You are not logged in.
Pages: 1
I've read about this today and I'm curious about my fellow archers opinion about it.
It seems that this move was made because nobody likes Ulrich Drepper very much, but I can't say much about it.
(lambda ())
Offline
Offline
The bugreports referenced in that blogpost are definitely an interesting read
Offline
The bugreports referenced in that blogpost are definitely an interesting read
/me agrees
Offline
New bug report today...
Warning: bad language:
http://sources.redhat.com/ml/glibc-bugs … 00028.html
Offline
I'm sure he's thick-skinned enough to take these insinuations as a joke, but I'm not so sure it's really appropriate for people to keep making references to Reiser's actions just because Drepper's a programmer who can be dickish at times.
...
Offline
I am currently playing with eglibc-2.9.90 (2.10 developement) on a chroot and VM.
The cool feature of eglibc is that can select group of options for build at runtime, so for example can remove support for ipv6, locales, etc. Of course this is great for embedded devices.
Offline
I'm sure he's thick-skinned enough to take these insinuations as a joke, but I'm not so sure it's really appropriate for people to keep making references to Reiser's actions just because Drepper's a programmer who can be dickish at times.
Heh, is true, I just had to let it off my head. I have seen worse opinions from other developers though.
Offline
New bug report today...
Warning: bad language:
http://sources.redhat.com/ml/glibc-bugs … 00028.html
The Summary and AssignedTo lines literally made me lol, even if the person who posted that is an immature idiot for spamming the bugtracker.
I thought eglibc was mainly for embedded devices. If Debian moves to that does it mean that non-embedded platforms will eventually suffer, even if only marginally?
Are any of the Arch devs considering a move to eglibc? I suppose that as long as it remains compatible with glibc it would be trivial to provide both packages with the correct provides and conflicts lines.
Last edited by Xyne (2009-05-08 00:30:57)
My Arch Linux Stuff • Forum Etiquette • Community Ethos - Arch is not for everyone
Offline
I'll stick with glibc for the time being. I am interested to see what actually happens because Debain wants to maintain ABI compatibility. So I see this becoming the equivalent of a big patchset for glibc rather than a big fork.
It is an interesting situation. I was not around (in the sense of a person who took real notice) during the gcc/egcs split (1997-1999) and the Xorg split (in 2004) was for a fairly clear cut reason (licensing issue).
Offline
Wow, that guy is a dick.
Offline
I'll stick with glibc for the time being. I am interested to see what actually happens because Debain wants to maintain ABI compatibility. So I see this becoming the equivalent of a big patchset for glibc rather than a big fork.
Yes, eglibc is not a fork of glibc. From what I can see, is a series of patches agains glibc that do some changes against the build system, and other fixes that are rejected in glibc.
For example a diffstat between glibc_2_9 and eglibc_2_9 is: "234 files changed, 12330 insertions(+), 7269 deletions(-)" (~ 1 MB) The big changes resides in the configure, changelog and other aditionals config files add to enable build components optionally.
I am quite curious, and now replace the package glibc-2.9 with one I maked with eglibc-2.9.90 [2.10] (trunk) in one VM. I am posting from here
Offline
New bug report today...
Warning: bad language:
http://sources.redhat.com/ml/glibc-bugs … 00028.html
Anyone else think WONTFIX would be a more accurate resolution than INVALID?
Note that I've not run into Drepper on the internets before this, only read one thing from him (http://lwn.net/Articles/250967/ , it's very good). So, I'm completely uneducated with regards to this situation. But I would say any situation where it's pretty much open-and-shut dude-is-an-***hole is at the very least a terrible image for the project, and discourages fresh blood in the development, and helpful bug reporters.
Anyways, points to whoever puts up an eglibc package in the AUR first.
Offline
I didn't know this guy before today .
But to be honest for someone maintaining a project for that long , he must be a nice guy If these bug reports are the best people can get to prove that he's an asshole .
English is not my native language .
Offline
Heh, is true, I just had to let it off my head. I have seen worse opinions from other developers though.
I know what you mean; I saw some of those comments to which you're referring just prior to my initial post. Your's in itself was rather benign, but I figured speaking up early on would steer away those who would have crossed a line.
...
Offline
Is there anything technically wrong with glibc, or technically superior about eglibc? Changing the code base purely due to the developer being an asshole is stupid. Emotion based decisions rarely work out for the best.
Are you familiar with our Forum Rules, and How To Ask Questions The Smart Way?
BlueHackers // fscanary // resticctl
Offline
Actually, it's not really a *so* big change. eglibc is basically glic plus patches and want to maintain binary compatibility. In fact, Debian was already using not least than 200 patches over glibc and Drepper clearly ( you can't say less ) stated that he doesn't want them in and not always for good reasons. If you are not convinced, just look at what he said on patches for the ARM architecture. It's childish at best.
So, I imagine that Debian developers want to move some code from repositories to mainstream and that eglibc is a good way to do it without breaking everything.
"Je verrais le monde de bas en haut, c'est peut-être plus rigolo.
Je n'y perdrai rien par surcroît : il est pas drôle à l'endroit."
Offline
Is there anything technically wrong with glibc, or technically superior about eglibc? Changing the code base purely due to the developer being an asshole is stupid.
I think the more serious people are arguing that he refuses good patches. Most of the noise right now though is coming from the melodrama nerdscape and that's being fanned by biased reporting of the whole thing (e.g. quotes deliberately taken out of context... not that the original context was great, but definitely not as bad as without it).
Emotion based decisions rarely work out for the best.
Hence the divorce rate.
My Arch Linux Stuff • Forum Etiquette • Community Ethos - Arch is not for everyone
Offline
fukawi wrote:Emotion based decisions rarely work out for the best.
Hence the divorce rate.
BZ
Last edited by fukawi2 (2009-05-09 11:29:37)
Are you familiar with our Forum Rules, and How To Ask Questions The Smart Way?
BlueHackers // fscanary // resticctl
Offline
Anyways, points to whoever puts up an eglibc package in the AUR first.
For *testing purposes only* http://aur.archlinux.org/packages.php?ID=26373
Offline
The idea of switching software based on anything but the merits of the technology alone is ridiculous. I hope no other distro gives this kind of action consideration.
Yes, he can be an ass, but he's also a genius.. hence why he is the glibc maintainer and paid to be so.
Favorite Software:
Openbox3 - Mutt - Vim - GNU Screen
Offline
The idea of switching software based on anything but the merits of the technology alone is ridiculous. I hope no other distro gives this kind of action consideration.
Yes, he can be an ass, but he's also a genius.. hence why he is the glibc maintainer and paid to be so.
It is not completely ridiculous. Sometimes you want/need to interact with the programmer: offer some patches, ask for some support, report bugs. Also he/she might impose some ridiculous license which might prevent you from using "technologically perfect" software (like ion3). This is especially true for distro maintainers - if the programmer refuses to fix something that causes problems, what options as a distro maintainer do you have?
Offline
jlimon wrote:The idea of switching software based on anything but the merits of the technology alone is ridiculous. I hope no other distro gives this kind of action consideration.
Yes, he can be an ass, but he's also a genius.. hence why he is the glibc maintainer and paid to be so.
It is not completely ridiculous. Sometimes you want/need to interact with the programmer: offer some patches, ask for some support, report bugs. Also he/she might impose some ridiculous license which might prevent you from using "technologically perfect" software (like ion3). This is especially true for distro maintainers - if the programmer refuses to fix something that causes problems, what options as a distro maintainer do you have?
I agree with this to a point.. if there was a serious matter in which glibc MUST be changed to accommodate a release and he was unwilling to make the changes, but still, I fail to see that really happening..
Plus, almost every distro patches glibc anyways. I fail to see how this could be a real show stopper.
Favorite Software:
Openbox3 - Mutt - Vim - GNU Screen
Offline
Pages: 1