Re: [Cooker] Missing fontconfig-devel
On 2003-01-24(Fri) 15:45:01 -0800, George Mitchell wrote: I am trying look over the latest version of Xfree but my system requires XFree-devel which in turn requires fontconfig-devel which is not currently available on the mirrors. Anyone know what is going on with this package? [root@mobile root]# urpmf --provides fontconfig-devel libfontconfig1-devel:libfontconfig-devel[== 2.1-5mdk] libfontconfig1-devel:fontconfig-devel[== 2.1-5mdk] -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg88015/pgp0.pgp Description: PGP signature
Re: [Cooker] Old packages not removed from the mirrors
On 2003-01-23(Thu) 18:35:36 -0800, Quel Qun wrote: Hi, New rpm are coming through but the old ones are not deleted (at least from uninett) so it's only a matter of time before they run out of space. As an Mandrake employee (I think it's either warly or gc) said, Mandrake has absolutely no control over the mirrors Abel -- Quel Qun [EMAIL PROTECTED] -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg87778/pgp0.pgp Description: PGP signature
Re: [Cooker] gnome-theme-manager stalls
On 2003-01-22(Wed) 17:35:34 +0100, Frederic Crozat wrote: On Wed, 22 Jan 2003 11:14:50 -0500, Tim Lee wrote: Are you, by any chance, having a high security level ? I don't think so. In /etc/sysconfig/msec, SECURE_LEVEL is set at 2. What file or directory would affect FAM like that? And what permissions should it be set at? the main problem is when msec is set to 4 or 5, portmap isn't accessible and fam doesn't like that at all :(( I've got the same hanging of gnome-theme-manager observed, even though I confirmed portmap/xinetd works flawlessly, without any iptables rule blocking, and without msec. Filed bugzilla #993. Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg87476/pgp0.pgp Description: PGP signature
Re: [Cooker] [Bug 795] [XFree86] Cursor shadow is annoying
On 2003-01-20(Mon) 11:11:35 +0100, Chmouel Boudjnah wrote: [Bug 795] [EMAIL PROTECTED] writes: --- Additional Comments From [EMAIL PROTECTED] 2003-01-18 23:42 --- Still not fixed in Beta 2. this is a feature request not a 'fix to be done' !! XFree86 only has shadow cursors shipped by default, AFAIK. Besides, if non-shadowed version of cursor is used in Mandrake, I'm afraid hundreds of complaint will pop up instead of this one. :) Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg86845/pgp0.pgp Description: PGP signature
Re: [Cooker] [Bug 795] [XFree86] Cursor shadow is annoying
On 2003-01-20(Mon) 11:39:43 +0100, Pascal Terjan wrote: XFree86 only has shadow cursors shipped by default, AFAIK. Besides, if non-shadowed version of cursor is used in Mandrake, I'm afraid hundreds of complaint will pop up instead of this one. :) BTW Is this a problem here or as the white cursor some strange vertical lines inside it and its shadow ? Yup, I see such a vertical line too, which is only found in whiteglass but not redglass. Worth investigation. Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg86898/pgp0.pgp Description: PGP signature
Re: [Cooker] [Bug 795] [XFree86] Cursor shadow is annoying
On 2003-01-20(Mon) 16:06:17 +0100, Benjamin Pflugmann wrote: I can see it, too. AFAICT, it's the border of the shadow. Both parts, the cursor itself and the shadow have their own transparency. Therefore, the left part of the cursor (without the shadow) is more transparent then the right part (where cursor and shadow overlay): Hold it over some colored, plain part of the screen and you will see what I mean - the left part lets through a fraction more of the color than the right part. And the transition line between those two parts appears like a line. Aha. However, the redglass cursor doesn't contain this artifact Regarding the grandparent post: X ships with a cursor set without shadows called handhelds (see /usr/lib/X11/icons). As has been already mentioned several times on this list, you can change the default by changing default/index.theme accordingly. The handheld cursor is so small, I can't recognise if it has shadow or not :) For another matter, the preferred place for changing index.theme is in ~/.icons/index.theme instead, much better than modifying system wide setting. Xcursor(3x) manpage has documented the search paths of cursors. Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg86919/pgp0.pgp Description: PGP signature
Re: [Cooker] New release of the skeleton spec-file
On 2003-01-19(Sun) 21:00:36 -0500, Austin Acton wrote: I don't see any advantage of packaging the three icons as one archive. I use the following... [] I don't see how that's any worse than one tarball. While it is longer, I don't think clarity should take a back-seat to brevity. Same as %buildroot vs $RPM_BUILD_ROOT, different people may have different practice. Also, the skeleton should definitely include lib-stuff. %define libname lib%name%major and all the rest (descriptions, file-lists, depends, etc.) Not only does this clarify the lib policy, but the packager doesn't have to figure out the depends/provides for the libs which was VERY confusing for me at first, and is the same 90% of the time. How about making another skeleton spec which contains the libname stuff? BTW, it's time to introduce the %mklibname macros... Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg86810/pgp0.pgp Description: PGP signature
Re: [Cooker] Re: [Contrib-Rpm] Geramik-0.17-2mdk
On 2003-01-16(Thu) 09:14:32 +0100, Thierry Vignaud wrote: i have read the above thread and i agree on the stupid scripts destruction but i strongly disagree with these stupid requires as i did use geramik theme even if some think it's just not another gtk theme So you should know that these stupid requires are used to mimic what David wants, but not what *I* want. you see, we're in a free software world, and freedom means that if i want to use that particular theme, there's no political reason to deny me this right. You have to convince David instead, not me. whisper I regret _so_ much I have been immersed in this thread. /whisper -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg86322/pgp0.pgp Description: PGP signature
Re: [Cooker] Re: [Contrib-Rpm] Geramik-0.17-2mdk
On 2003-01-16(Thu) 09:14:25 +, Adam Williamson wrote:
So you should know that these stupid requires are used to mimic what
David wants, but not what *I* want.
That's simply not true, though. David said that Geramik shouldn't be
*used* by people who use GNOME, which is perfectly true - it's kind of
silly (no offence Thierry, but it is!) He never said it shouldn't be
installed on a machine which has GNOME on it. Linux is MULTIUSER,
remember. It should be possible to install Geramik on a machine used by
people who use GNOME and people who use KDE, for use only by the people
who use KDE.
I must admit I have forgotten this case: a multiuser machine, where some
use KDE and others use GNOME. But my argument (it invades users who use
default theme in GNOME) still holds true. Everything boils down to the
scriptlets:
%post
if [ $1 != 2 ]; then
if [ ! -f /etc/gtk/gtkrc ]; then
ln -s %{_datadir}/themes/Geramik/gtk/gtkrc /etc/gtk/gtkrc
fi
if [ ! -f /etc/gtk/gtkrc-2.0 ]; then
ln -s %{_datadir}/themes/Geramik/gtk-2.0/gtkrc-2.0 /etc/gtk-2.0/gtkrc
fi
fi
%preun
if [ $1 == 0 ]; then
if grep %{name} /etc/gtk/gtkrc /dev/null 21 ;then
rm -f /etc/gtk/gtkrc
fi
if grep %{name} /etc/gtk-2.0/gtkrc /dev/null 21 ;then
rm -f /etc/gtk-2.0/gtkrc
fi
fi
If one says that:
since sysadmins installed Geramik on those systems, and sysadmins are
supposed to be in control, users who use default theme in GNOME-only
environment shouldn't complain about their default theme changed, as
sysadmins are always correct.
Then I'll give up.
Hope GTK2_RC_FILE can be a solution. I'll wait and see.
--
Abel Cheung
GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
msg86336/pgp0.pgp
Description: PGP signature
Re: [Cooker] Re: [Contrib-Rpm] Geramik-0.17-2mdk
On 2003-01-14(Tue) 16:48:46 -0800, David Walser wrote: Now I see the point. Geramik is trying to do what Bluecurve did, right? Basically. More specifically, the color, font, and related settings that you configure for Keramik in KDE Control Center are what Geramik uses, rather than reading .gtkrc* files and being configured with Gnome tools. It's a Gtk+ theme for KDE users who have to use Gtk+ apps, and it keeps the look of the two consistent 100% of the time. I'm not entirely sure if Bluecurve is set up that way to be configured in one place, and always look the same no matter the toolkit, but if it doesn't do that it probably will in the future. I think I get your entire point now. So Geramik is for those who solely uses KDE as desktop, but don't touch GNOME desktop at all. They usually use KDE apps, but occasionally use GTK+ apps, and want to make those GTK+ apps look consistant with the whole Keramik theme. Isn't it? However, I see why Fred has removed your scripts, since the effect of your %post/%postun scripts has gone too far -- it affected *ALL* persons who use GTK+ software. Besides, it still remains correct if no %post/%postun scripts is used, since people who want it can still use it anytime, and nothing is lost. If they don't want it (I'd want to hear this from Laurent first, as it's really for KDE users. Crozat is the Gnome guy and he really has no business messing with it), then they should delete it from contrib, and probably additionally configure KDE to not use Keramik by default. As I have said, you've gone too far. At least you have to write the scripts in such a way that, only those who use desktop in THAT way is affected. Yes, you have to make sure that people installed KDE desktop only, and if people install GNOME desktop later, your changes have to be undone immediately and nicely. That means you need to make Geramik cope with people's change in behavoir. If the above is not achieved, Geramik will just be a normal theme, nothing more, nothing less (quoted). No, that's still wrong. I think you'll understand from my explanation what Geramik really is. If you just want a Gtk+ theme that looks like Keramik, Geramik is not what you want. Such normal themes do exist. Yes, I do understand your point now, but still think what you proposition (its intended usage) is flawed: can knife manufacturers say that knives are only for slicing food, and they prohibit people to use it differently? Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg86172/pgp0.pgp Description: PGP signature
Re: [Cooker] Re: [Contrib-Rpm] Geramik-0.17-2mdk
On 2003-01-15(Wed) 10:57:06 -0800, David Walser wrote: That means you need to make Geramik cope with people's change in behavoir. Still doesn't make sense. I'll agree with one point you've made...if there were a way to have it install itself by default, but only activate for users that are in KDE (kind of like .gtkrc-kde), that would be better. Is this possible? The simplest way I can think of is: Requires: kdebase Conflicts: gnome-desktop gnome-session This should fit your intended usage. You can try convincing Fred if your scriptlets can be added back if the above 2 lines are added as well. Your arguments still failed to convince me, so this is the most I can help. Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg86239/pgp0.pgp Description: PGP signature
Re: [Cooker] Re: [Contrib-Rpm] Geramik-0.17-2mdk
On 2003-01-16(Thu) 07:11:43 +0100, Thierry Vignaud wrote: R.I.P. Deaddog [EMAIL PROTECTED] writes: Is this possible? The simplest way I can think of is: Requires: kdebase Conflicts: gnome-desktop gnome-session and if one want both kde and gnome ? and if one does not want kde but does want geramik theme for its gtk+ apps ? David said that, Geramik is only for those people who only use KDE as desktop, and just occasionally use GTK+ apps under KDE. Those who use GNOME desktop shouldn't ever use it. That's why I propose these absurd Requires/Conflicts, if he insist getting his %post/%postun scripts back. But he keep telling me I misunderstood him, while I just quoted his intended usage of Geramik. His original %post/%postun links /etc/gtk/gtkrc and /etc/gtk-2.0/gtkrc to Geramik theme. This act invades people who use default theme on GNOME desktop, and Crozat removed his scripts. You can see the earlier thread. Somehow I become tired of this thread. :) -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg86311/pgp0.pgp Description: PGP signature
Re: [Cooker] Re: [CHRPM] freetype2-2.1.3-4mdk
On 2003-01-14(Tue) 15:07:53 +0200, Buchan Milne wrote: Guillaume Rousse wrote: I'd like to have other people feedback on this point. If everyone agrees, the plf package could be dropped. For =9.0 yes. Yup, even multibyte TTF work well here. Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg85963/pgp0.pgp Description: PGP signature
Re: [Cooker] Re: [Contrib-Rpm] Geramik-0.17-2mdk
On 2003-01-14(Tue) 05:17:05 -0800, David Walser wrote: --- Frederic Crozat [EMAIL PROTECTED] wrote: Name: Geramik Version : 0.17 * Tue Jan 14 2003 Frederic Crozat [EMAIL PROTECTED] 0.17-2mdk - Drop the gtkrc (un)installation stuff in %post/%postun, this is completely broken.. Gah!!! You destroyed it! All that post/postun stuff was heavily tested and works perfectly. All bugs reported were squashed. Expect complaints now. Would you mind explaning all of the below anyway? I think that this is not only broken, but *abusive* too. Theme is a per-user setting, why is it trying to make itself as a global theme by default? -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg85965/pgp0.pgp Description: PGP signature
Re: [Cooker] Re: [QA] cooker main changes
On 2003-01-14(Tue) 09:08:48 -0500, Brian J. Murrell wrote: On Tue, Jan 14, 2003 at 03:03:51PM +0100, Mandrakesoft packages database wrote: better tool in e2fsprogs Which tool is it that is better? I guess it is refering to /sbin/resize2fs. Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg85971/pgp0.pgp Description: PGP signature
Re: [Cooker] Re: [Contrib-Rpm] Geramik-0.17-2mdk
On 2003-01-14(Tue) 06:33:17 -0800, David Walser wrote: I think that this is not only broken, but *abusive* too. Theme is a per-user setting, why is it trying to make itself as a global theme by default? The idea is if you install this you *want* that. BUT, ^^^ *THIS* idea is broken in this case. How many people will install *only* 1 theme? Generally people install many themes, and switch one by one later on if they like. If Geramik can establish itself as the default theme, why others can't? Or let them fight -- the first one installed wins? Abel if you don't, it also is very careful to not get in your way. If you set something else as default, it won't override it. If you just delete that gtkrc it puts in place (to disable this being the default) then upgrade the package, it *won't* put it back. (Here I'm talking about the old version before fcrozat ruined it. The new version may not be as careful as the old one was). -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg85973/pgp0.pgp Description: PGP signature
Re: [Cooker] Re: [Contrib-Rpm] Geramik-0.17-2mdk
On 2003-01-14(Tue) 07:11:34 -0800, David Walser wrote: --- Frederic Crozat [EMAIL PROTECTED] wrote: This is right.. Theme packages should not screw up default theme.. There is no point for discussion on that.. Geramik is not a normal theme. The only useful purpose it serves is being the default theme, alongside KDE's default Keramik theme. Looks-wise neither are all that wonderful, and noone's gonna choose to use them over something else. Their only usefulness is in their consistency, it's basically the same idea as Bluecurve. Now I see the point. Geramik is trying to do what Bluecurve did, right? If this is so, then the current changes are not enough. You have to persuade Mandrake developers to use G/Keramik to replace the default mdk theme, and adopt G/Keramik as the default instead. In additional to this, you also need to get the GNOME/KDE behavior sync'ed. To sum up, Mandrake have to follow RedHat's decision to make GNOME/KDE similar enough. If the above is not achieved, Geramik will just be a normal theme, nothing more, nothing less (quoted). Having two desktops looking similar but behave differently is worse than what we currently have, IMHO. Moreover, Gemarik is more than buggy Yes, there's only one author, I'm sure he'd appreciate help. I would think having a well-working Geramik would be a desirable thing for MDK. Yes, not only for Mandrake, but good for everybody who like K/Geramik too. Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg86024/pgp0.pgp Description: PGP signature
Re: [Cooker] Meta-Packages ?
On 2003-01-13(Mon) 22:13:32 +0100, Steffen Barszus wrote: urpmi kde-desktop would installs a bunch of kde-packages urpmi mandrake-development installs gcc / gcc-c++ kernel-header and so on. urpmi mandrake-simple-desktop could provide an small kde-desktop with only one programm for each task . I'd support doing this for the devel packages, e.g. bison/flex/gcc/binutils and so on. This has been discussed for several times but nobody is in a position that can take action. GNOME has this already done, as gnome2 package. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg85839/pgp0.pgp Description: PGP signature
Re: [Cooker] Re: [CHRPM] automake1.6-1.7.2-2mdk
On 2003-01-02(Thu) 16:48:11 -0500, Yura Gusev wrote: Stefan van der Eijk said: --=-=-= Name: automake1.6 Version : 1.7.2 URL : http://sources.redhat.com/automake/ --=-=-= Please update URL http://www.gnu.org/directory/GNU/automake.html Bugzilla can be a better place for such suggestions... so that they will not be forgotten. No? -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg84872/pgp0.pgp Description: PGP signature
Re: [Cooker] libgda2 and libgnomedb2
On 2003-01-02(Thu) 21:47:42 -0500, Quel Qun wrote: Is there any reason why the last 0.9.0 version is not used in cooker instead of the old version. It seems that only gnumeric requires libgda0, but nothing prevents libgda2 to be installed besides it. As a side note, libgnomedb2 should be required to build glade2 and provide full gnome support. As libgnomedb RPM is already in /incoming, it can be put into cooker after some polishing. Fred here? Listening? One note though. Should the RPM be renamed to meageant? -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg84880/pgp0.pgp Description: PGP signature
[Cooker] More package can be removed from cooker
Hi all, Lately some packages were nuked from cooker (was it gc or warly who did it?), so I may miss the boat. Another good candidate to be removed is GXedit, which is unmaintained since Jan 2000. I suppose nobody is using it? -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg84513/pgp0.pgp Description: PGP signature
Re: [Cooker] More package can be removed from cooker
On 2002-12-26(Thu) 07:04:45 -0500, Charles A Edwards wrote: I suppose nobody is using it? I use it almost exclusively. It is mush faster to load than gedit. Even so, since the project itself is no longer being maintained and the URL for it no longer exists (only the 1998 sf dl page) and since I can build/rebuild an rpm for it if wanted/needed, I would have no serious objection to its removal. I think cvs.mandrakesoft.com will have the spec preserved. Besides, there are various RPMs and tarballs downloadable, so rebuilding it can be easy. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg84524/pgp0.pgp Description: PGP signature
Re: [Cooker] More package can be removed from cooker
On 2002-12-26(Thu) 15:14:20 +0100, Stefan van der Eijk wrote: How about SVGATextMode? No new release since 2000-09-02 (on: http://www.ibiblio.org/pub/Linux/utils/console/). Yup, resizecons can almost replace it, and frame buffer console is another (IMHO better) alternative. So the list grows: * GXedit * SVGATextMode Anymore? R.I.P. Deaddog wrote: [..] Another good candidate to be removed is GXedit, which is unmaintained since Jan 2000. I suppose nobody is using it? -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg84525/pgp0.pgp Description: PGP signature
Re: [Cooker] Building athlon kernel rpm from src.rpm
On 2002-12-25(Wed) 04:16:21 -0600, Lonnie Borntreger wrote: When I give the command: rpmbuild --rebuild --target=athlon kernel-2.4.20.2mdk-1-1mdk.src.rpm the process crashes after a couple of minutes... Is this the correct command? I have also tried using --target=athlon-gnu-linux with the same results :( You may need to rearrange the options, with --target first, followed by (re)build options (such as -bb, -ba, --rebuild), and finally the --with/without options. Not confirmed all combination of options, but with -ba and --target, this worked for me. Here's what I do. 1) create a file called ~/.rpmrc with the following: buildarchtranslate: athlon: athlon buildarchtranslate: i686: athlon buildarchtranslate: k6: athlon buildarchtranslate: i586: athlon buildarchtranslate: i486: athlon buildarchtranslate: i386: athlon 2) Build using: rpm --rebuild --with smp --with source --without secure --without enterprise --without BOOT --without doc kernel-2.4.20.2mdk-1-1mdk.src.rpm -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg84494/pgp0.pgp Description: PGP signature
Re: [Cooker] what's /usr/bin/autom4te-2.13?
On 2002-10-28(Mon) 16:06:55 +0100, Guillaume Cottenceau wrote:
ac-wrapper: ouch, couldn't call binary (/usr/bin/autom4te-2.13).
autoheader-2.5x: /usr/bin/autom4te failed with exit status: 2
I've had this problem before. Not sure what causes the problem. Try
When my script wrongly detects 2.1 and actually the things try to
call autom4te which is a 2.5 only thingy.
[...]
Not all config tools comply with the rules of the version detection
script (/usr/bin/autoconf) by GC.
Patches welcome :).
GC, IMHO your autoconf version detecting rules are
fairly complete; usually it's configure.{in,ac} that
need to be patched instead :)
--
Abel Cheung
GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
msg84386/pgp0.pgp
Description: PGP signature
Re: [Cooker] %configure2_5x rpm macro
On 2002-12-19(Thu) 10:22:51 +0100, Gwenole Beauchesne wrote: I had this problem with gstreamer-plugins, without an explicit libtoolize call the C++ plugins wouldn't link to libstdc++. Well, what we do in libtool is simply a workaround. You have to fix gstream-plugins there too. ;-) Basically, C++ code or programs have to be linked with g++. However, if it doesn't use anything from libstdc++ you can consider linking with gcc but with an extra -lsupc++. BTW, point granted for %configure2_5x additions, will do. OK, put them into bugzilla #700 and #701, so they won't be forgotten. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc msg84393/pgp0.pgp Description: PGP signature
Re: [Cooker] %configure2_5x rpm macro
On 2002-12-12(Thu) 13:18:02 +0100, G?tz Waschk wrote:
I think the %configure2_5x macro should be fixed. ATM it expands to:
[goetz@klama SRPMS]$ rpm --eval %configure2_5x
CONFIGURE_TOP=${CONFIGURE_TOP:-.};
CFLAGS=${CFLAGS:--O2 -fomit-frame-pointer -pipe -march=i586 -mcpu=pentiumpro\} ;
export CFLAGS ;
CXXFLAGS=${CXXFLAGS:--O2 -fomit-frame-pointer -pipe -march=i586
-mcpu=pentiu\mpro} ; export CXXFLAGS ;
FFLAGS=${FFLAGS:--O2 -fomit-frame-pointer -pipe -march=i586 -mcpu=pentiumpro\} ;
export FFLAGS ;
(cd $CONFIGURE_TOP; [ -f configure.in ] libtoolize --copy --force) ;
[schnipp]
With autoconf 2.5 the autoconf file can be named configure.ac instead
of configure.in. So the libtoolize call should check also check for
configure.ac
Yup, though not many packages uses configure.ac solely, I think rpm
maintainer (flepied?) should consider adding this bit now.
Yet another problem:
[...snip...]
$CONFIGURE_TOP/configure --build %{_target_platform} --host
%{_target_platform} --target %{_target_platform} \\\
[...snip...]
Adding --host and --target will put autoconf 2.5x into cross compile
mode. The result is, some packages will contain funny named binaries
such as i586-mandrake-linux-foobar, and need to use ugly hacks to fix
it. It's good for all to turn --host and --target off. This has been
discussed in various places, I think.
--
GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
msg84188/pgp0.pgp
Description: PGP signature
RE: [Cooker] More intellegent logging postprocessing
On Mon, 29 Apr 2002, Borsenkow Andrej wrote:
In addition to the failure caused by the config bundled
inside RPM :(
Oh, I forgot to ask - what is the problem with distributed config?
Without detailed study it looks more or less O.K.
-andrej
It's this line:
source remote { udp(); };
If it's not commented out, syslog-ng simply refuses to start.
Even running syslog-ng -d on command line doesn't help tracing
the problem.
--
Abel Cheung
GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Gnome has close to no menu items
On Sun, 28 Apr 2002, Alexander Skwar wrote: When starting Gnome 2, I've got close to no menu items in the Gnome menu. That's with a blank user I created for testing. Is this expected? Alexander Skwar Yes, I think fcrozat has posted many times on cooker list that at least a completely working Gnome 2 desktop is to be packaged before adding any mandrake customization. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Re: Why are .rpmnew files created?
On Sun, 28 Apr 2002, Brian J. Murrell wrote: On Sun, Apr 28, 2002 at 09:45:24PM +0200, Guillaume Cottenceau wrote: The user or any script or something.. In my case anyway, no. Nothing has edited these config files for which .rpmnew files are being created. Next time I do an update I will rpm -Va first and verify that none of the .rpmew files had and md5 errors during the rpm -Va. b. Probably check the access/modification time as well. I'm afraid that problem is deeper than mere md5 sum. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] gconf-editor missing?
On 28 Apr 2002, Steve Fox wrote: Any idea why gconf-editor wouldn't be on my system? It used to be. I noticed something on one of the gnome lists about a build fix, so maybe that's why it's not in the current packages? libGConf2_4-1.1.9-2mdk libGConf1-devel-1.0.9-2mdk pkgconfig-0.12.0-1mdk GConf-1.0.9-2mdk GConf2-1.1.9-2mdk libGConf2_4-devel-1.1.9-2mdk libGConf1-1.0.9-2mdk It has never been in cooker (yet), gconf-editor is a seperate package. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] More intellegent logging postprocessing
On Sun, 28 Apr 2002, Geoffrey Lee wrote: some known cases. http://www.balabit.hu/en/products/syslog-ng/ There are also packages which will watch output from (ordinary or NG) syslog and trigger on what they see. Better if you list out the available of the RPM in contrib. :-) -- Geoff. In addition to the failure caused by the config bundled inside RPM :( syslog-ng documentation is virtually zero -- otherwise it's a great replacement for syslog. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] More intellegent logging postprocessing
On Sun, 28 Apr 2002, Geoffrey Lee wrote: There is some documentation here : http://www.balabit.hu/static/syslog-ng/reference/book1.html But it's not written in an easy-to-undertand format .. it seems to be quite technical. -- Geoff. This reference is created around the time during syslog-ng 1.1.x series. There are whole slew of features/config formats not documented since then. Bad. :( Besides, it's true that the above doc is for technies -- I still remembered struggling for many hours before working out a usable config during those time... -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] can't build scrollkeeper .src.rpm
On Fri, 26 Apr 2002, Stefan van der Eijk wrote: Any idea's? Build fine here.. Try look better at errors during compilation (errors you posted are not causing any problem.. Doesn't build on my system either. http://eijk.homelinux.org/build/cooker/i586/problem/scrollkeeper-0.3.7-1mdk.src.rpm.txt Probably some BuildRequires missing... Hi Stefan, This problem is very likely not BR related. %doc error is non-fatal; but the excerpt below is suspicious. It seems to indicate a failure to search for scrollkeeper-* executables in $PATH, hence unable to rebuild scrollkeeper database. However, it's worthy to note that, Bryan's problem sounds to be different from yours -- scrollkeeper-rebuilddb runs fine, and the reason of error is unknown. I have encountered such a failure myself (and mysterically second build succeeds). Anyway, we may wait for fcrozat to build 0.3.8. This is an important release, since all Gnome 2 OMF files are heading for scrollkeeper 0.3.8 compliance :) Abel # Build/rebuild the catalog rm -rf /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/var/lib/scrollkeeper /bin/sh ./mkinstalldirs /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/var/lib/scrollkeeper mkdir /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/var mkdir /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/var/lib mkdir /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/var/lib/scrollkeeper /bin/sh ./mkinstalldirs /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/var/log mkdir /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/var/log echo `date +\%b %d %X\` Installing ScrollKeeper `scrollkeeper-config --version`... /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/var/log/scrollkeeper.log /bin/sh: scrollkeeper-config: command not found /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/usr/bin/scrollkeeper-rebuilddb -q -p /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/var/lib/scrollkeeper /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/usr/bin/scrollkeeper-rebuilddb: scrollkeeper-config: command not found /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/usr/bin/scrollkeeper-rebuilddb: /var/log/scrollkeeper.log: Permission denied /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/usr/bin/scrollkeeper-rebuilddb: scrollkeeper-update: command not found /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/usr/bin/scrollkeeper-rebuilddb: /var/log/scrollkeeper.log: Permission denied make[3]: Leaving directory `/home/cooker/RPM/BUILD/scrollkeeper-0.3.7' [...] RPM build errors: File not found: /home/cooker/tmp/scrollkeeper-0.3.7-1mdk-buildroot/var/lib/scrollkeeper Bad exit status from /home/cooker/tmp/rpm-tmp.59095 (%doc) = -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Why is xmatrix missing from xscreensaver?
On Mon, 22 Apr 2002, Alan wrote: In looking at the spec file for xscreensaver I notice that xmatrix and extrusion are set not to build. extrusion I can understand because it requires a specialized GL extrusion library. As a matter of fact, the gle library is in cooker too. Abel Why is xmatrix disabled? If it is copyright nonsense, then lament should not be there either. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Why is xmatrix missing from xscreensaver?
On Tue, 23 Apr 2002, Alan wrote: In looking at the spec file for xscreensaver I notice that xmatrix and extrusion are set not to build. extrusion I can understand because it requires a specialized GL extrusion library. As a matter of fact, the gle library is in cooker too. OK... Then why are either disabled? xmatrix: of course it's the stupid copyright extrusion: gle is in contrib (and things on main distro shouldn't depend on contrib stuff), and gle is not used by anything else for now. Perhaps adding a toggle on xscreensaver RPM is a good thing. But xscreensaver may have some bigger change (choose either the official xscreensaver or gnome-branched one), so it's for fcrozat to consider it... -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Why is xmatrix missing from xscreensaver?
On Tue, 23 Apr 2002, Alan wrote: xmatrix: of course it's the stupid copyright If this is correct, then someone is stretching copyright paranoia a little too far. During the old days, it's pretty dangerous. When things settle down, people tend to ignore it. But if it's brought out again, then you're doomed. Are there any other distributions that disable this one? Dunno. Why are there other undisabled screensavers based on films? (Starwars and lament, for example.) Not sure, have to ask mdksoft people here. extrusion: gle is in contrib (and things on main distro shouldn't depend on contrib stuff), and gle is not used by anything else for now. Perhaps adding a toggle on xscreensaver RPM is a good thing. But xscreensaver may have some bigger change (choose either the official xscreensaver or gnome-branched one), so it's for fcrozat to consider it... Or just add a xscreensaver-extrusion rpm in contrib. I'd guess it's quite impossible here, due to complications in compiling it, but not 100% sure (haven't checked it). -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Why is xmatrix missing from xscreensaver?
On Tue, 23 Apr 2002, Goetz Waschk wrote: Am Dienstag, 23. April 2002, 17:10:08 Uhr MET, schrieb R.I.P. Deaddog: On Tue, 23 Apr 2002, Alan wrote: extrusion: gle is in contrib (and things on main [...] Why hasn't anyone of you looked at the spec file of latest xscreensaver? There's an option to enable extrusion and xmatix at build time. Try something like --with xmatrix --with extrusion and it will work. Thanks for the pointer. My knowledge is based on pre-4.00 days, so you indicated what I know is outdated. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Some man pages are shown incorrectly
On Mon, 22 Apr 2002, Thierry Vignaud wrote: Guillaume Rousse [EMAIL PROTECTED] writes: here is a temporary fix: [...] While on this topic: could you please decomment /usr/lib/perl5/man line in /etc/man.conf ? too late for 2mdk. now, export LESSCHARSET=latin1 and less won't owerwrite it with utf-8. i've to figure why utf-8 is broken. i'll decomment perl man pages in next release :-) Seems utf-8 is not broken. It doesn't work just because it is using ISO-10646-1 charset; if the input stream contains something not belonging to UTF-8, it is 'rolled back', thus causing lost character. Of course I can be wrong, this is just a uneducated interpretation of less source. If a sane default value is needed, I'd suggest LESSCHARSET=koi8-r instead of latin1 or utf-8, since koi8-r is very close to a 'raw' charset -- almost any char is display as is, except control characters. Even latin1 excludes the 0x80-0x9F range. I still happen to remember some comment like Mandrake is the most iso-8859-x'ed Linux distro, but not i18n'ed. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Galeon needs to be rebuild
On Sun, 21 Apr 2002, Charles A Edwards wrote: Regards. Or you can get an unofficial mandrake based rpm at : ftp://ftp.ibiblio.org/pub/Linux/distributions/contrib/texstar/Mandrake-8.2-i586 nautilus-mozilla needs to be updated for a successful texstar upgrade. I do not know if you meant a mdk update but there will be no new mdk pkg for nautilus-mozzila. That package belongs with old Gnome and has been superseded by pkg nautilus-gtkhtml which is new Gnome. Then probably texstar need to release another nautilus 1.0.x package which is recompiled against mozilla 1.0rc1. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Strange conflicts with new packages
On Sun, 21 Apr 2002, Han wrote: file /usr/lib/gtk-2.0/modules/libgail.so from install of libgail13-0.13-1mdk conflicts with file from package libgail11-0.11-1mdk Please search cooker archive, there has been discussion on this problem. This must be an error: file /usr/share/vim/lang/fr from install of vim-common-6.1-5mdk conflicts with file from package libgimpprint1-4.2.1-0.pre5.1mdk file /usr/share/vim/lang/pl from install of vim-common-6.1-5mdk conflicts with file from package libgimpprint1-4.2.1-0.pre5.1mdk file /usr/share/vim/lang/sk from install of vim-common-6.1-5mdk conflicts with file from package libgimpprint1-4.2.1-0.pre5.1mdk Both vim-common and libgimpprint1 share the problem. Have been extensively discussed as well. For vim-common, please remove it and re-install. For libgimpprint, I have asked Till why it owns /usr/share/locale/*, but was denied. Till, here? -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] libgail13-0.13-1mdk
On Fri, 19 Apr 2002, Frederic Crozat wrote: gail will be freezed for GNOME 2.0 release since it is part of the platform.. Oh, I have thought that it is similar to libgda etc -- won't freeze. If it's freezing like other components do, then it's worthy to wait. Then no more fix is necessary :) Besides, I really think it's reasonable to get at least libgail.so and libferret.so splitted out, since: [.] But we can't be sure using new GTK input libs (libferret and libgail) with old libgailutils will work.. Actually very likely to break (I have tried it before). Sounds like this boils down to packaging question: should non-versioned libraries/modules bundled with the main library as one package, or split it out? -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] vim-common-6.1-4mdk is dangerous
On 18 Apr 2002, Guillaume Cottenceau wrote:
Since:
[gc@bi ~] ls -l /usr/share/vim/lang
lrwxrwxrwx1 root root 17 Apr 3 05:22 /usr/share/vim/lang -
/usr/share/locale/
And now:
[gc@bi ~] rpm -qplv /RPMS/vim-common-6.1-4mdk.i586.rpm | grep /usr/share/vim/lang$
drwxr-xr-x2 rootroot0 Apr 17 18:43 /usr/share/vim/lang
Installing vim-common-6.1-4mdk screwed up all my
/usr/share/locale directory (thx Titi).
Luckily I stopped when urpmi complains about conflict between
vim and locales-*. Otherwise all my customized translations will
be burned in hell as well. Such changes ought to put into
real machine testing before releasing.
Attached a patch to check carefully whether $datadir/vim/lang
is a real directory before removing it.
Besides, I remembered somebody suggested to keep /bin/vi link
to /bin/vim-minimal forever, so that one can still use vi if
/usr is not mounted. Is it acceptable?
--
Abel Cheung
GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
--- vim.spec2002-04-18 00:42:58.0 +0800
+++ vim.spec.new2002-04-18 17:59:18.0 +0800
-268,7 +268,11
# So we've to symlink locales there
# But to prevent update faillure, we must first be sure a link
# creation won't fail because old directory is still there
-rm -fr %_datadir/vim/lang/*/||:
+if test -d %{_datadir}/vim/lang -a ! -L %{_datadir}/vim/lang; then
+ rm -fr %{_datadir}/vim/lang
+else
+ rm -f %{_datadir}/vim/lang
+fi
%post minimal
update-alternatives --install /bin/vi vi /bin/vim-minimal 10
[Cooker] CUPS owns /usr/share/locale/* [Was: vim-common-6.1-4mdk isdangerous]
On Thu, 18 Apr 2002, Pascal Terjan wrote: Guillaume Cottenceau wrote: Installing vim-common-6.1-4mdk screwed up all my /usr/share/locale directory (thx Titi). Didn't it conflict with anything ? Here I got file conflicts with : kde-i18n-fr-2.2.2-1mdk koffice-i18n-fr-1.1.1-1mdk cups-common-1.1.14-3mdk locales-fr-2.3.1.3-1mdk locales-ja-2.3.1.3-1mdk libgimpprint1-4.2.1-0.pre5.1mdk So I didn't install it ! That means the above packages try to own /usr/share/locale/*, which should belong to locales-* packages instead. I'm not sure if KDE people will fix it at all, but for cups-common and libgimpprint1, perhaps Till will fix this soon? PS Hence CC'ed to Till -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] vim-common-6.1-4mdk is dangerous
On Thu, 18 Apr 2002, Geoffrey Lee wrote: Attached a patch to check carefully whether $datadir/vim/lang is a real directory before removing it. Besides, I remembered somebody suggested to keep /bin/vi link to /bin/vim-minimal forever, so that one can still use vi if /usr is not mounted. Is it acceptable? One can still invoke vim-minimal manually even if /usr is not mounted .. the only difference is that they wouldn't have the vi link. -- G. Yeah you're correct. I missed that :) -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] multiload-applet-2 keeps crashing.
On 18 Apr 2002, Brice Figureau wrote: Installed it a few minutes ago, and it seems to work now. Another question: it seems that the bonobo-activation and the gconf2 servers are not stopped when I log out, is it normal ? Yes, just like what oafd and gconfd behave for Gnome 1.x. :/ -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
[Cooker] Re: CUPS owns /usr/share/locale/* [Was: vim-common-6.1-4mdk isdangerous]
On Thu, 18 Apr 2002, Till Kamppeter wrote: cups-common does not own /usr/share/locale. I have checked it. Till Sorry I didn't make it clear. I mean /usr/share/locale/*, not /usr/share/locale itself. Here's the output: # rpm -qf /usr/share/locale/sv locales-sv-2.3.1.3-1mdk cups-common-1.1.4-3mdk libgimpprint1-4.2.1-5.pre5.1mdk Abel R.I.P. Deaddog wrote: That means the above packages try to own /usr/share/locale/*, which should belong to locales-* packages instead. I'm not sure if KDE people will fix it at all, but for cups-common and libgimpprint1, perhaps Till will fix this soon? PS Hence CC'ed to Till -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Re: [CHRPM] nautilus-gtkhtml-0.3.1-2mdk
On Thu, 18 Apr 2002, Goetz Waschk wrote: Am Donnerstag, 18. April 2002, 15:45:30 Uhr MET, schrieb Frederic Crozat: - Obsoletes nautilus-mozilla Why? Is the current nautilus-mozilla package too unstable? nautilus-mozilla is a Gnome 1.x thing, and nautilus-gtkhtml is a Gnoem 2.x thing. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Re: [CHRPM] nautilus-gtkhtml-0.3.1-2mdk
On Thu, 18 Apr 2002, Goetz Waschk wrote: - Obsoletes nautilus-mozilla Why? Is the current nautilus-mozilla package too unstable? Is it not unstable, it doesn't work with nautilus 1.1.x ... (GNOME 1 vs GNOME 2 ...) I thought I saw a Gnome2 package named nautilus-mozilla, what happened to it? Fcrozat.. is it that nautilus-gtkhtml shouldn't obsolete nautilus-mozilla? Goetz is correct, nautilus-mozilla has been porting to Gnome 2 too, and is different from nautilus-gtkhtml. However, I just compiled nautilus-mozilla, and it bombs upon starting seems not quite usable yet :-| -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Re: [CHRPM] vim-6.1-5mdk
On Thu, 18 Apr 2002, Charles A Edwards wrote: --=-=-= Name: vim Relocations: (not relocateable) Version : 6.1 Vendor: MandrakeSoft Release : 5mdk Build Date: Thu Apr 18 14:51:39 2002 file /usr/share/vim/lang/fr from install of vim-common-6.1-5mdk conflicts with file from package libgimpprint1-4.2.1-0.pre5.1mdk file /usr/share/vim/lang/pl from install of vim-common-6.1-5mdk conflicts with file from package libgimpprint1-4.2.1-0.pre5.1mdk file /usr/share/vim/lang/sk from install of vim-common-6.1-5mdk conflicts with file from package libgimpprint1-4.2.1-0.pre5.1mdk Installation failed If there's any vim version = 6.1-3mdk , remove it first before upgrading to newer version. Then there will be no more conflict. :) Otherwise your /usr/share/locale will be trashed like gc does... -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] libgail13-0.13-1mdk
On Wed, 17 Apr 2002, Frederic Crozat wrote: I don't plan to split gail more : gail is not yet API/ABI frozen but it will be frozen for GNOME 2 release.. So, it won't add anything good to split gail now.. I'll wait for gail 1.0 and after, I'll see if we really need to slip it.. In the mean time... it's cooker :)) Heh heh, this starts to be the magical reason for every problem in the world :) But I think it start to be pressing now. If gail doesn't freeze anytime soon, then you'll receive more and more complaints whenever new releases are done... Besides, I really think it's reasonable to get at least libgail.so and libferret.so splitted out, since: 1. They are there since day one 2. They are moved only once, from /usr/lib to /usr/lib/gtk-2.0/something - So they are unlikely to move or change again. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] rpm issues from this past weekend's updates
On Mon, 15 Apr 2002, Todd Lyons wrote: A couple of rpms produced some strange output: scrollkeeperwarning: /etc/scrollkeeper.conf created as /etc/scrollkeeper.conf.rpmnew ## /usr/share/gnome/help/mailcheck/C/mailcheck.xml:308: error: End tag : expected '' para ^ Then a little further down: gnome-applets ## error: db4 error(-30988) from dbcursor-c_get: DB_PAGE_NOTFOUND: Requested page not found libdb4 wasn't installed. I removed gnome-applets, installed libdb4, then reinstalled gnome-applets and received no errors. Should gnome applets have libdb4 as a Requires? Gnome-applets has nothing to do with libdb4. The Requested page not found solely comes from the static libdb4 inside rpm-4.0.4. Abel It might be important that on this machine I don't have KDE3 installed. Blue skies... Todd -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] libgail13-0.13-1mdk
On Tue, 16 Apr 2002, Charles A Edwards wrote:
file /usr/lib/gtk-2.0/modules/libgail.so from install of
libgail13-0.13-1mdk conflicts with file from package libgail11-0.11-1mdk
fcrozat, probably it's appropriate to split out
/usr/lib/gtk-2.0/modules/lib{ferret,gail}.so into
another subpackage (something like gail or libgail-modules)?
Then make libgail13 Requires: libgail-modules etc.?
--
Abel Cheung
GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
[Cooker] Re: [CHRPM] gnome-games-1.90.2-1mdk
On Thu, 11 Apr 2002, Frederic Crozat wrote:
--=-=-=
Name: gnome-games Relocations: (not relocateable)
Version : 1.90.2Vendor: MandrakeSoft
Release : 1mdk Build Date: Thu Apr 11 11:36:49 2002
--=-=-=
* Thu Apr 11 2002 Frederic Crozat [EMAIL PROTECTED] 1.90.2-1mdk
- Release 1.90.2 (GNOME 2)
- Remove patch0 (no longer needed)
Noticed a couples of problems in gnome-games:
1.
Typo in the %%pre section, that caused failure in
ghostfile (scores file) generation:
%create_ghostfile %{gamesdir}/$i.scores games.games 0664
create_ghostfile needs 4 arguments, so the 2 games should
probably be seperated by a space here?
2.
why remove gnometris? There were some legal issues in the
old days; but now there are still so many *tris clone in
both Cooker and contrib, so I guess adding gnometris wouldn't
hurt?
3.
Score files are only generated once during first time install.
So if newer version of gnome-games is installed (which may contain
some new games), new score files are never added. This problem
exists in Gnome 1.x days too (sorry that I forgot to report).
4.
glines need to be suid, and glines score file is missing...
--
Abel Cheung
GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Postfix Vs. qmail?
On Tue, 16 Apr 2002, civileme wrote: Not trying to start a flame war I would really like to find a good mta and that doesn't include sendmail. First I would like to know if this vulnerability still exists or if it has been fixed in mandrake? http://cr.yp.to/maildisasters/postfix.html You will see there that postfix has been fixed on 2-28-2002 and on 11-29-2001. We keep up with real exploits, and our security team issues fixes. Normally a question like that would not get an answer on this list. May I suggest aubacribiung to a help list like [EMAIL PROTECTED] or [EMAIL PROTECTED] For even more appropriate list, you can subscribe to [EMAIL PROTECTED] or [EMAIL PROTECTED] -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] More scrollkeeper errors
On Sat, 13 Apr 2002, Alexander Skwar wrote: gnome-media ## /usr/bin/scrollkeeper-update: /usr/local/share/omf: Datei oder Verzeichnis nicht gefunden /usr/bin/scrollkeeper-update: /opt/gnome/share/omf: Datei oder Verzeichnis nicht gefunden /usr/bin/scrollkeeper-update: /opt/gnome-2.0/share/omf: Datei oder Verzeichnis nicht gefunden /usr/bin/scrollkeeper-update: /opt/kde/omf: Datei oder Verzeichnis nicht gefunden Normal. scrollkeeper is trying to search for various default OMF directories. kwar@teich askwar]$ rpm -q gnome-media gnome-media-1.287.113-1mdk What kind of a version number is that??? You have to ask Gnome people about this. Abel Alexander Skwar -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Error when installing scrollkeeper-0.3.6-1mdk
On Sat, 13 Apr 2002, Alexander Skwar wrote: I just got this when installing scrollkeeper-0.3.6-1mdk: scrollkeeper## /usr/share/gnome/help/mailcheck/C/mailcheck.xml:308: error: End tag : expected '' para ^ [askwar@teich askwar]$ rpm -qf /usr/share/gnome/help/mailcheck/C/mailcheck.xml gnome-panel-1.5.16-1mdk This has been fixed in Gnome CVS. Try the attached patch against gnome-panel. Just a typo. Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc Index: gnome-panel/applets/gen_util/help/C/mailcheck/mailcheck.xml === RCS file: /cvs/gnome/gnome-panel/applets/gen_util/help/C/mailcheck/mailcheck.xml,v retrieving revision 1.9 retrieving revision 1.10 diff -u -r1.9 -r1.10 --- gnome-panel/applets/gen_util/help/C/mailcheck/mailcheck.xml 28 Mar 2002 11:34:25 - 1.9 +++ gnome-panel/applets/gen_util/help/C/mailcheck/mailcheck.xml 11 Apr 2002 00:47:37 +- 1.10 -304,7 +304,7 to check for mails automatically. This option will be turned off if an error (eg. wrong password) occured. - /para + /para para You can adjust the range for checking whether mail has arrived between 1440 minutes (once a day) to never. Setting it to never
Re: [Cooker] More scrollkeeper errors
On Sat, 13 Apr 2002, Alexander Skwar wrote: »Alexander Skwar« sagte am 2002-04-13 um 01:18:30 +0200 : /usr/bin/scrollkeeper-update: /usr/local/share/omf: Datei oder Verzeichnis nicht gefunden /usr/bin/scrollkeeper-update: /opt/gnome/share/omf: Datei oder Verzeichnis nicht gefunden /usr/bin/scrollkeeper-update: /opt/gnome-2.0/share/omf: Datei oder Verzeichnis nicht gefunden /usr/bin/scrollkeeper-update: /opt/kde/omf: Datei oder Verzeichnis nicht gefunden [root@teich RPMS]# less /etc/scrollkeeper.conf OMF_DIR=/usr/share/omf:/usr/local/share/omf:/opt/gnome/share/omf:/opt/gnome-2.0/share/omf:/opt/kde/omf No idea how these directories got into the conf file. They are default settings. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] gnome-applets install error
On Sat, 13 Apr 2002, Alexander Skwar wrote:
gnome-applets ##
warning: failed to load external entity
/etc/gconf/schemas/{battstat*,cdplayer*,charpick*,drivemount*,geyes*,gkb*,gtik*,gweather*,mini-commander*,
»/etc/gconf/schemas/{battstat*,cdplayer*,charpick*,drivemount*,geyes*,gkb*,gtik*,gweather*,mini-commander*,«
konnte nicht geöffnet werden: Datei oder Verzeichnis nicht gefunden
warning: failed to load external entity modemlights*,panel-menu*}
»modemlights*,panel-menu*}« konnte nicht geöffnet werden: Datei oder Verzeichnis
nicht gefunden
Fehler: execution of %post scriptlet from gnome-applets-1.98.0-1mdk failed, exit
status 1
[askwar@teich askwar]$ rpm -q gnome-applets
gnome-applets-1.98.0-1mdk
Hi fcrozat,
Sounds like gnome-applets.spec has genuine problem in %%post here:
the shell expension doesn't work correctly. Attached a patch that
hopefully fix it by avoiding shell expension completely. The source
of problem seems to be an extra space; anyway I haven't tested it
thoroughly so not sure about that yet.
Abel
--
Abel Cheung
GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
--- gnome-applets.spec.orig 2002-04-10 21:34:38.0 +0800
+++ gnome-applets.spec 2002-04-13 13:48:36.0 +0800
-47,17 +47,14
%{find_lang} %{name}-2.0 --with-gnome --all-name
-rm -f $RPM_BUILD_ROOT%{_datadir}/applets/Utility/gnotes_applet.desktop
-rm -f $RPM_BUILD_ROOT%{_datadir}/applets/Utility/bug-applet.desktop
-
%clean
rm -rf $RPM_BUILD_ROOT
%post
/sbin/ldconfig
if [ -x /usr/bin/scrollkeeper-update ]; then /usr/bin/scrollkeeper-update -q || true
; fi
-for SCHEMA in
%{_sysconfdir}/gconf/schemas/{battstat*,cdplayer*,charpick*,drivemount*,geyes*,gkb*,gtik*,gweather*,mini-commander*,
modemlights*,panel-menu*} ; do
- GCONF_CONFIG_SOURCE=`gconftool-2 --get-default-source` gconftool-2
--makefile-install-rule $SCHEMA /dev/null
+for SCHEMA in battstat cdplayer charpick drivemount geyes gkb gtik gweather
+mini-commander modemlights panel-menu; do
+ GCONF_CONFIG_SOURCE=`gconftool-2 --get-default-source` gconftool-2
+--makefile-install-rule %{_sysconfdir}/gconf/schemas/${SCHEMA}.schemas /dev/null
done
%postun
Re: [Cooker] what about pre-version policy ?
On Wed, 27 Mar 2002, Guillaume Rousse wrote: openssh-3.1p1 openssl-0.9.6c libhdf5_0-1.4.2p1 These are not preversion : p is for patch level ... (ie openssh 3.1 patch level 1..) Nope. p for portable, http://www.openssh.org/portable.html You're right, I confused with other programs.. Anyway, it is not preversion... OK, i was wrong for openssh, as it won't ever change, but not for other packages: as alphanumeric characters prevent proper version ordering, they should be moved to release part. No, fcrozat is still correct. E.g. for openssl: 0.9.6c 0.9.6b 0.9.6a 0.9.6 For HDF5: 1.4.2p1 1.4.2 Concisely, versionsomething is always greater than version only. Hence it creates problem for pre/alpha/beta versions, but not patched versions. That's what the package versioning doc in Mandrake meant. -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] what about pre-version policy ?
On Wed, 27 Mar 2002, Guillaume Rousse wrote: http://www.linux-mandrake.com/howtos/mdk-rpm/advanced.html#PREVERSIONS-NAMING but they are actually plenty of counter-examples in distro: openssh-3.1p1 openssl-0.9.6c libhdf5_0-1.4.2p1 etc... No, the aforementioned problem only applies to pre/alpha/beta versions. Since e.g. 1.0pre1 1.0 , one needs to use 1.0-0.pre1.1mdk to distinguish the _pre_ status. However, in the above cases, it's patch level, not pre. So we have: openssh: 3.1 3.1p1 openssl: 0.9.6 0.9.6a 0.9.6b 0.9.6c libhdf5: 1.4.2 1.4.2p1 Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] status of GNOME 2
On Wed, 27 Mar 2002, Frederic Crozat wrote: here is a quick status about GNOME 2 integration in cooker : the latest part of GNOME 2 main desktop package has been uploaded (gnome-panel) which replace the old gnome-core package... Actually the situation is: gnome-core has been splitted into gnome-panel, gnome-desktop and gnome-session. gnome-core itself is obsoleted. Currently, almost no Mandrake customization has been integrated in cooker packages, since I prefer having a working stock GNOME 2 before patching it :)) It means that Mandrake menu is not visible from GNOME 2 panel.. Sounds good, at least people will have a working Gnome 2 for testing before the Mandrake customization is done. Oh, thanks a million for your great work... Abel Nautilus 2 (ie nautilus based on GNOME 2) has not been uploaded yet.. Same thing for gnome-applets and other desktop packages.. I'll be off for this week and I'll be at GUADEC 3 next week (maybe I'll see some cooker folks there...) so don't expect any GNOME 2 modifications until Monday April 8th ... Don't report missing features in current packages, there are not there :)) -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Re: Re: gtk+2.0-2.0.0-1mdk
On Mon, 25 Mar 2002, Frederic Crozat wrote: No way because GTK+ team has not said it is supported yet or no way for some other reason, like you don't like gtk-x11 and gtk-linux-fb conflicting with each other? Because it is not supported yet.. And why wait until it is officially supported for Cooker? For release, I understand, but this is the unstable testing branch. Because I'll have enough work (and trouble) with GNOME2 and GTK+2 X11 without adding linuxfb... If you want, you can build linuxfb part on your system and send patches/fixes to GTK+ team.. I've thought about this for quite some time; but now I think it's quite impossible to use alternative or any sort of trick here. Whether applications link with gtk-linuxfb or gtk-x11 is decided on *BUILD* time, so it can possibly break everything if you switch backend without rebuilding apps. (it's just my claim, actually I can't get through the gtk+ barrier; dare not build anything beyond gtk using linux-fb target) Abel PS I'll try it when I have some time -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Re: libxml2-devel: wrong #include in libxml.m4
On 29 Mar 2002, [ISO-8859-1] Frédéric Crozat wrote: In /usr/share/aclocal/libxml.m4 one can find #include xmlversion.h two times. This is wrong, it has to be #include libxml/xmlversion.h. For libxml2, when using AM_PATH_XML macro, xml2-config --cflags is used and add -I/usr/include/libxml2/libxml -I/usr/include/libxml2 to compile test program.. mr@nibbler ~ $ xml2-config --cflags -I/usr/include/libxml2 There is nothing wrong here.. OK, then maybe xml2-config is broken? In fact, this is a problem with version 2.4.18 and above (and I was testing with Mdk 8.2 version..) I'll check with libxml author to see if this is the right fix to do.. I vaguely remembered this change is intentionally done by libxml author. Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Re: [CHRPM] libgnomeprintui2.0-1.110.0-1mdk
On Mon, 25 Mar 2002, Charles A Edwards wrote: urpmi is not handling all the depends needed to install, or more exactly the depend have depend which urpmi is not handling. # urpmi libgnomeprintui2.0_0 libgnomeprintui2.0_0-devel To satisfy dependencies, the following packages are going to be installed (27 MB): [..] Is it OK? (Y/n) y error: failed dependencies: gtk+2.0-backend-devel = 2.0.0 is needed by libgtk+2.0_0-devel-2.0.0-1mdk gtk+2 = 1.3.13 is needed by libgtk+2-linuxfb-1.3.13-2mdk libgdk_pixbuf-1.3.so.13 is needed by libgtk+2-linuxfb-1.3.13-2mdk libgtk+2-devel = 1.3.13 is needed by libgtk+2-linuxfb-devel-1.3.13-2mdk Installation failed You're quite correct here, since the package splitting for glib/gtk+ is quite different between 1.3.x and 2.0. For now, it's suggested to completely remove any packages that depend on glib 1.3.x before installing glib-2.0. Abel -- Abel Cheung GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] gpg error, recent versions of gnupg
On Sat, 23 Mar 2002, Todd Lyons wrote: I'm getting errors uploading/downloading keys to *.keyserver.net with gnupg. The same keys work just fine with *.pgp.net (which eliminates variables like firewalls, diskspace, mailer configuration, etc). Running in debug mode, it works successfully until the very last step where it's expecting a response back indicating success or failure. The remote just closes the connection. I've tried it on an 8.0, an 8.1, an 8.2, and a fully cookerized box and all do the same thing. It works with no problems on a RH 6.2 box (and of course a correspondingly old and probably_vulnerable_somehow gnupg). What I have is much more absurd then your case. I have to retrieve keys *by chance*. If the key is really present in keyserver.net, then I usually got a Connection reset by peer for first time. After repeated execution of gpg, it will succeed, but the number of execution is random! Once I need to execute it 10 times before the key retrieval. OTOH, seems this observation is specific to Mandrake. I have never seen such problem under Debian Woody. Abel Here's what an 8.0 box looks like without debugging: [todd@barracuda todd]$ gpg --keyserver search.keyserver.net --recv-keys 9716DB0 A gpg: requesting key 9716DB0A from search.keyserver.net ... gpg: [fd 4]: read error: Connection reset by peer gpg: Total number processed: 0 gpg: no valid OpenPGP data found. [todd@barracuda todd]$ gpg --keyserver wwwkeys.us.pgp.net --recv-keys 9716DB0A gpg: requesting key 9716DB0A from wwwkeys.us.pgp.net ... gpg: key 9716DB0A: public key imported gpg: /home/todd/.gnupg/trustdb.gpg: trustdb created gpg: Total number processed: 1 gpg: imported: 1 -- GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Re: New GnomeICU release
On Thu, 21 Mar 2002, Brian J. Murrell wrote: On Fri, Mar 22, 2002 at 04:21:27AM +0800, R.I.P. Deaddog wrote: On Thu, 21 Mar 2002, Brian J. Murrell wrote: Have you checked the CDs? 0.98.2 is already present in CD#2... Oh wow. There it is. Too bad it don't work. I have only been able to add one contact. Guess I will go Goolin' to see what is up. Yes, it's *very* unstable; it crashes from time to time in my box. Don't have time to trace it; hope future relase can be more stable. But somehow all crashes are related to people sending me message. Abel -- GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Re: [CHRPM] gtk+2.0-2.0.0-1mdk
On Fri, 22 Mar 2002, Frederic Crozat wrote: On Thu, Mar 21, 2002 at 07:47:55PM +0100, Frederic Crozat wrote: Name: gtk+2.0 Relocations: (not relocateable) Version : 2.0.0 Vendor: MandrakeSoft Release : 1mdk Build Date: Thu Mar 21 19:05:= 59 2002 =20 - Disabled linux-fb target temporarily, causing HUGE memory leak=20 (anyway, x11 and linux-fb target can't coexist for now.) Oh no! Is this HUGE memory leak a known bug to the GTK+ developers? Why can't x11 and linux-fb coexist? The library names are non-conflicting. At least in the 1.3 releases, there was libgtk-x11 and libgtk-linux-fb, and so on. It would be even nicer to have run-time selection, but compile time is good enough for now. Ask Abel, I just adapted his rpms.. They REALLY cann't coexist for now. Try compiling both x11 target and linux-fb target seperately, and compare their gtk+-2.0.pc and gdk-2.0.pc. For gtk+ 1.3.1x package, the x11 copy of gtk+-2.0.pc overwrites the one from linux-fb target, that means any program linked against gtk+ would be unconditionally linked against libX11 as well. Abel PS Actually, x11 is still the only 'blessed' target in gtk+ 2.0.0. You can't even compile the linux-fb target from vanilla source. -- GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] New GnomeICU release
On Thu, 21 Mar 2002, Brian J. Murrell wrote: OK, so now that 8.2 out and Cooker is cooking again can we get GnomeICU updated to the 0.98.2 release? Have you checked the CDs? 0.98.2 is already present in CD#2... Abel -- GPG Key: (0xC67186FF) http://deaddog.org/gpg.asc
Re: [Cooker] Which cooker SRPM contains popt
On 21 Mar 2002, William Kenworthy wrote: Which cooker SRPM contains popt ? a dir *popt* on uio shows nothing, but popt* binary rpm's exist. BillK It's the rpm-4.0.x source RPM. Abel
Re: [Cooker] Mandrake way of life
On Sun, 17 Mar 2002, Warly wrote: So the datum is: release date is March the 15th. Now we need to do it in this timeframe (and we are the 17th), and we have no other choice. In a few days 8.2 updates will be released, and them you will have the real stable and polished distro you want. Debian has no real realease date, as a consequence doing a stable release is just a piece of cake, anybody can do it. Maybe our model is not good. Debian's model has it's fault too. Recently one famous Debian hacker has left, because he thinks nobody is eager to stabilize woody. The current woody and sid works well, so nobody thinks an 'absolutely stable' distro is necessary. Thinking of a better model, I am not sure what I could choose. No deadline means no hurry, and if the last moments didn't exist, nothing good would be done. Moreover if we wait too much, new versions of nearly all the tools in the distro will be released, and their respective authors will not care anymore fixing bugs in the old version we would have included. I do think that our model is not so bad, and that we are reaching a good compromise between cutting edge and stability, but /this/ compromise _is needed_. An unrelated suggestion: I think it's better to let cooker subscribers here know more about the time frame, so that testers will know: OK, the time is near, we should do more testing or work harder to submit bug reports. Abel
[Cooker] [Contrib-RPM] jasper uploaded
The following signed package is uploaded to /incoming. = Name: jasper Relocations: (not relocateable) Version : 1.500.4 Vendor: MandrakeSoft Release : 1mdk Build Date: Sun Mar 17 08:22:02 2002 Install date: (not installed) Build Host: mobile.maddog.lan Group : System/Libraries Source RPM: (none) Size: 645482 License: BSD-like Packager: Abel Cheung [EMAIL PROTECTED] URL : http://www.ece.ubc.ca/~mdadams/jasper/ Summary : A library that can manipulate JPEG 2000 format images Description : JasPer is a software-based implementation of the codec specified in the emerging JPEG-2000 Part-1 standard (i.e., ISO/IEC 15444-1). The JasPer software is written in the C programming language. This software has also been submitted to the ISO for inclusion in the JPEG-2000 Part-5 standard (as an official reference implementation). = I've put a copy of source RPM in http://deaddog.org/files/Mandrake/contrib/SRPMS/ as well. Abel
[Cooker] [Contrib-Rpm] gcal uploaded
New gcal source RPM is uploaded to /incoming: === Name: gcal Relocations: (not relocateable) Version : 3.01 Vendor: MandrakeSoft Release : 5mdk Build Date: Mon Mar 11 23:47:33 2002 Install date: Mon Mar 11 23:50:54 2002 Build Host: mobile.maddog.lan Group : OfficeSource RPM: gcal-3.01-5mdk.src.rpm Size: 2369363 License: GPL Packager: Abel Cheung [EMAIL PROTECTED] URL : http://www.gnu.org/software/gcal/gcal.html Summary : Print calendars Description : Gcal is a program for printing calendars. Gcal displays a calendar for a month or a year, eternal holiday lists and fixed date lists, in many ways. The program correctly omits the dates that were skipped when the current Gregorian calendar replaced the earlier Julian calendar. Additionally, it can generate air line distance between over 400 cities on earth, moon phases in these cities, whole annual holidays, etc. === * Mon Mar 11 2002 Abel Cheung [EMAIL PROTECTED] 3.01-5mdk - Patch1: Misc fixes taken from debian - Add missing manpages - find_lang - Use _install_info and _remove_install_info === Also available under http://www.deaddog.org/files/Mandrake/contrib/SRPMS/gcal-3.01-5mdk.src.rpm . Abel
Re: [Cooker] 8.2 bug - junkbuster
On Sat, 16 Mar 2002, Hoyt wrote: When installed, gives message that group nobody does not exist, running as root. Yes, nobody has already been replaced with nogroup, I assume. You can have a look at setup-2.2.0 changelog. BTW, I'd want to ask if you can help me test a new junkbuster. My submission of new junkbuster has been deleted without reason, and I will plan to submit it again. Here it is: http://www.deaddog.org/files/Mandrake/contrib/SRPMS/junkbuster-2.9.10-1mdk.src.rpm Note that the config file format changed a lot, and old config is not quite usable.. Abel
Re: [Cooker] GNOME 2.0
On Thu, 14 Mar 2002, Frederic Crozat wrote: hi, what about GNOME 2.0 ?! Shall it be included in MDK8.2 it is currently Beta2 And final version will be May 1st 2002 ... No comment.. Please blame me here, I promised fcrozat that I can package it, but not delivered any of my promise here :-( Just started to package newest Gnome 2, hope it can be done within 1 week -- please aim at my brain and shoot otherwise :( Abel
Re: [Cooker] Tonight: rebuilding an entire Mandrake distro fromscratch!
On 12 Mar 2002, Bryan Paxton wrote: locales: localedef is hrm broken (at least from my end), segfaults, reason unknown. Search the mail archives, I made a few posts about this. armagetron: eh, bad code or patch iirc ImageMagick: Compiles fine, but there's some wierd perl regex going on inside that spec file which makes it impossible for RPM to find Magick++-config and other scripts/programs/whatever in the devel package. Seems it's only you who faces problem here. If RPM fails to find *-config scripts, then it's usually because installation already fails, not because the perl regex caused them to disappear. Did you check the whole build log? Abel perl-DB: Fails during tests, thus it exits : ) hdf5: Hmm I don't remember why on this one
Re: [Cooker] AfterStep 1.8.11
On Tue, 12 Mar 2002, Kimmo Hovi wrote: I just uploaded AfterStep 1.8.11-2mdk source rpm. Working on updating the default start menu. Sadly none of the active developers use the default menu (wonder why ;), so it's ages old. Wonder if this will be included in 8.2 Better yet, provide it on some other sites so that others can try it. /incoming directory is not readable :( Abel
Re: [Cooker] lynx and apache?
On Tue, 12 Mar 2002, Vincent Meyer, MD wrote:
Why does apache need lynx? Trying to unselect lynx during package selection
in RC1 says it will de-select about a dozen packages, including apache,
apache-conf, apache-mod-perl, etc.
excerpt from /etc/init.d/apache:
LYNX=lynx -dump
stop)
stop
;;
extendedstatus)
$LYNX $STATUSURL | awk ' /process$/ { print; exit } { print } '
echo
.
Abel
Re: [Cooker] update-menus core dumping
On 9 Mar 2002, Bryan Paxton wrote: It perhaps might be the latest gcc, specifically libstdc++, but it could be my optimizations that made something in libstdc++ fubar... I'm going to compile for i686, but with the default rpm opt flags and see what's what : ) I'll email back with results : ) Those being that either libstdc doesn't like compiling with the below flags, or gcc is happy with them (2.9.6 from cooker): i686 -03 -funroll-loops -ffast-math -malign-double -mcpu=pentiumpro -march=pentiumpro -fomit-frame-pointer -fno-exceptions -fno-strength-reduce I've encountered quite a few 'risky' programs that segfaults with too much optimization, turning off -ffast-math and/or -fno-strength-reduce usually helps. Just my experience. Abel (PIII - 850) Anyway, so no problem with libstdc++ : )
Re: [Cooker] galeon and URL completion
On 9 Mar 2002, Borsenkow Andrej wrote: I can complete list of previous URLs by hitting TAB in URL input field. I can also select one of them with ENTER but Galeon starts download immediately. How can I select for editing, so that I can change it. A quick workaround is to press ESC quick enough immediately after pressing ENTER. But of course this is not a good way to solve such issue. Abel TIA -andrej
Re: [Cooker] LHA (was: rpm option for removing wrong dependency?)
On Sat, 9 Mar 2002, huug wrote: is it possible to remove a wrong dependency from a rpm package? I'd like to test further but can't install KDE because kdeutils-2.2.2-24mdk.i586.rpm misses lha and I won't wait for the update. ;-) Well, LHA ain't that big: The point is, *why* is lha needed? KDE packages are known to contain weird dependencies without reason. Abel
Re: [Cooker] rc1
On Sat, 9 Mar 2002, Warly wrote: Tomorrow we will be in deep freeze. That's mean that no package upload will be allowed except for severe showstoppers. RC1 contains 2 important pbs that are now nearly fixed, symlinks are broken on CDs with supermount enable (fixed in the latest cooker kernel), and USB hub device detection (especially for printers) is non 100% functional (removing /usr/sbin/usbmodules fix the pb). For this last week of testing, we will concentrate on really important bugs, letting other ones to be corrected in 8.2 updates. I have some concerns about the soft on CD1, everyone want its program to be on CD1, as a consequence some other more important may be rejected to CD2, if you think that crucial package for any installation are on CD2, whine. Questions? I have sent email to fcrozat about dysfunctional gtm along with a patch, waiting for his reply. Abel
[Cooker] Re: [CHRPM] textutils-2.0.17-5mdk
--=-=-= Name: textutilsRelocations: (not relocateable) Version : 2.0.17Vendor: MandrakeSoft Release : 5mdk Build Date: Thu Mar 7 15:54:46 2002 --=-=-= * Thu Mar 07 2002 Thierry Vignaud [EMAIL PROTECTED] 2.0.17-5mdk - add infrastructure to ease update This is a wish unrelated to messed info handling; I hope the outdated traslations contained in textutils be removed (and removing patch100 as well). Almost all translations contained in textutils tarball are newer than those bundled specially in Mandrake textutils package which are dated 1998-xx-xx. Abel
Re: [Cooker] libwmf not obsoleting previous version
On 5 Mar 2002, Guillaume Cottenceau wrote: $ rpm -qa | grep libwmf02 libwmf02_1-021-1mdk libwmf02_2-022-1mdk That's the point of the new lib policy So we just get more and more and more libs forever? Is this why I have Yes, we have decided that this problem is far better than previous situation where upgrade basically broke things because of missing/conflicting libraries to keep manually uninstalling libfoo18-devel packages whenever I need to -Uvh libfoo19-devel? Surely something more elegant can be conceived Yeah, some time ago I was thinking what you are thinking now, and even created packages that worked like what you have described But later I discovered that, this will usually break things in subtle way Let's consider gal and gtkhtml as example; usually gtkhtml needs to be built against some specific version of gal, and if non-matching devel packages are mixed (eg by upgrading libgal-devel without upgrading libgtkhtml-devel), this usually results in a broken libgtkhtml-devel Abel
Re: [Cooker] fileutils update strange output
On Tue, 5 Mar 2002, Borsenkow Andrej wrote: install-info: warning: no entries found for `/usr/share/info/coreutilsinfobz2'; nothing deleted install-info: menu item `yes' already exists, for file `sh-utils' install-info: menu item `yes' already exists, for file `sh-utils' Yes, confirmed this for newest fileutils and sh-utils: fileutils-415-3mdk sh-utils-2011-7mdk Abel
Re: [Cooker] macro file for po files in vim
On Mon, 4 Mar 2002, Keld [iso-8859-1] Jørn Simonsen wrote: Nice to have: I see that there are a number of different supports for different file types with vim It would be nice if the po macros were also supported out of the box Pablo even advertises the po-macros on the Mandrake translation status page Did you notice the po syntax highlighting in vim? If not, please have a look at /usr/share/vim/syntax/povim Abel Best regards keld
Re: [Cooker] too much i18n...
On Mon, 4 Mar 2002, Guillaume Rousse wrote: Is it *really* necessary to translate 'frozen bubble' to 'bulle glacées' in menu :-) ? That really depends on the translator's taste or translation team's policy If you really dislike it, you can talk to the translator of menu-messages Abel
Re: [Cooker] wish: gettext 0.11
On Mon, 4 Mar 2002, Keld [iso-8859-1] Jørn Simonsen wrote: The gettext version delivered with beta3 was 01040 It would be nice if the new gettext version was delivered instead The recent version is 011something, pablo would know more keld I guess nobody would like such change in cooker The major number of libintlso has changed in gettext 011 (libintlso2) But I must say that gettext 011 has many sweet additions compared with gettext 010x Yet one problem is, it needs to search for installed JDK before it has java support :( Abel
Re: [Cooker] too much i18n...
On 4 Mar 2002, andre wrote: Is it *really* necessary to translate 'frozen bubble' to 'bulle glacées' in menu :-) ? That really depends on the translator's taste or translation team's policy If you really dislike it, you can talk to the translator of menu-messages Names shouldn be translated morte chien Again, this also depends There are many cases that, even though a brand name has an official translation, people just ignore the translation and uses English brand name On the contrary, I've seen other cases that, once a name has a translation, nobody uses the original name anymore What I want to express is, I believe no single party or no single person can have absolute decision on this It depends on the all people's practice for that particular language Abel PS Should we move this discussion to cooker-i18n?
Re: [Cooker] wish: gettext 0.11
On Mon, 4 Mar 2002, Keld [iso-8859-1] Jørn Simonsen wrote: The gettext version delivered with beta3 was 01040 It would be nice if the new gettext version was delivered instead The recent version is 011something, pablo would know more I guess nobody would like such change in cooker The major number of libintlso has changed in gettext 011 (libintlso2) Could it be optional, then? At some time we need to go forward It might be too big a chenge now After 82 is out, I guess But I must say that gettext 011 has many sweet additions compared with gettext 010x yes, that was what I was thinking of Just packaged it, under http://cleorghk/~baddog/files/Mandrake/modified-packages/gettext-011-1mdksrcrpm You may give it a try :) Abel
Re: [Cooker] too much i18n...
On 4 Mar 2002, andre wrote: There are many cases that, even though a brand name has an official translation, people just ignore the translation and uses English brand name On the contrary, I've seen other cases that, once a name has a translation, nobody uses the original name anymore What I want to express is, I believe no single party or no single person can have absolute decision on this It depends on the all people's practice for that particular language Language is what you speak and understand Not something that has been semi-arbitrary been decided what is correct and incorrect If you use the name frozen bubble than it is the name of the game Exactly -- it's the first sentence Language is what you speak and understand, but it's not you to decide what others speak and understand Hope you get my point -- it's really a bit semi-arbitrary, but once the trend comes, nobody can reverse it You can still argue that Frozen-bubble is still new, and people can still decide what it's name should be Once the original name or translated name is established, nobody can change others' usage globally PS Should we move this discussion to cooker-i18n? No So, everybody please don't blame me if the discussion should belong to elsewhere Abel
Re: [Cooker] [patch] xmms.spec
On 2 Mar 2002, Bryan Paxton wrote:
This patch changes part of the buildrequires in the latest xmms.spec.
[...]
Anyway, here's the patch, nice and small : )
-BuildRequires: libpanel_applet-devel libglib-devel libgtk+-devel libxml-devel
libvorbis-devel libogg-devel
+BuildRequires: gnome-core-devel
+BuildRequires: libglib-devel libgtk+-devel libxml-devel libvorbis-devel libogg-devel
BuildRequires: db1-devel libmikmod-devel XFree86-devel XFree86-static-libs
libesound-devel
BuildRequires: gettext ORBit-devel gnome-libs-devel
Requires: %{lib_name} = %{version}-%{release}
@@ -325,6 +327,10 @@
%endif
%changelog
+* Sat Mar 2 2002 Bryan Paxton [EMAIL PROTECTED] 1.2.6-3mdk
+- Revert from libpanel_applet0-devel to gnome-core-devel on Build Requires.
+ (Keep it Ximian GNOME friendly)
If you want to work on reverting the BR, then I guess you can
clean up the BuildRequires completely? It sounds to me that
there are so many unnecessary BuildRequires.
Abel
Re: [Cooker] Re: [CHRPM] libgtop-1.0.13-2mdk
On 23 Feb 2002, Daouda LO wrote: Name: libgtop Relocations: (not relocateable) Version : 1.0.13Vendor: MandrakeSoft Release : 2mdk Build Date: Sat Feb 23 05:12:26 2002 --=-=-= * Fri Feb 22 2002 Frederic Lepied [EMAIL PROTECTED] 1.0.13-2mdk - added BuildRequires guile-devel Sorry, seems this is wrong; guile interface for libgtop is not built at all by default. Even if guile is enabled explicitly, all it does is no more than building one more example executable. It doesn't heart. Leave it out in your build if you want (--nodeps does the trick). However, if it's for me, I'd rather get things corrected, rather than leaving them wrong (unnecessarily) :( Abel
Re: [Cooker] Re: [CHRPM] libgtop-1.0.13-2mdk
On 24 Feb 2002, Daouda LO wrote: Name: libgtop Relocations: (not relocateable) Version : 1.0.13Vendor: MandrakeSoft Release : 2mdk Build Date: Sat Feb 23 05:12:26 2002 --=-=-= * Fri Feb 22 2002 Frederic Lepied [EMAIL PROTECTED] 1.0.13-2mdk - added BuildRequires guile-devel Sorry, seems this is wrong; guile interface for libgtop is not built at all by default. Even if guile is enabled explicitly, all it does is no more than building one more example executable. I agree but for packagers/developers i think that it's better having all libs needed ( even if not mandatory) installed. You're true if there are any visible changes after adding libguile as BuildRequires; however in this case it isn't true. There's no change at all; you still need to give extra configure flags before guile support is activated. And even after guile support is manually added, it does compile one (or 2) more example binaries, which are not bundled with libgtop binary package at all. Hence, adding guile-devel is not much different than adding, say, libusb-devel. Just no effect at all. If there's some unspoken special need to add guile-devel, then I wouldn't object to this; just that I feel it's completely redundant. Abel
[Cooker] Re: [CHRPM] libgtop-1.0.13-2mdk
On Sat, 23 Feb 2002, Frederic Lepied wrote: Name: libgtop Relocations: (not relocateable) Version : 1.0.13Vendor: MandrakeSoft Release : 2mdk Build Date: Sat Feb 23 05:12:26 2002 --=-=-= * Fri Feb 22 2002 Frederic Lepied [EMAIL PROTECTED] 1.0.13-2mdk - added BuildRequires guile-devel Sorry, seems this is wrong; guile interface for libgtop is not built at all by default. Even if guile is enabled explicitly, all it does is no more than building one more example executable. Abel
Re: [Cooker] Re: [Contrib-Rpm] multi-gnome-terminal-1.3.11-2mdk
My mail was rejected before. This is a re-send. On Wed, 20 Feb 2002, Stefan van der Eijk wrote: These are the BR for the current (1.3.12-1mdk) package: [..] gdk-pixbuf-devel = 0.7.0 gtk+-devel = 1.2.5 gnome-libs-devel = 1.0.59 Unless there is a REAL need to specify the required version of these packages. I'm wondering what the effect of this is going to be. I'd prefer to put versioning only if it's REALLY required (it really can't Versioned dependencies should never never never be removed without valid reasons. Usually software maintainers have need to specify versions for required libraries, due to incompatible API or bug fixed on newer libraries. But in *THIS* particular case, I think the above versioned dependencies can be removed safely without harm, since these versions are quite old. The use of %%configure2_5x already voided the version specifications, since %%configure2_5x exists much later than the above versioned libraries. be built with another version, etc). Otherwise things are going to get overly complicated and harder to maintain when time passes. On the other I'd prefer fixing things correctly, rather than get things done easily by merely avoiding trouble. hand, if you put a version requirement on the libglade-devel and gdk-pixbuf-devel won't that then garantee that you have the correct version of gtk+-devel and gnome-libs-devel installed (because they get pulled in)? We could possibly also add version requirements on the other packages that are installed due to these BuildRequires: [. long list ..] gnome-print libimlib1-devel libgnome32-devel oaf liboaf0 libglade-bonobo0 libGConf1 libgda0 GConf libgnome-db0 libglade-gnome-db0 libglade0-devel libscrollkeeper0 scrollkeeper No version needed for all these except the few specified by software maintainer, I suppose. Abel But wouldn't that be a bit too much? (anybody any comment on this)? Should a package have no BuildRequires/Requires and just rely on the library names? See above.
Re: [Cooker] Re: [Contrib-Rpm] multi-gnome-terminal-1.3.11-2mdk
My email was rejected. This is a re-send. On Wed, 20 Feb 2002, Alexander Skwar wrote: Because common-licenses is required by basesystem. It should always be on your system, so there is really no need to list it. Well, if I remove basesystem (which is required by nothing), I could also easily remove common-licenses if every packager would follow your rule. This would cause that the user would have no copy of the GPL on his system. However, if every package requires common-licenses, this cannot happen (excluding cases where the user --nodeps a package removal). I think it'll be fine if glibc requires basesystem. One would argue that non-glibc dependent software would have no dependency against common-licenses, but that's quite sufficient for today, since more than 99% of software needs glibc, right? Abel
Re: [Cooker] MySQL-devel is a prereq, but there's no such package
On Fri, 22 Feb 2002, Schlomo Schwartz wrote: surely libmysql10-devel-3.23.47-4mdk.i586.rpm Why has this package been renamed? Even Red Hat's rawhide still calls it mysql-devel (not to mention what RPMs you can download from mysql.com). I just don't understand why these package names can't stick to the standard of what everyone else in the linux world knows them as. Makes Mandrake look like they're the oddballs. Thanks for the reply. Have a look at http://people.mandrakesoft.com/~warly/files/policy/index.html#N39 before making more comments. Abel
Re: [Cooker] Buildrequires
On Thu, 14 Feb 2002, Frederic Crozat wrote: On Thu, 14 Feb 2002 10:34:20 +0100, Alexander Skwar wrote: »Frederic Crozat« sagte am 2002-02-14 um 10:26:55 +0100 : No, you are wrong.. We are trying to switch to implicit BuildRequires.. I don't understand. What do you mean by implicit? It means all -devel package should requires all needed -devel package : libgnome32-devel should require libORBit-devel, libgtk+1.2-devel (libgtk+-1.2-devel will require libglib1.2-devel) and so on.. Therefore, a GNOME application will only need to have libgnome32-devel BuildRequires.. Yes, Alexander is exactly talking about this. Anybody care to fix all these from the ground up? I can help a bit. Abel
Re: [Cooker] configure warnings
On Thu, 14 Feb 2002, Alexander Skwar wrote: Hi! Recently, I often see this warning when configure is run: + ./configure i686-mandrake-linux-gnu --prefix=/usr --exec-prefix=/usr --bindir=/usr/bin --sbindir=/usr/sbin --sysconfdir=/etc --datadir=/usr/share --includedir=/usr/include --libdir=/usr/lib --libexecdir=/usr/lib --localstatedir=/var/lib --sharedstatedir=/usr/com --mandir=/usr/share/man --infodir=/usr/share/info configure: WARNING: you should use --build, --host, --target Could the %configure macro please be changed to also include --build, --host and --target? Use %configure2_5x instead. Abel Alexander Skwar
