cooker] The gcc-2.96 stupidity!
At 12 October, 2000 Thierry Vignaud wrote: And as for us, we _never_ release a distro with gcc-2.96. To quote Guillaume Cottenceau: "if unfortunately gcc3 will not be out for our next release we will have to find a solution, and it will not be to revert to 2.95 for sure." Have you people at mandrakesoft even discussed this matter? As I always have said, if you are just going to use it for cooker, and _never_ do a release using this snapshot (you can call it what ever, but it's not a official gcc release), I have no problems at all. I still dont see why you have to rush into things without even knowing if it's needed. I dont think the right way to prepair for a kernel is to use the hackerkernel as default for a unknown period of time before the expected stable release, the same apply on compilers! //Snaggen -- Mattias Eriksson E-mail: [EMAIL PROTECTED] Tvistevägen 26 [EMAIL PROTECTED] 907 36 UMEA Tel:090-198800 SWEDEN 070-5636769 'I don't fight for a cause Hemsida: http://www.acc.umu.se/~snaggen I fight for the fight' PGP: http://www.acc.umu.se/~snaggen/snaggen.asc
Re: [Cooker] The gcc-2.96 stupidity!
Mattias Eriksson [EMAIL PROTECTED] writes: At 12 October, 2000 Thierry Vignaud wrote: And as for us, we _never_ release a distro with gcc-2.96. To quote Guillaume Cottenceau: "if unfortunately gcc3 will not be out for our next release we will have to find a solution, and it will not be to revert to 2.95 for sure." Have you people at mandrakesoft even discussed this matter? not really, and as for me i don't agree with gc on this but changing back to gcc-2.95 will not be that easy, that's the pb. For example 2.95 has pbs with "-O3 -mpentiumpro" on C++, whereas 2.95 may work...
Re: [Cooker] The gcc-2.96 stupidity!
At 12 October, 2000 Pixel wrote: Mattias Eriksson [EMAIL PROTECTED] writes: At 12 October, 2000 Thierry Vignaud wrote: And as for us, we _never_ release a distro with gcc-2.96. To quote Guillaume Cottenceau: "if unfortunately gcc3 will not be out for our next release we will have to find a solution, and it will not be to revert to 2.95 for sure." Have you people at mandrakesoft even discussed this matter? not really, and as for me i don't agree with gc on this but changing back to gcc-2.95 will not be that easy, that's the pb. For example 2.95 has pbs with "-O3 -mpentiumpro" on C++, whereas 2.95 may work... I would like to see a official statement from mandrakesoft that stated that we are _never_ going to make a release with an unstable compiler! But this would require that this matter is discussed by the people at mandrakesoft. Until then I suggest to remove gcc-2.96 from cooker! //Snaggen -- Mattias Eriksson E-mail: [EMAIL PROTECTED] Tvistevägen 26 [EMAIL PROTECTED] 907 36 UMEA Tel:090-198800 SWEDEN 070-5636769 'I don't fight for a cause Hemsida: http://www.acc.umu.se/~snaggen I fight for the fight' PGP: http://www.acc.umu.se/~snaggen/snaggen.asc
Re: [Cooker] The gcc-2.96 stupidity!
Pixel [EMAIL PROTECTED] writes: Mattias Eriksson [EMAIL PROTECTED] writes: At 12 October, 2000 Thierry Vignaud wrote: And as for us, we _never_ release a distro with gcc-2.96. To quote Guillaume Cottenceau: "if unfortunately gcc3 will not be out for our next release we will have to find a solution, and it will not be to revert to 2.95 for sure." Have you people at mandrakesoft even discussed this matter? not really, and as for me i don't agree with gc on this I was explaining what Chmouel told: -=-=-=-=- It mean we going to take a highly patched/fixed/modified version of gcc2.96 if gcc3.0 is not ready, we going to usethe save version that the leader has, period. [...] What he says is that if unfortunately gcc3 will not be out for our next release we will have to find a solution, and it will not be to revert to 2.95 for sure. -=-=-=-=- And I can hardly imagine us working on fixing compiles with gcc-2.96 during 3 months, then revert back to gcc-2.95 :-((. -- Guillaume Cottenceau -- Distribution Developer for MandrakeSoft http://us.mandrakesoft.com/~gc/
Re: [Cooker] The gcc-2.96 stupidity!
Mattias Eriksson [EMAIL PROTECTED] writes: At 12 October, 2000 Thierry Vignaud wrote: And as for us, we _never_ release a distro with gcc-2.96. To quote Guillaume Cottenceau: "if unfortunately gcc3 will not be out for our next release we will have to find a solution, and it will not be to revert to 2.95 for sure." and to quote me: 'we've never released a distro with gcc-2.96' :-) (see above ??) past!='possible future', ok? Have you people at mandrakesoft even discussed this matter? As I always have said, if you are just going to use it for cooker, and _never_ do a release using this snapshot (you can call it what ever, but it's not a official gcc release), I have no problems at all. I still dont see why you have to rush into things without even knowing if it's needed. I dont think the right way to prepair for a kernel is to use the hackerkernel as default for a unknown period of time before the expected stable release, the same apply on compilers! then 2.4.x drivers won't be tested, then new gcc generated code'll never be tested. Is this you want from us ? whatever gcc-3 is availlable or not for 7.3/8, we'll provide a tested and fixed gcc. in fact, gcc-2.95.3 may have more bugs that the new one ...
Re: [Cooker] The gcc-2.96 stupidity!
On Thu, 12 Oct 2000, you wrote: Mattias Eriksson [EMAIL PROTECTED] writes: At 12 October, 2000 Thierry Vignaud wrote: And as for us, we _never_ release a distro with gcc-2.96. To quote Guillaume Cottenceau: "if unfortunately gcc3 will not be out for our next release we will have to find a solution, and it will not be to revert to 2.95 for sure." and to quote me: 'we've never released a distro with gcc-2.96' :-) (see above ??) past!='possible future', ok? Have you people at mandrakesoft even discussed this matter? As I always have said, if you are just going to use it for cooker, and _never_ do a release using this snapshot (you can call it what ever, but it's not a official gcc release), I have no problems at all. I still dont see why you have to rush into things without even knowing if it's needed. I dont think the right way to prepair for a kernel is to use the hackerkernel as default for a unknown period of time before the expected stable release, the same apply on compilers! then 2.4.x drivers won't be tested, then new gcc generated code'll never be tested. Is this you want from us ? whatever gcc-3 is availlable or not for 7.3/8, we'll provide a tested and fixed gcc. in fact, gcc-2.95.3 may have more bugs that the new one ... This is all I want to hear ;) I agree with the need to test for gcc-3, I just don't think 2.96 in it's current state should be on a release distro. People hold of on new versions when MS releases, while linux people look forward to the new releases because they are usually more robust while still as stable, I don't want to see than change and I'm sure you don't want to release a distro like redhat did. All in all I trust and like Mandrake enough to push it on all my friends and family, and I've tried quite a few distro's. Redhat, Suse, slack, debian, even non linux freebsd and netbsd (a while back), I chose to use Mandrake because it's a superior product for my purposes at home and work. Whatever happens with gcc, even it it's in 7.3/8 I will at least give it a try with Mandrake.
Re: [Cooker] The gcc-2.96 stupidity!
At 12 October, 2000 Thierry Vignaud wrote: Mattias Eriksson [EMAIL PROTECTED] writes: At 12 October, 2000 Thierry Vignaud wrote: And as for us, we _never_ release a distro with gcc-2.96. To quote Guillaume Cottenceau: "if unfortunately gcc3 will not be out for our next release we will have to find a solution, and it will not be to revert to 2.95 for sure." and to quote me: 'we've never released a distro with gcc-2.96' :-) (see above ??) past!='possible future', ok? Ehhh se above... OK "we _never_ release" != 'we've never released Better learn hot to quote... to rush into things without even knowing if it's needed. I dont think the right way to prepair for a kernel is to use the hackerkernel as default for a unknown period of time before the expected stable release, the same apply on compilers! then 2.4.x drivers won't be tested, then new gcc generated code'll never be tested. Is this you want from us ? whatever gcc-3 is availlable or not for 7.3/8, we'll provide a tested and fixed gcc. in fact, gcc-2.95.3 may have more bugs that the new one ... I said for a unknown period of time, when I know that the time to the expected release of a stable kernel is less than the time to the distributions release I think it's ok to use unstable stuff to make it stable. In that case you never risk to be in a situation where you have to release a unstable kernel. The same goes for compilers If I dont have a clue about when the expected releases will happen, the best choise is to wait and gather more info. //Snaggen -- Mattias Eriksson E-mail: [EMAIL PROTECTED] Tvistevägen 26 [EMAIL PROTECTED] 907 36 UMEA Tel:090-198800 SWEDEN 070-5636769 'I don't fight for a cause Hemsida: http://www.acc.umu.se/~snaggen I fight for the fight' PGP: http://www.acc.umu.se/~snaggen/snaggen.asc
Re: [Cooker] The gcc-2.96 stupidity!
Millions of electrons died to bring me this message. Was it worth it, Mattias Eriksson? As I always have said, if you are just going to use it for cooker, and _never_ do a release using this snapshot (you can call it what ever, but it's not a official gcc release), I have no problems at all. Mandrake doesn't use *any* official releases. $ uname -r 2.2.17-21mdksecure $ rpm -qa | grep glibc glibc-2.1.3-16mdk glibc-devel-2.1.3-16mdk compat-glibc-5.3-2.0.7.9mdk glibc-profile-2.1.3-16mdk See those "-21mdksecure", "-16mdk", and the like? It means that the Official Release (tm) has been modified by Mandrake. Life goes on. -- Graham Percival
