Re: Bug#616317: base: commit= ext3 mount option in fstab has no effect.
On Sun, May 08, 2011 at 05:24:09AM +0100, Ben Hutchings wrote: > On Sat, 2011-05-07 at 22:43 -0400, Ted Ts'o wrote: > [...] > > Should we try to make this work (at best badly) since a change in > > mount options in /etc/fstab would only take effect at the next > > mkinitramfs and/or update-grub invocation? Or should we just close > > out this bug and say, "tough luck, kid; if you want to change the root > > file system's mount options, you need to edit your kernel's boot > > options using whatever bootloader you might happen to be using"? > [...] > > Could we not have init remount root based on /etc/fstab? It already > handles remounting read-write. I suppose the problem then is that some > mount options can't practically be changed when remounting. (Worse, the > failure to change them is silent in some cases. And that is definitely > a bug.) See also: #520009 Also, with the version of initscripts in experimental, the domount mount helper can remount filesystems with the options from /etc/fstab. This is used to remount the filesystems mounted in the early initramfs with the options the sysadmin defined (if any). This could also be extended to do the same for the rootfs. The existing read_fstab could also be refactored to use the generic read_fstab_entry to pull out the entire set of mount options for remounting rather than just using the ro option. As both this and #520009 note, this won't work for all mount options such as data= for ext3, but it would allow any other options to from fstab to be applied to the root mount. Regards, Roger -- .''`. Roger Leigh : :' : Debian GNU/Linux http://people.debian.org/~rleigh/ `. `' Printing on GNU/Linux? http://gutenprint.sourceforge.net/ `-GPG Public Key: 0x25BFB848 Please GPG sign your mail. signature.asc Description: Digital signature
Re: Bug#616317: base: commit= ext3 mount option in fstab has no effect.
On Sat, 2011-05-07 at 22:43 -0400, Ted Ts'o wrote: [...] > Should we try to make this work (at best badly) since a change in > mount options in /etc/fstab would only take effect at the next > mkinitramfs and/or update-grub invocation? Or should we just close > out this bug and say, "tough luck, kid; if you want to change the root > file system's mount options, you need to edit your kernel's boot > options using whatever bootloader you might happen to be using"? [...] Could we not have init remount root based on /etc/fstab? It already handles remounting read-write. I suppose the problem then is that some mount options can't practically be changed when remounting. (Worse, the failure to change them is silent in some cases. And that is definitely a bug.) Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Re: Bug#616317: base: commit= ext3 mount option in fstab has no effect.
reassign 616317 base thanks This isn't a bug in e2fsprogs; e2fsprogs has absolutely nothing to do with mounting the file system. Debian simply doesn't support the mount options for the root file system in /etc/fstab having any effect on how the root file system is mounted. The root file system is mounted by the kernel, and the mount options used by the kernel are specified by the rootflags= option on the kernel's boot command line. This is effectively a feature request, and I debated what was the best way to deal with this bug. I could close it, and say, "not a bug", since Debian has never worked this way, and I suspect it was deliberate. Or, I could assign it to initramfs-tools, since what some other distributions do is look in /etc/fstab, parse out the mount options in for the root file system in /etc/fstab, and then insert into initrd image the appropriate root mount options. The problem with this is, (a) it's a bit of a hack, (b) it only takes effect the next time you install a new kernel, or if you deliberately and explicitly run mkinitramfs, which has fairly baroque options that most users would never figure out, and (c) not all Debian installations use an initrd, so whether or not it works would depend on how the boot sequence was set up. If you don't use an initrd, you'd have to edit it into the grub's configuration file. But then, not all Debian systems use grub as their boot loader. Neither these seemed obviously the right choice. So I'm going to do the cowardly thing, and choose the third option, which is to reassign this back to base, cc'ing debian-devel. I'm not sure what the right thing is to do here, since honoring this feature request would require making changes to multiple different packages: initramfs-tools, all of the bootloaders, etc. Should we try to make this work (at best badly) since a change in mount options in /etc/fstab would only take effect at the next mkinitramfs and/or update-grub invocation? Or should we just close out this bug and say, "tough luck, kid; if you want to change the root file system's mount options, you need to edit your kernel's boot options using whatever bootloader you might happen to be using"? I have a slight preference for the latter, since it's a lot less complexity that won't really work right anyway, but let's see what other people think. Regards, - Ted -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110508024329.ga15...@thunk.org
Re: Bits from the Release Team - Kicking off Wheezy
Le Sat, May 07, 2011 at 03:43:11PM +0200, Enrico Zini a écrit : > > a GR was needed to be able to proceed, because the hands of FD were > rather tied by this other GR: http://www.debian.org/vote/2008/vote_002 Hi Enrico, the 2008 GR invites to seek consensus, and in my opinion, what prove to be anti-consensual is to divide developers into formal categories. I have not seen such a vigourous opposition in the recent years to the idea of accepting non-packaging developers in Debian. Cheers, -- Charles Plessy Tsurumi, Kanagawa, Japan -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110508014635.gc12...@merveille.plessy.net
Re: Best practice for cleaning autotools-generated files?
On Sat, 07 May 2011, Tollef Fog Heen wrote: > | The problem is that autoreconf offers NO command line options for you to > | pass the required -I parameters for aclocal, nor is there a way to encode > | that information in the one place where it could conveniently live > | (configure.ac) AFAIK. > > Can't you use AC_CONFIG_MACRO_DIR? Note that this requires > ACLOCAL_AMFLAGS = -I m4 too in the top level Makefile.am if you're using > automake, though. Hey, thanks! Yes, it looks like it does exactly what I wanted for cyrus-imap. -- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110508002207.gb31...@khazad-dum.debian.net
Re: Best practice for cleaning autotools-generated files?
On Sat, 07 May 2011 at 13:33:53 +0200, Enrico Weigelt wrote: > c) does _NOT_ call configure As much as I wish this had been the convention, it isn't - the convention is that autogen.sh *does* call ./configure (often with options suitable for developers of the project, whereas the ./configure defaults are more suitable for packagers). Not doing so will just confuse people (and build systems like jhbuild, for that matter). It's mainly intended to be used to get started with development from a VCS checkout. Many upstreams do provide some way to not run ./configure (something like "./autogen.sh --no-configure" or "NOCONFIGURE=1 ./autogen.sh"). For many (most? all?) autoconf/automake projects, running "autoreconf" is enough; that's essentially what dh_autoreconf does. S -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507230300.ga7...@reptile.pseudorandom.co.uk
Bug#626017: ITP: ruby-shoulda-context -- context framework for Test::Unit
Package: wnpp Severity: wishlist Owner: Antonio Terceiro * Package name: ruby-shoulda-context Version : 1.0.0.beta1 Upstream Author : Several authors * URL : https://github.com/thoughtbot/shoulda-context * License : MIT Programming Lang: Ruby Description : context framework for Test::Unit From upstream README: Shoulda’s contexts make it easy to write understandable and maintainable tests for Test::Unit. It’s fully compatible with your existing tests in Test::Unit, and requires no retooling to use. -- Antonio Terceiro http://softwarelivre.org/terceiro signature.asc Description: Digital signature
Bug#626016: ITP: ruby-shoulda-matchers -- Test helpers for Rails applications, compatible with Test::Unit and RSpec
Package: wnpp Severity: wishlist Owner: Antonio Terceiro * Package name: ruby-shoulda-matchers Version : 1.0.0.beta2 Upstream Author : Several authors * URL : https://github.com/thoughtbot/shoulda-matchers * License : MIT Programming Lang: Ruby Description : Test helpers for Rails applications, compatible with Test::Unit and RSpec From upstream README: Test::Unit- and RSpec-compatible one-liners that test common Rails functionality. These tests would otherwise be much longer, more complex, and error-prone. -- Antonio Terceiro http://softwarelivre.org/terceiro signature.asc Description: Digital signature
Re: Best practice for cleaning autotools-generated files?
* Henrique de Moraes Holschuh schrieb: > > I'm (as upstream) using serval macros in their own .m4 files (eg. > > in ./m4/, maybe even sorted into subdirs). Can autoreconf figure > > out the required search pathes all on its own ? > > The problem is that autoreconf offers NO command line options for you to > pass the required -I parameters for aclocal, nor is there a way to encode > that information in the one place where it could conveniently live > (configure.ac) AFAIK. So, more precisely: autoreconf lacks an important feature would like to see. So, until this feature is available, we'll still need another way to cope with those situations. Sure, most of the packages probably could be changed to be fine with simple autoreconf call, but I've seen several cases where it's not that easy. In the end we have two options: a) maintain specific build rules for virtually any package in virtually any distro (yes, I'm not just looking at one single distro), indefinitely b) fix it in the source (the package itself) once and for all > I sure hope it will NEVER decide to actually search for .m4 files at > non-standard directories on its own, that would make things much worse. > > Anyway, you have to work around it using something like: > > ACLOCAL='aclocal -I foo -I bar' autoreconf Well, that would again be a package specific workaround which has to be maintained in the distro's build descriptors (whichever distro you're looking at). And if that changes, the distro's package maintainer has to find out early enough to (manually) adapt properly. Leaves too much manual works and chances of breaks for my taste. cu -- -- Enrico Weigelt, metux IT service -- http://www.metux.de/ phone: +49 36207 519931 email: weig...@metux.de mobile: +49 151 27565287 icq: 210169427 skype: nekrad666 -- Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme -- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507195421.GA25222@nibiru.local
Re: Best practice for cleaning autotools-generated files?
]] Henrique de Moraes Holschuh Hiya, | The problem is that autoreconf offers NO command line options for you to | pass the required -I parameters for aclocal, nor is there a way to encode | that information in the one place where it could conveniently live | (configure.ac) AFAIK. Can't you use AC_CONFIG_MACRO_DIR? Note that this requires ACLOCAL_AMFLAGS = -I m4 too in the top level Makefile.am if you're using automake, though. cheers, -- Tollef Fog Heen UNIX is user friendly, it's just picky about who its friends are -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/8762pmuwoy@qurzaw.varnish-software.com
Re: Best practice for cleaning autotools-generated files?
On Sat, 07 May 2011, Enrico Weigelt wrote: > * Henrique de Moraes Holschuh schrieb: > > Yes. I think it was Cyrus IMAP that required -I in places where > > autoreconf doesn't reach, so I called each tool separately. Which is > > obviously a problem in autoreconf. > > Is it really a problem of autoreconf ? Yes. > Imagine the following situation: > > I'm (as upstream) using serval macros in their own .m4 files (eg. > in ./m4/, maybe even sorted into subdirs). Can autoreconf figure > out the required search pathes all on its own ? The problem is that autoreconf offers NO command line options for you to pass the required -I parameters for aclocal, nor is there a way to encode that information in the one place where it could conveniently live (configure.ac) AFAIK. I sure hope it will NEVER decide to actually search for .m4 files at non-standard directories on its own, that would make things much worse. Anyway, you have to work around it using something like: ACLOCAL='aclocal -I foo -I bar' autoreconf (assuming that even works, I didn't test). I consider that way too icky. At that point, I just call each required tool directly in autogen.sh, instead of relying on autoreconf. -- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507191840.ga31...@khazad-dum.debian.net
Re: glibc: causes segfault in Xorg
On Sat, May 07, 2011 at 12:25:15PM +0200, Aurelien Jarno wrote: > On Wed, May 04, 2011 at 02:30:35PM +0200, Aurelien Jarno wrote: > > Le 04/05/2011 07:42, Steve M. Robbins a écrit : > > > P.S. I tried rebuilding glibc myself locally, but gcc also segfaults > > > in the process :-( > > > > Are you sure it is something related? Which gcc version are you using? > > Do you have a backtrace point to the same issue? I was careless in my initial report; I should have specified that I tried rebuilding the *old* glibc and got a segfault. At this point, all I really know is that building the eglibc 2.11.2-11 Debian source package on my up-to-date sid amd64 machine fails: make[3]: Entering directory `/home/steve/tmp/old-eglibc/eglibc-2.11.2/sunrpc' CPP='gcc-4.4 -E -x c-header' /home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/elf/ld-linux-x86-64.so.2 --library-path /home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc:/home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/math:/home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/elf:/home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/dlfcn:/home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/nss:/home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/nis:/home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/rt:/home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/resolv:/home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/crypt:/home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/nptl /home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/sunrpc/rpcgen -Y ../scripts -c rpcsvc/bootparam_prot.x -o /home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/sunrpc/xbootparam_prot.T make[3]: *** [/home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/sunrpc/xbootparam_prot.stmp] Segmentation fault (core dumped) The segfault is actually in the "ld-linux-x86-64.so.2" binary produced during the build, not gcc as I had earlier written. The backtrace is: (gdb) bt full #0 0x in ?? () No symbol table info available. #1 0x2b4d84d7e990 in call_init (l=, argc=7, argv=0x7fff26bf50a0, env=0x7fff26bf50e0) at dl-init.c:85 j = 1 jm = 4 init_array = 0x2b4d852ebb50 #2 0x2b4d84d7ea87 in _dl_init (main_map=0x2b4d84f90178, argc=7, argv=0x7fff26bf50a0, env=0x7fff26bf50e0) at dl-init.c:134 preinit_array = preinit_array_size = 0x0 i = 0 #3 0x2b4d84d71b2a in _dl_start_user () from /home/steve/tmp/old-eglibc/eglibc-2.11.2/build-tree/amd64-libc/elf/ld-linux-x86-64.so.2 No symbol table info available. #4 0x7fff26bf6574 in ?? () No symbol table info available. #5 0x0007 in ?? () No symbol table info available. #6 0x7fff26bf6825 in ?? () No symbol table info available. #7 0x7fff26bf6872 in ?? () No symbol table info available. #8 0x7fff26bf6875 in ?? () No symbol table info available. #9 0x7fff26bf6880 in ?? () No symbol table info available. #10 0x7fff26bf6883 in ?? () No symbol table info available. #11 0x7fff26bf689b in ?? () No symbol table info available. #12 0x7fff26bf689e in ?? () No symbol table info available. #13 0x in ?? () No symbol table info available. Hope this clarifies the issue somewhat. Thanks, -Steve signature.asc Description: Digital signature
Re: Bug#625865: ITP: ocportal -- ocPortal is a Content Management System for building and maintaining a dynamic website
On Fri, 6 May 2011, Chris Warburton wrote: On Fri, 2011-05-06 at 11:29 -0400, Scott Kitterman wrote: On Friday, May 06, 2011 11:23:50 AM Tshepang Lekhonkhobe wrote: On Fri, 2011-05-06 at 09:11 -0400, Scott Kitterman wrote: On Friday, May 06, 2011 08:56:21 AM Chris Warburton wrote: Programming Lang: PHP Description : ocPortal is a Content Management System for building and maintaining a dynamic website How many content management systems written in php does Debian need? It's not kool that you didn't even ask about how good it is. Maybe it's better than whatever exists in Debian currently, have you checked? My point is your question isn't helpful. It smacks of flaming. The question I should have asked is what is it's security record like. This is an area that's rife with applications that have 'poor' security records. Adding more to that pile would be an unfortunate burden on the security team. That's probably the most significant of the project wide costs adding a package like this brings with it. Scott K Hi Scott. ocPortal isn't massively widespread compared to other systems, so there's obviously less experimental proof of security. We had a security hole a few years ago; this was before I got involved, but there's details here http://en.wikipedia.org/wiki/OcPortal#Criticisms Hi Chris and the ITP and debian-devel, I think that if you are willing to work to make this a high-quality package, and be a responsive maintainer to bugs reported by users, I think it will be great to have you maintain it in Debian. The security work that you've described sounds great, and I hope that other PHP app upstreams hold their apps to such a high standard. If not, maybe you can use your tools to start filing bugs left and right against them. (-: For that reason, I will review your packaging when it's ready, and sponsor it into Debian if it passes muster. Keep me posted. -- -- Asheesh. http://asheesh.org/ Life is to you a dashing and bold adventure. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/alpine.deb.2.00.1105071405370.7...@rose.makesad.us
Re: Bits from the Release Team - Kicking off Wheezy
Enrico Zini wrote: > On Fri, May 06, 2011 at 02:04:20PM -0400, Michael Gilbert wrote: > > > It wasn't the GR itself. It was the fact that these changes to the NM > > process were actually made. I suppose it is arguable that those changes > > simply would not have happened without the GR, but that indicates more > > of a lack of direct motivation within the new maintainer team. > > > > So, if it required the GR to motivate them, then I suppose it was a > > necessity and ultimately a good thing, but my point is simply that its > > better when motivation comes from within; rather than an applied > > external force. > > I do not see how talking about the NM process or that GR is at all > relevant in this thread, and please do not consider this message of mine > an intent to contribute to it in any other way but to clarify a > misrepresentation of a team I'm a member of. > > From the point of view of Front Desk, motivation has always been there: > http://lists.debian.org/debian-devel-announce/2008/10/msg5.html > but a GR was needed to be able to proceed, because the hands of FD were > rather tied by this other GR: http://www.debian.org/vote/2008/vote_002 Then the core problem is the hand-tying itself, and that is the consequence of the GR process itself. Thus, my point remains: GRs are the wrong way to achieve change; they have long-term and unintended consequences. Ideally, changes should be adopted based on technical merit itself; rather than forced. > Please do not try to provide facts about the motivation or intentions of > others unless you really know them, otherwise you run into the risk of > misrepresenting other people, which is bad. OK, I actually tried to avoid presenting that as a fact, and more as an interpretation of the situation. "...if it required the GR to motivate them..." makes that intent pretty clear I think. I didn't intend the word "motivation" to be interpreted as lack of interest or in any kind of negative connotation, but more in the sense of overcoming some kind of barrier/inertia; i.e. definition 1 in a google search: "The reason or reasons one has for acting or behaving in a particular way". Best wishes, Mike -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507132223.d2deb096.michael.s.gilb...@gmail.com
Re: Integrating aptosid?
On Tue, May 03, 2011 at 03:05:34PM +0200, Pierre Habouzit wrote: > I know it's not simple, but it's not necessarily harder than making > testing usable, I think Joss made a pretty good case about that on his > blog. FWIW (#1), for the non-planet readers this is at http://np237.livejournal.com/31868.html > aptosid is just an example, I don't even know the distro, they may not > be the best choice, I just try to find alternates ideas. Note that I > don't think it takes more than 10 people to do the rolling distro like > Joss propose it: snapshot unstable every month and fix the worst > breakages. It takes more than 10 people to do the same in testing > because each individual fix is tangled in the migration issue, hence > rapidly needs to update things that are at first totally unrelated to be > fixed too first. FWIW (#2). Even though we seem to have consensus on going in a different direction for a Debian rolling suite, I've mimicked what Lucas did and reached out to a couple of aptosid people I've had a chance to meet at a past FOSS event in Dublin a few months ago and asked them to comment on this. You can find attached a mail from Niall Walsh, aptosid developer, which I've asked to comment on the issues of: 1) who they believe their target users are and 2) on the possibility of merging efforts with Debian. The mail is forwarded with permission from Niall, although it has been written initially for a single recipient rather than for a list. Take it with a grain of salt. Niall has also clarified that his views are only his and not necessarily representative of all aptosid developers, but I've found them interesting and useful for this debate nonetheless. Thanks a lot to Niall and to all the other developers for aptosid! Cheers. -- Stefano Zacchiroli -o- PhD in Computer Science \ PostDoc @ Univ. Paris 7 zack@{upsilon.cc,pps.jussieu.fr,debian.org} -<>- http://upsilon.cc/zack/ Quando anche i santi ti voltano le spalle, | . |. I've fans everywhere ti resta John Fante -- V. Capossela ...| ..: |.. -- C. Adams --- Begin Message --- Hi Stefano, Paul forwarded your e-mail onto me (not in a way I could really reply to inline, sigh). First off, obviously you now have my e-mail address. Secondly I permalurk on OFTC with the nick bfree, feel free to ping me there any time. You probably won't be surprised at all to know that I and all the other aptosid people are aware of the rambling (rolling) thread on debian-devel. Yesterday we even had lucas come join us in #aptosid where he spoke to a few of the aptosid team and had most of his questions answered as best they could be on the spot by Kel Modderman. You also probably won't be surprised to know that it's certainly not easy to even begin to try and answer the questions in your mail. I've thrown the mail in front of the other members of the aptosid team in case it helps prompt any of them to put on their asbestos underpants and join in. Both Kel and Stefan Lippers-Hollmann (CC'd here at his request as he hopes to send you an answer of his own) have shown their faces in the monster thread on debian-devel, so aptosid hasn't been totally silent on it. Some initial comments (I won't call them answers) to the questions posed in your mail: 1) I think many aptosid users are using it simply because they like to be closer to the "bleeding edge" then debian stable would allow without losing all the advantages of Debian. This mainly means simply getting newer versions of their applications. But as a group I guess I would say they must also be driven by enjoying staying closer to current software development, seeing things when they are newer and not waiting years before they get their hands on the features they hear about being released by the upstream projects. I guess I would charecterise the vast majority of aptosid users as people who have prior Linux experience on one distribution or another. Many seem drawn by the chance to use Debian's repository either because they know or have heard about the scope and quality of the software it contains. Obviously though they want (or at least think they want) newer things then Debian stable will give them. Quite a few will say they want a rolling release (and many mention having tried or considered Arch) though I've never tried to find out why so I can only answer that one for myself. I personally like a rolling release as a user because it means you don't get to face a day where you suddenly upgrade your entire system at once. Obviously the downside to this is you get many more opportunities for an upgrade to break something, but if it happens it's usually much easier to pick out the problem. I also like to feel that I am contributing to improving Debian even if I am only using unstable and do not contribute anything. The reality is though that the more people who use unstable, the more bugs will be found earlier and the better Debian will be for all, w
Re: New Packages generator
> As we already noted in our meeting minutes, there is one difference in > the files that you might notice: The order of the fields within one > entry is different. This should not create any trouble with compliant > 822 parsers, as order does not matter, but in case you do some weird > parsing on your own relying on any special order you might want to check > it still does what you expect. Just to clarify: The order of the fields is static, it doesn't change per call. It is just different to what we had. (And, you shouldnt rely on the order anyways :) ) -- bye, Joerg -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/871v0aedwr@gkar.ganneff.de
Re: Best practice for cleaning autotools-generated files?
On Sat, 7 May 2011 16:51:01 +0200 Enrico Weigelt wrote: > * Neil Williams schrieb: > > > Nonsense. It is not the job of ./autogen.sh to revert to the VCS state > > It is it's job to produce a clean state where *all* generated > files have been regenerated and the next stage (configure) > can start from here, w/o any manual intervention or workarounds. s/clean// ./autogen.sh does not clean anything. It may refresh stuff but it's not about cleaning things. It's not the job of ./autogen.sh to convert a semi-built broken setup back to the original VCS state. It simply regenerates the autotools stuff and provides the ./configure script. Any ./autogen.sh script which deletes .o files or messes with existing content in .libs/ directories is RC buggy IMHO. Fair enough if those are changed when the next 'make' is issued but that's the point, it's up to $(MAKE), not ./autogen.sh. > > and there's no harm in ./autogen.sh calling configure with whatever > > It *is*, as soon as this cannot run through without special > flags (eg. if some features have to be switched off, etc). The options are just passed on unchanged. No problem with that. Can still call ./configure directly. Maybe it wastes a little bit of time but if you're cross-building, you're using a really fast machine so this is hardly of concern. > > arguments are passed to ./autogen.sh - as long as those options have > > the same effect when passed to ./configure for subsequent runs during > > development. > > Adds additional complexity to add proper parameters here, for each > individual package. (and, of course, find out all this first). > > This way, you add unnecessary burden to all the maintainers of all > the distros out there (that might be interested in your package). and all the other build systems out there are suddenly compatible across all packages?? Dream on. > > There is no technical reason for any of these suggestions, > > Actually, there is. (okay, maybe more an organisatoric reason). That's not technical and there is no one organisation which bridges all of the various upstream teams. If you want to change things, you must persuade each team to adopt your preferences and you have singularly failed to convince me - because these ARE preferences, not best practice. > Exactly the same reason why things like AC plugs, voltages and > frequencies are standardized. Not true. Those things MUST work together in every permutation within a specific jurisdiction or people can die. Debian and autotools are nowhere near that level of importance. (My previous field {for >20yrs} was medical, so don't go assuming I'm unaware of the risks of non-standard electrical services. I've treated many patients for the kind of burns which result when things are even slightly outside the standard.) -- Neil Williams = http://www.linux.codehelp.co.uk/ pgpmACNhtDYCm.pgp Description: PGP signature
Re: Best practice for cleaning autotools-generated files?
* Neil Williams schrieb: > Nonsense. It is not the job of ./autogen.sh to revert to the VCS state It is it's job to produce a clean state where *all* generated files have been regenerated and the next stage (configure) can start from here, w/o any manual intervention or workarounds. > and there's no harm in ./autogen.sh calling configure with whatever It *is*, as soon as this cannot run through without special flags (eg. if some features have to be switched off, etc). > arguments are passed to ./autogen.sh - as long as those options have > the same effect when passed to ./configure for subsequent runs during > development. Adds additional complexity to add proper parameters here, for each individual package. (and, of course, find out all this first). This way, you add unnecessary burden to all the maintainers of all the distros out there (that might be interested in your package). > Again, this has nothing to do with Debian in most cases. True. Applies to virtually any distro. And they all have to cope with that crap, again, and again, and again. Total waste of human workforce, which would go away by simply fixing a few lines once. > There is no technical reason for any of these suggestions, Actually, there is. (okay, maybe more an organisatoric reason). Exactly the same reason why things like AC plugs, voltages and frequencies are standardized. cu -- -- Enrico Weigelt, metux IT service -- http://www.metux.de/ phone: +49 36207 519931 email: weig...@metux.de mobile: +49 151 27565287 icq: 210169427 skype: nekrad666 -- Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme -- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507145101.GG25423@nibiru.local
Re: Best practice for cleaning autotools-generated files?
On Sat, 7 May 2011 15:41:22 +0200 Enrico Weigelt wrote: > * Neil Williams schrieb: > > > No. Reverting the build to the point where it is equivalent to only > > what is in the VCS adds completely unnecessary build dependencies and > > build stages. > > Which ones exactly (besides the usual autotools) ? bison, flex and a few others. Also that certain stages of the transition from VCS checkout to tarball are not actually scripted anywhere, just described in a file. 'make distcheck' is an extremely useful sanity check with autotools packages. To do the check and then not distribute the results is just odd. To not do the check is negligent. > I don't want to repeat the last decades if SCM theory and common > practies here, but one of the major goals of an SCM infrastructure > is that you always get can grab the source of some product in some > specific version and run a fully rebuild, anytime. _Reliable_ > reproducability. Tarballs that are not fully-automatically produced > from the appropriate SCM tag (and can be reproduced anytime) simply > dont meet that requirement. Disagree. A tarball is easier to get hold of than some weird tag in an unfamiliar VCS. Common format, common methods to download and obtain it, simple URL's become possible instead of weird ones with ?=_, and other characters which rely on active parsing by a server instead of a simple directory listing. > Maybe that all worked for you over many years. Fine. It has indeed - about 10yrs at last count. > For me it always never worked, starting with common things like > sysroot'ed crosscompiling (yes, I've propbably got a different > scope than you, I'm coming from the embedded front) Umm, please check www.emdebian.org - I am very deeply involved in all of the embedded activity in Debian and elsewhere. It's been my geek obsession since before I was a DD and it's now my full time employment. Personally, I dislike sysroot because it doesn't actually make my cross-building work any easier. I work mainly on a dpkg-cross based system, I wrote or improved nearly all of the scripts which currently support cross-building in Debian and prepared the only completely cross-built Debian distribution for ARM and released it alongside Lenny. I did some early work on the cross-building toolchains before handling that task over to Hector. I think I qualify as "embedded". Don't make assumptions - my embedded work isn't exactly hard to find. > > I wouldn't want to work with any upstream which does not produce > > stable tarballs which are available for download without going > > anywhere near the VCS. I shouldn't even need to know what VCS is > > being used if all I want to do is build the package for my own use. > > Guess what, I wouldn't work with upstreams that don't provide > reliable release tags in the VCS (an VCS that I can import to git). > Actually, I sometimes do, but it adds a lot of extra burden. Fine. That means that any difference on that issue is personal preference and has no place in any "best practice" guide or specification. Each preference works for those who prefer that method. Either the guide / spec steers well clear of the entire area or it gives 100% equal coverage to all of the possible preferences. > My buildsystem (called Briegel), doesn't even bother with things > like tarballs or patches (still supported, but unused for years now). > It simply grabs the source tree from git by a normalized ref (eg. tag). > Normalized means: there's a simple rule to transform a vector of > package name and (also normalized) version number into a ref name. > Period. Then your build system cannot cope with any of my packages. The point with Emdebian is that it must cope with all packages which might be useful, across most of Debian. > Any changes needed in the source tree are done through SCM. > (the processes behind may vary between individual packages). No, those should be done via patches, managed by dpkg-source format 3 or directly incorporated into the upstream tarball (which is how I prefer to do it). Depending on the purpose of the patches, the patch can go upstream or into the Debian package. > > Overall, this has little to do with Debian. > > True. It's a distro-agnostic issue. It's about how to be a good > upstream to make distro's lifes easier and extends into the whole > QM topics. And that includes the ability to compare the checksums of tarballs included in the distributed sources of an embedded cross-built distribution to check that you really do have the same version as other distributions. It helps limit the amount of things which need checking when a bug only appears in one side. Normally, the cross-building process takes the most blame but there can be other issues. > > If those people are happy with their own system, then arguing > > about tags versus tarballs is just bikeshedding. > > This argument is only valid as long as you only consider official > upstream releases as your base. For example, I'
Re: Best practice for cleaning autotools-generated files?
* Sean Finney schrieb: IOW: all the fun of indeterministically self-modifying code ... cu -- -- Enrico Weigelt, metux IT service -- http://www.metux.de/ phone: +49 36207 519931 email: weig...@metux.de mobile: +49 151 27565287 icq: 210169427 skype: nekrad666 -- Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme -- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507135537.GF25423@nibiru.local
Re: Bits from the Release Team - Kicking off Wheezy
On Fri, May 06, 2011 at 02:04:20PM -0400, Michael Gilbert wrote: > It wasn't the GR itself. It was the fact that these changes to the NM > process were actually made. I suppose it is arguable that those changes > simply would not have happened without the GR, but that indicates more > of a lack of direct motivation within the new maintainer team. > > So, if it required the GR to motivate them, then I suppose it was a > necessity and ultimately a good thing, but my point is simply that its > better when motivation comes from within; rather than an applied > external force. I do not see how talking about the NM process or that GR is at all relevant in this thread, and please do not consider this message of mine an intent to contribute to it in any other way but to clarify a misrepresentation of a team I'm a member of. From the point of view of Front Desk, motivation has always been there: http://lists.debian.org/debian-devel-announce/2008/10/msg5.html but a GR was needed to be able to proceed, because the hands of FD were rather tied by this other GR: http://www.debian.org/vote/2008/vote_002 Please do not try to provide facts about the motivation or intentions of others unless you really know them, otherwise you run into the risk of misrepresenting other people, which is bad. Ciao, Enrico -- GPG key: 4096R/E7AD5568 2009-05-08 Enrico Zini signature.asc Description: Digital signature
Re: Best practice for cleaning autotools-generated files?
* Neil Williams schrieb: > No. Reverting the build to the point where it is equivalent to only > what is in the VCS adds completely unnecessary build dependencies and > build stages. Which ones exactly (besides the usual autotools) ? > VCS is not the final source, it's a tool for upstream to create the > final source. Lets replace "VCS" by "SCM" to get to the whole picture. I don't want to repeat the last decades if SCM theory and common practies here, but one of the major goals of an SCM infrastructure is that you always get can grab the source of some product in some specific version and run a fully rebuild, anytime. _Reliable_ reproducability. Tarballs that are not fully-automatically produced from the appropriate SCM tag (and can be reproduced anytime) simply dont meet that requirement. Maybe that all worked for you over many years. Fine. For me it always never worked, starting with common things like sysroot'ed crosscompiling (yes, I've propbably got a different scope than you, I'm coming from the embedded front) > I wouldn't want to work with any upstream which does not produce > stable tarballs which are available for download without going > anywhere near the VCS. I shouldn't even need to know what VCS is > being used if all I want to do is build the package for my own use. Guess what, I wouldn't work with upstreams that don't provide reliable release tags in the VCS (an VCS that I can import to git). Actually, I sometimes do, but it adds a lot of extra burden. My buildsystem (called Briegel), doesn't even bother with things like tarballs or patches (still supported, but unused for years now). It simply grabs the source tree from git by a normalized ref (eg. tag). Normalized means: there's a simple rule to transform a vector of package name and (also normalized) version number into a ref name. Period. Any changes needed in the source tree are done through SCM. (the processes behind may vary between individual packages). And these efforts, prodiving (long term) stable source trees is scope of the OSS-QM project (which acts as an intermediate between the upstream and consumers like my Briegel-based special-purpose distros). > Overall, this has little to do with Debian. True. It's a distro-agnostic issue. It's about how to be a good upstream to make distro's lifes easier and extends into the whole QM topics. > If those people are happy with their own system, then arguing > about tags versus tarballs is just bikeshedding. This argument is only valid as long as you only consider official upstream releases as your base. For example, I'm running automated git imports and cleanups for certain packages where upstream doesn't bother to provide a clean and stable source. You could use this, or copy the machinery and run it on your own (IOW: join OSS-QM ;-p) > Personal preferences have no place in a "best practice" statement - > that way lies the hell of mandating which VCS gets used, pandering to > whichever fanboy shouts the loudest. Well, I wouldn't set any rule which VCS upstreams should use. That's their choice. But they should provide anything that can be easily imported (eg. into git). But as distro maintainers (from whatever distro some might come from), we could agree to import all our required source trees into some standardized SCM where anyone can grab the (maybe cleaned) trees for a particular package/version automatically, and additionally maintain our bugfix and/or tailoring changes there. (-> OSS-QM ;-p) > > A good rule, IMHO, is: take *only* the *actual* source files (those > > written by the developers) and always regenerate all intermediate > > steps. Of course this means, _never ever_ use prebuild configure > > scripts etc. > > Fundamentally disagree. Obviously, you almost never have to change the input files. Or need to fix/tailor autotools. For mainline distros this might work well, but in the embedded world you're easily out of luck with this approach. > That is a very, very, bad rule IMHO. ./autogen.sh is a valuable > aid and it is pointless to try and assert that it should called > in all builds. It can become necessary if the latest upstream > tarball has gone stale but that's about it. And it *IS* necessary, as soon as you have to change some of the input files. In my daily projects, this happens on the majority of the packages. > > To go even one step further: dont use tarballs, instead upstream's > > VCS release tags (assuming upstream provides that ;-o) > > ... and that is simply rubbish. Just because some teams want to use VCS > tags does not mean that all upstreams would consider such a change. True, not every upstream wants to do that. And not every upstream wants to maintain stable releases. Well, that's life. Essentially we have two options to cope with that: a) everybody does the required fixups and workarounds all on its own b) collaborate and provide a stabelized midstream together and so save a lot of human workforce. > As an upstream
Bug#625971: ITP: wims-help -- help files for wims
Package: wnpp Severity: wishlist Owner: Georges Khaznadar * Package name: wims-help Version : 4.01 Upstream Authors : Bernadette Perrin-Riou , Marie-Claude David , Association WIMSEDU * URL : http://www.wimsedu.info * License : GPL-2 Programming Lang: phtml, and oef languages (parsed by Wims) Description : help files for wims Wims' modules implement every user interface beyond its main page, this package features the help modules. . WIMS is an acronym for WWW Interactive Mathematics Server. Nowadays WIMS serves much more than mathematic contents (physics, chemistry, biology, languages). . The WIMS educational platform features a rich set of resources and exercises either with free access or for personalised study. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507130617.29119.41897.report...@photos.khaznadar.fr
Re: 0-day NMUs for RC bugs without activity for 7 days?
On Sat, May 07, 2011 at 12:51:46PM +0200, Jakub Wilk wrote: > This works both ways. If a NMUer uploaded my package without a delay > and without a good reason[0], I want to be able to yell at him „you > are > a jerk (according to Developers Reference)!” > > Unfortunately, clueless NMUers do exist. :( It's all a matter of trade-offs. You're right that clueless NMUers do exist. OTOH very good NMUers that won't go ahead because a policy tell them not to go ahead exist as well. The question is whether there are more people in the first camp or people in the second camp. In my experience with NMUs---both as NMUers and as NMUed maintainer---I tend to believe in Debian we have way more people on the cautious side than clueless NMUers. Ultimately, having an overcautious policy for NMUs has the potential of blocking bug fixes and evolution in our project. I think we really need to fight that and that we should tolerate the risk of a handful clueless NMUers going ahead. Of course, when that happen, we should take good care of explaining to them why their approach was suboptimal and not in the best interest of Debian. I believe that way we can in the long run both increase our culture of doing good NMUs and avoiding overzealous blockers that will simply delay our procedures and increase frustration on people that know they can just go ahead and fix a broken package. But I agree that this policy should not force maintainers of several packages to ping their bug logs every 7 days, although at the very minimum I do expect a maintainer to post to an RC bug log at least once an "I'm on it" message. I've proposed before, for this change, to stress that the NMUer should do a best effort attempt to verify whether the maintainer is working on the fix, for instance by looking at VCS head of the package, asking on IRC, or the like. Finally, considering this policy has been in effect for 5 years or so, and considering that devref states guidelines rather than hard rules, I believe the _practical_ impact of this change would be very low. Cheers. -- Stefano Zacchiroli -o- PhD in Computer Science \ PostDoc @ Univ. Paris 7 zack@{upsilon.cc,pps.jussieu.fr,debian.org} -<>- http://upsilon.cc/zack/ Quando anche i santi ti voltano le spalle, | . |. I've fans everywhere ti resta John Fante -- V. Capossela ...| ..: |.. -- C. Adams signature.asc Description: Digital signature
Re: 0-day NMUs for RC bugs without activity for 7 days?
On Sat, 7 May 2011 12:51:46 +0200 Jakub Wilk wrote: > [0] No, 7 days without activity in the bug log is not a good enough > reason. Let's turn our empathy on and face it: we are not bug-fixing > monkeys and 7 days is a very short time frame. It's 7 days without maintainer response - if you're busy, as well get from time to time, it just means that you explain that in the bug report with just a single email. Let people know - we're not mind readers. RC bugs are on the radar of every DD, you may have told those you work with often that things are busy in real life, doesn't mean that someone else affected by the RC bug will know, unless you put just a brief note in the actual bug report. It's an open development model, so communicate openly - even if you're busy, sending a short reply to the bug report shouldn't be beyond what Debian can justifiably expect from any maintainer. 7 days without email is a good enough reason to assume that the maintainer isn't going to uploading a fix any time soon. -- Neil Williams = http://www.linux.codehelp.co.uk/ pgpccjyUvP9oi.pgp Description: PGP signature
Re: Best practice for cleaning autotools-generated files?
On Sat, 7 May 2011 13:33:53 +0200 Enrico Weigelt wrote: > * Josue Abarca schrieb: > > > From: /usr/share/doc/autotools-dev/README.Debian.gz > > "Example autogen.sh and debian/rules files can be found in > > /usr/share/doc/autotools-dev/examples. Do not use them as-is. Rather, > > properly customize your own." > > > > and from /usr/share/doc/autotools-dev/examples/autogen.sh > > " ...Still, this script is capable of cleaning > > # just about any possible mess of autoconf files." > > Aprops autogen.sh: > > IMHO the upstream always should supply a proper autogen.sh script, > which: > > a) removes all autogenerated files > b) calls all the autotools commands, which can be overridden >via environment (eg. $AUTOCONF, ...) > c) does _NOT_ call configure Nonsense. It is not the job of ./autogen.sh to revert to the VCS state and there's no harm in ./autogen.sh calling configure with whatever arguments are passed to ./autogen.sh - as long as those options have the same effect when passed to ./configure for subsequent runs during development. Again, this has nothing to do with Debian in most cases. Fix the few occasions where it might matter, don't make it a general requirement. > Perhaps we could write some spec how an "good" upstream autogen.sh > script should look like, fix them where necessary (so no individual > logic in the control files anymore for this part) and push the > patches to upstreams. Try that with my upstreams and I'll do the opposite. There are no good reasons to make such a spec which elevates personal preference to best practice. There is no technical reason for any of these suggestions, no benefit from these being turned into "best practice" and no gain in Debian from wider adoption of such *preferences*. > I've already done that for several packages in the OSS-QM project. > (which is not meant for Debian, but completely distro agnostic) I've no problem with that as long as you leave my upstreams alone. ;-) -- Neil Williams = http://www.linux.codehelp.co.uk/ pgpfyXBNj66gP.pgp Description: PGP signature
Re: Best practice for cleaning autotools-generated files?
* Henrique de Moraes Holschuh schrieb: > Yes. I think it was Cyrus IMAP that required -I in places where > autoreconf doesn't reach, so I called each tool separately. Which is > obviously a problem in autoreconf. Is it really a problem of autoreconf ? Imagine the following situation: I'm (as upstream) using serval macros in their own .m4 files (eg. in ./m4/, maybe even sorted into subdirs). Can autoreconf figure out the required search pathes all on its own ? cu -- -- Enrico Weigelt, metux IT service -- http://www.metux.de/ phone: +49 36207 519931 email: weig...@metux.de mobile: +49 151 27565287 icq: 210169427 skype: nekrad666 -- Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme -- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507125017.GD25423@nibiru.local
Re: Best practice for cleaning autotools-generated files?
On Sat, 7 May 2011 13:48:00 +0200 Enrico Weigelt wrote: > * Neil Williams schrieb: > > > Not every generated file needs to be cleaned. The list can vary if the > > package is very old and depends on how the package itself handles the > > clean internally. > > IMHO, *all* generated files should be cleaned away, just to be sure. > In the last decade I've encountered a lot of nasty problems w/ > forgotten generated files and similar stuff. No. Reverting the build to the point where it is equivalent to only what is in the VCS adds completely unnecessary build dependencies and build stages. VCS is not the final source, it's a tool for upstream to create the final source. I wouldn't want to work with any upstream which does not produce stable tarballs which are available for download without going anywhere near the VCS. I shouldn't even need to know what VCS is being used if all I want to do is build the package for my own use. I don't care what people say here, I am not putting the pre-tarball build stages into the upstream build system of my upstream packages. I usually document them in the README.svn but if it doesn't work, meh, use the tarball which I ensure does work. Overall, this has little to do with Debian. There is no consensus amongst upstream teams, no benefit in breaking stable projects by forcing a change in the build system and it just ends up as a simple preference by whoever has commit access to the upstream. If those people are happy with their own system, then arguing about tags versus tarballs is just bikeshedding. Personal preferences have no place in a "best practice" statement - that way lies the hell of mandating which VCS gets used, pandering to whichever fanboy shouts the loudest. > A good rule, IMHO, is: take *only* the *actual* source files (those > written by the developers) and always regenerate all intermediate > steps. Of course this means, _never ever_ use prebuild configure > scripts etc. Fundamentally disagree. That is a very, very, bad rule IMHO. ./autogen.sh is a valuable aid and it is pointless to try and assert that it should called in all builds. It can become necessary if the latest upstream tarball has gone stale but that's about it. > To go even one step further: dont use tarballs, instead upstream's > VCS release tags (assuming upstream provides that ;-o) ... and that is simply rubbish. Just because some teams want to use VCS tags does not mean that all upstreams would consider such a change. As an upstream developer, I simply refuse. The VCS is for upstream, including the tags. What gets into distributions is the tarball because then I can be sure that the package builds with 'make distcheck' without using a new £$ %$£ branch for every "£$£" change between releases. Lunacy. It's elevating a tool into a stick to hit people with. I've only got one answer to that and it's NO. One method does not fit all. I've got no problem with teams who do want to use VCS tags as a release process but don't push your model on others. All you'll get from my upstream projects is tarballs and if the VCS tag doesn't build, I really don't care. There is no such thing as a "release tag" in my upstreams, there might be tags but those are often just experiments. > > Building direct from VCS is a nice idea but sometimes, it just isn't > > worth the pain - let the build system build the tarball and package > > from the tarball. It's just easier. Or just change the build system > > and/or the tool. > > Actually, when building directly from VCS release tag causes big pain, > upstream is doing something _seriously_ wrong. So: fix the upstream > or put in an intermediate project for the (distro-independent) QM > works. (-> OSS-QM project). NO. I am one of those upstreams and I do not care if any VCS tag causes pain to anyone when building. The tarball is the released code, use it or find someone else to package it. Indeed, in a few upstream projects, the only files which ever get tagged are the debian/ files. ;-) -- Neil Williams = http://www.linux.codehelp.co.uk/ pgpeIVTViOyjJ.pgp Description: PGP signature
Re: Best practice for cleaning autotools-generated files?
* Henrique de Moraes Holschuh schrieb: > 1. No spawn from autotooling allowed in the VCS. EVER. .gitignore it >away at once. Autogenerated files inside the VCS repo are almost >always a bad idea. It was madness with CVS, it was bad mojo with >SVN, and it is certainly at least an annoyance with git. ACK, couldn't aggress more. BTW: last week, in a commercial project, a collegue actually proposed checking in prebuilt binaries into the VCS ... (well, not a VCS, but TFS) ... i've almost spit out my coffe when he said this ;-o > 2. Add an autogen.sh script that makes it trivial to retool everything as >needed. Keep it up-to-date. ACK. I'm already doing so for several packages in the OSS-QM project. Such steps heavily reduce the complexity of individual distro's build systems/package descriptors. It's all about standardized interfaces. > 3. I always hook the debian package to retool on package build. I have >no reason to trust whatever cruft upstream shipped, even if I happen to >be upstream at that moment in time. ACK. Preshipped autotools-generated files tend to be unreliable. The goal of "generated once - run everywhere" is simply missed. Just consider that for any reason just one of the input files have to be changed - it always needs a full regeneration, to be sure. At this point we can question whether we want the whole complexity (not just technologically, but also on human workforce) for deciding whether to do the regeneration or just always do it per default. >It also means I will instantly notice if upstream started doing very >"unpure" things to the build system, such as directly modifying >autogenerated files, and it lets me try to contact them (or fix my ways, >if I am upstream) and contain the braindamage before it causes any Debian >victims. ACK. And the downstream hopefully doesn't have similar insane ideas (eg. patching autogenerated files) ... cu -- -- Enrico Weigelt, metux IT service -- http://www.metux.de/ phone: +49 36207 519931 email: weig...@metux.de mobile: +49 151 27565287 icq: 210169427 skype: nekrad666 -- Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme -- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507123607.GC25423@nibiru.local
Re: 0-day NMUs for RC bugs without activity for 7 days?
Hi, On Sat, 07 May 2011, Jakub Wilk wrote: > This works both ways. If a NMUer uploaded my package without a delay > and without a good reason[0], I want to be able to yell at him „you > are a jerk (according to Developers Reference)!” No. First off, I never said that the rules are there to be able to badmouth people. So calling someone a jerk is never a good idea. And whatever mistake made, the NMU is still someone trying to help you, so you should respond calmly: "Thank you for the NMU but the issue really did not warrant a direct upload, you would have saved us some troubles by contacting me or by uploading to DELAYED." > [0] No, 7 days without activity in the bug log is not a good enough > reason. Let's turn our empathy on and face it: we are not bug-fixing > monkeys and 7 days is a very short time frame. We're speaking of RC bugs, you should not have plenty of RC bugs on your shoulders... Cheers, -- Raphaël Hertzog ◈ Debian Developer Follow my Debian News ▶ http://RaphaelHertzog.com (English) ▶ http://RaphaelHertzog.fr (Français) -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507121459.gb8...@rivendell.home.ouaza.com
Re: 0-day NMUs for RC bugs without activity for 7 days?
Le Fri, May 06, 2011 at 11:14:55PM +0200, Lucas Nussbaum a écrit : > > - The dev-ref documents the "default" choice. While there are cases where I > agree that uploading the fixed package ASAP is necessary, in most cases, we > can probably survive two more days with the bug, if we already survived more > than 7 days. Dear Lucas and everybody, I also think that not all RC bugs are equal. What Lucas wrote above is reflected by that some RC bugs will be uploaded with a “low” urgency, while some others will have a higher one. In http://bugs.debian.org/625449#85, I am proposing the following: - urgency=emergency: no communication needed. - urgency=high: normal upload if no activity for 7 days. - urgency=medium: DELAYED/2 if no activity for 7 days. - urgency=low: ask first. Cheers, -- Charles Plessy Tsurumi, Kanagawa, Japan -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507121003.gc15...@merveille.plessy.net
Re: Best practice for cleaning autotools-generated files?
* Neil Williams schrieb: > Not every generated file needs to be cleaned. The list can vary if the > package is very old and depends on how the package itself handles the > clean internally. IMHO, *all* generated files should be cleaned away, just to be sure. In the last decade I've encountered a lot of nasty problems w/ forgotten generated files and similar stuff. A good rule, IMHO, is: take *only* the *actual* source files (those written by the developers) and always regenerate all intermediate steps. Of course this means, _never ever_ use prebuild configure scripts etc. To go even one step further: dont use tarballs, instead upstream's VCS release tags (assuming upstream provides that ;-o) > >Since I use (or plan to use) git-buildpackage, I don't have a > > tarball which could serve as an authoritative whitelist. Thus an > > additional whitelist refresh step would be required every time I > > merge the upstream branch into the debian branch. That's bad. > > Is this a case of a problem with a tool causing more work? Fix the > tool? Change your choice of tool? No, git-buildpackage (IMHO) is the way to go. But at this point I wouldn't do any sourcetree manipulation outside of git, do all this within the vcs (eg. via tailored import/rebase mechanisms, etc). I'm using a similar approach w/ my Briegel embedded build system and OSS-QM source repositories. I could go more into detail if you like to. > Building direct from VCS is a nice idea but sometimes, it just isn't > worth the pain - let the build system build the tarball and package > from the tarball. It's just easier. Or just change the build system > and/or the tool. Actually, when building directly from VCS release tag causes big pain, upstream is doing something _seriously_ wrong. So: fix the upstream or put in an intermediate project for the (distro-independent) QM works. (-> OSS-QM project). cu -- -- Enrico Weigelt, metux IT service -- http://www.metux.de/ phone: +49 36207 519931 email: weig...@metux.de mobile: +49 151 27565287 icq: 210169427 skype: nekrad666 -- Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme -- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507114800.GB25423@nibiru.local
Re: Best practice for cleaning autotools-generated files?
* Josue Abarca schrieb: > From: /usr/share/doc/autotools-dev/README.Debian.gz > "Example autogen.sh and debian/rules files can be found in > /usr/share/doc/autotools-dev/examples. Do not use them as-is. Rather, > properly customize your own." > > and from /usr/share/doc/autotools-dev/examples/autogen.sh > " ...Still, this script is capable of cleaning > # just about any possible mess of autoconf files." Aprops autogen.sh: IMHO the upstream always should supply a proper autogen.sh script, which: a) removes all autogenerated files b) calls all the autotools commands, which can be overridden via environment (eg. $AUTOCONF, ...) c) does _NOT_ call configure Perhaps we could write some spec how an "good" upstream autogen.sh script should look like, fix them where necessary (so no individual logic in the control files anymore for this part) and push the patches to upstreams. I've already done that for several packages in the OSS-QM project. (which is not meant for Debian, but completely distro agnostic) cu -- -- Enrico Weigelt, metux IT service -- http://www.metux.de/ phone: +49 36207 519931 email: weig...@metux.de mobile: +49 151 27565287 icq: 210169427 skype: nekrad666 -- Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme -- -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507113353.GA25423@nibiru.local
Re: 0-day NMUs for RC bugs without activity for 7 days?
* Raphael Hertzog , 2011-05-07, 09:12: A patch was proposed (#625449) to implement in dvelopers-reference the "DELAYED/0 for upload fixing only release-critical bugs older than 7 days, without maintainer activity for 7 days" policy. I don't think that this policy is a good idea. But, since I was one of the drivers of the DEP that resulted in the current NMU policy (http://dep.debian.net/deps/dep1.html), I'm a bit biased, and I thought I would bring this proposed change to the attention of -devel@, so that we can get more feedback. I wonder how many people made use of that 0-day NMU rule. I have always interpreted those "rules" as “go ahead with NMUs, if the maintainer is a jerk and complain, we'll be on your side”. This works both ways. If a NMUer uploaded my package without a delay and without a good reason[0], I want to be able to yell at him „you are a jerk (according to Developers Reference)!” Unfortunately, clueless NMUers do exist. :( [0] No, 7 days without activity in the bug log is not a good enough reason. Let's turn our empathy on and face it: we are not bug-fixing monkeys and 7 days is a very short time frame. -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507105146.ga3...@jwilk.net
Re: glibc: causes segfault in Xorg
On Wed, May 04, 2011 at 02:30:35PM +0200, Aurelien Jarno wrote: > Le 04/05/2011 07:42, Steve M. Robbins a écrit : > > On Wed, May 04, 2011 at 12:10:48AM -0500, Jonathan Nieder wrote: > > > >> Sounds like http://sourceware.org/bugzilla/show_bug.cgi?id=12518 > >> which is fixed (sort of) by commit 0354e355 (2011-04-01). > > > > Oh my word. So glibc 2.13 breaks random binaries that happened to > > incorrectly use memcpy() instead of memmove()? What's wrong with the > > glibc developers (and Ulrich Drepper in particular)? > > > > I'm with Linus on this: let's just revert to the old behaviour. A > > tiny amount of clock cycles saved isn't worth the instability. > > > > Thanks, > > -Steve > > > > P.S. I tried rebuilding glibc myself locally, but gcc also segfaults > > in the process :-( > > > > Are you sure it is something related? Which gcc version are you using? > Do you have a backtrace point to the same issue? > > I am using this libc version for two months (on a CPU having ssse3 > instruction set), it is also used by other distributions, so I find > strange it breaks something so common than gcc. For XOrg it can be due > to the difference in configuration, that's why the problem stayed unnoticed. > Any news about that? Which GCC version is affected? Can you please send us the backtrace? -- Aurelien Jarno GPG: 1024D/F1BCDB73 aurel...@aurel32.net http://www.aurel32.net -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507102515.ga24...@hall.aurel32.net
Re: Bug#625865: ITP: ocportal -- ocPortal is a Content Management System for building and maintaining a dynamic website
On Saturday 07 May 2011 09:41:34 Raphael Hertzog wrote: > Hi, Hi, > On Fri, 06 May 2011, George Danchev wrote: > > * writing a meaningful ITP helps to grab attention, especially if there > > are multiple alternatives. Prove your point (ref: I'm upstream and I > > want to maintain it, doesn't magically buy you a slot into the archive) > > There's nothing to buy... only people offering to maintain packages in > Debian. But we should certainly not turn out upstream who are willing to > maintain the package in Debian. > > In fact I want more upstream involved in Debian! I didn't write exactly that. You simply twisted the meaning of what I wrote. Please, re-read, the keyword is *magically*. > (Unless someone does a serious review and has enough credit to convince > many people that the software is crap and would really be a big burden) We will accumulate tons of PHP CMSes that way, which doesn't seem to scale. In case of multiple alternatives, I'd rather prefer inclusion if enough arguments exist that it is better than already included ones. > > * writing lengthy rebuttals for well known facts from the past are quite > > unlikely, people has more important things to do. > > We're not speaking of lengthy rebuttals. I agree with Tshepang that the > answers were rather aggressive when you consider that you speak with > someone who is starting in the Debian community. > > Something like this would have perfectly done the job: > "We already have many PHP CMS in the archive, what does this one offer > that the other don't? Also PHP software tends to have a bad security track > record, is ocPortal any better in that regard?" That would have been better. I agree. > > * recognize the fact when someone says that chances are high you are > > about to be wasting your own time packaging $something. > > Everybody is free to do what they want with their own time, so you should > certainly not say anyone that they are wasting their time. If you believe > they are, you can certainly hint at better alternatives and let people > see by themselves if they wish to spend their time differently now that > they know of a possible alternative. Okay, I just gave a hint from my mind, let's see what happens. -- pub 4096R/0E4BD0AB -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201105071036.51004.danc...@spnet.net
Re: 0-day NMUs for RC bugs without activity for 7 days?
On Fri, 06 May 2011, Lucas Nussbaum wrote: > A patch was proposed (#625449) to implement in dvelopers-reference the > "DELAYED/0 for upload fixing only release-critical bugs older than 7 days, > without maintainer activity for 7 days" policy. > > I don't think that this policy is a good idea. But, since I was one of the > drivers of the DEP that resulted in the current NMU policy > (http://dep.debian.net/deps/dep1.html), I'm a bit biased, and I thought I > would > bring this proposed change to the attention of -devel@, so that we can get > more > feedback. I wonder how many people made use of that 0-day NMU rule. I have always interpreted those "rules" as “go ahead with NMUs, if the maintainer is a jerk and complain, we'll be on your side”. In other words, I think that the release team wants to encourage NMU to fix RC bugs more than saving 2 more days in the usual NMU process. So I believe the change is not very relevant. But I have no reason to object. That said I truly believe that RC bugs should be treated as high-priority issues by maintainers and that we should have the required information in the bug log ("I'll take care of it at ", "I have no idea how to fix this, any help welcome" + tag help, etc.). Cheers, -- Raphaël Hertzog ◈ Debian Developer Follow my Debian News ▶ http://RaphaelHertzog.com (English) ▶ http://RaphaelHertzog.fr (Français) -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110507071212.gd...@rivendell.home.ouaza.com