link on your site
Dear owner of http://qa.debian.org, I am the advertising representative of a number of leading online companies. They operate in various fields, that include the gaming industry, sports activities, international flower delivery and various leisure activity fields . I was impressed by your internet site http://qa.debian.org and would like to discuss the possibility of you hosting advertising space for my clients on your site. We are interested in various options that include textural advertisements, and Banner space. If you are interested please contact me and we can discuss all the possibilities in detail. Please contact me for farther details. Best regards, Jim Lawrence Client side acquisition http://www.luckychip-media.info This is NOT SPAM -- this is a one-time link request. We have NO INTENTION to email you again. You can also reply to this email with REMOVE in the subject line to make sure we'll NEVER send you any more e-mails in the future.
Packages not making it into testing
I am not the maintainer, but... zope-pythonmethod uploaded 180 days ago, out of date by 170 zope-pythonmethod is essentially superceded by Script (python), which entered zope as a non-optional component in release 2.3. This is a good candidate for removal. (There are some minor API changes, but pythonmethods are now receiving no upstream support.) Jim Penny
Re: List of bugs that *must* be fixed before releasing Hamm
Brian White <[EMAIL PROTECTED]> writes: > 20980: kaffe- kaffe: kaffe depends on jdk-common [0] (Vincent > Renardias <[EMAIL PROTECTED]>) I had filed a bug against ftp.debian.org (#20165) to take it out of hamm and orphan it several weeks ago. It's beyond my comprehension to understand why it takes so long to resolve simple fixes to the ftp archive. I will be uploaded kaffe and japhar packages soon (for slink). Cheers, - Jim pgpEqga43n2Ef.pgp Description: PGP signature
Bug#25573: packaging-manual: Small typo
Package: packaging-manual Version: 2.4.1.1 Severity: wishlist >From the packaging manual: 12.2.4 How to use dpkg-shlibdeps and the shlibs files? __ 12.2.4.1 If your package doesn't provide a shared library Put a call to dpkg-shlibs into your debian/rules file. If your package ^^^ Shouldn't that be "dpkg-shlibdeps"? Cheers, - Jim -- System Information Debian Release: 2.0 Kernel Version: Linux fleming 2.0.34 #1 Wed Jul 15 23:36:01 PDT 1998 i586 unknown
Re: Unexpected behaviour of tsort
Thanks for the report. This will be fixed in the textutils, eventually. I don't have time right now. I'll gladly consider patches. Torsten Landschoff <[EMAIL PROTECTED]> writes: | Hi textutils developers, | | I am writing to you because we (Debian) got a release critical bug | against our textutils package. It seems that tsort does not behave as | most people would expect. | | Note: I am not the maintainer of that package but a member of the | Quality Assurance team. | | As the behaviour of tsort conforms to the documentation this is not | really a bug. Problem is that the other unices I tried this on report | the error but nonetheless they try to generate senseful output on | stdout. | | This is from the manpage of tsort on SuSE Linux: | | If the graph contains a cycle (and therefore cannot be properly | sorted), one of the arcs in the cycle is ignored and the sort | continues. Cycles are reported on standard error. | | SuSE seems to use the BSD version of tsort. On Solaris there is no | reference to this problem in the manpage but output is generated anyway. | | As Debian tries to stay compatible with other Linux versions we are | thinking of replacing tsort with the BSD version but as we ship | GNU/Linux we would like a new GNU version which fixes this better. | | So please consider changing the behaviour of tsort. If you lack the | manpower I could help with the patch but I am out of reach starting this | Friday for a week. | | Here is the original bugreport: | | > Date: Fri, 27 Aug 1999 12:53:44 +0200 | > From: "J.H.M. Dassen (Ray)" <[EMAIL PROTECTED]> | > To: [EMAIL PROTECTED] | > Subject: textutils: tsort behaves unexpectedly | > | > Package: textutils | > Version: 2.0-2 | > Severity: important | > | > [Severity motiviation: this breaks behaviour expected by some source builds, | > e.g. of netkit-telnet-ssl.] | > | > Consider the following data: | > auth.o auth.o | > misc.o misc.o | > ssl.o ssl.o | > sslapp.o sslapp.o | > ssl.o misc.o | > ssl.o auth.o | > auth.o misc.o | > ssl.o auth.o | > misc.o auth.o | > auth.o ssl.o | > sslapp.o ssl.o | > auth.o misc.o | > ssl.o sslapp.o | > | > Running tsort on this results in these errors on stderr | > [this is correct: a true topological sort does not exists when there are | > cycles] | > tsort: -: input contains a loop: | > | > tsort: auth.o | > tsort: misc.o | > and no output no stdout. | > | > In 2.1, tsort was built from bsd sources rather than the GNU textutils. That | > version also gives warnings on stderr | > tsort: cycle in data | > tsort: ssl.o | > tsort: sslapp.o | > tsort: cycle in data | > tsort: misc.o | > tsort: auth.o | > tsort: ssl.o | > tsort: cycle in data | > tsort: auth.o | > tsort: misc.o | > but does produce output on stdout | > [namely a fewest-cycles sort of the input] | > sslapp.o | > ssl.o | > misc.o | > auth.o | > | > This behaviour is expected by a number of source builds, which contain an | > optimisation in library generation: | > ar [some flags] `lorder *.o | tsort` | > | > Please make GNU's tsort behave like the BSD one, or coordinate with the | > bsdmainutils maintainer to switch back to the BSD version. | > | > -- System Information | > Debian Release: potato | > Kernel Version: Linux jcn206a 2.2.12-int2.2.11.2 #3 Fri Aug 27 | > 10:02:05 CEST 19 99 i586 unknown | > | > Versions of the packages textutils depends on: | > ii libc6 2.1.2-0pre11 GNU C Library: Shared libraries and timezone | | Torsten Landschoff for the Debian QA Group
Re: Release problems with slang?
On Tue, Oct 05, 1999 at 01:34:51AM +0100, Edward Betts wrote: > Christian Hammers <[EMAIL PROTECTED]> wrote: > > Format: 1.6 > > Date: Sun, 3 Oct 1999 13:17:33 +0200 > > Source: jed > > Binary: jed-sl jed-common xjed jed > > Architecture: source i386 all > > Version: 0.99.9-1 > > Distribution: unstable > > Urgency: low > > Maintainer: Christian Hammers <[EMAIL PROTECTED]> > > Description: > > jed- Editor for programmers. (textmode version) > > jed-common - Byte compiled SLang runtime files for jed and xjed. > > jed-sl - Sources of Slang runtime files for jed and xjed > > xjed - Editor for programmers. (x11 version) > > Changes: > > jed (0.99.9-1) unstable; urgency=low > > . > >* New upstream version ! Fixes a lot of bugs. > >* This one's static linked against the slang1_1.3.8 since the maintainer > > of > > slang didn't upload the latest version within the last two month :-( > > Static linking, because the maintainer of another library has not updated > it? This is bad. Have you contacted the maintainer? Is it time for Quality > Assurance to step in? Anybody want to do a NMU of slang? I have been talking to Christian. The new packages of slang are ready to go. I have been waiting for other people to upgrade their packages to slang1_1.2.2-3 which fixes a dependancy problem. Most people have updated their packages but there were a few more that had promised to upload soon. As soon as 1.3.8 is uploaded those packages will break (and since a couple of them and slang are all default packages this means most upgrades of potato will break) I announced most of this before. I don't mean to deflect all the criticism as I have been on and off on holiday and haven't been bugging the other people as much as I should. I will do an upload today. Jim -- Jim Mintha Email: [EMAIL PROTECTED] System Administrator Work: +31 20 525-4919 Informatiseringscentrum Home: +31 20 662-3892 University of Amsterdam Debian GNU/Linux: [EMAIL PROTECTED] _There are always Possibilities_ http://jim.ultralinux.org
Re: Release problems with slang?
On Tue, Oct 05, 1999 at 05:49:26AM -0400, Raul Miller wrote: > On Tue, Oct 05, 1999 at 09:33:05AM +0200, Jim Mintha wrote: > > I have been talking to Christian. The new packages of slang are ready > > to go. I have been waiting for other people to upgrade their packages > > to slang1_1.2.2-3 which fixes a dependancy problem. Most people have > > updated their packages but there were a few more that had promised to > > upload soon. As soon as 1.3.8 is uploaded those packages will break > > (and since a couple of them and slang are all default packages this > > means most upgrades of potato will break) I announced most of this > > before. > > Does this mean that the new slang package should be named slang1.3? > > If backwards compatability isn't maintained, what's the rational for > keeping the same package name? No there is backward compatibility, but the earlier versions of slang 1.2.2 had a shlibs.depend that created a dependency on slang1 << 1.3 which was wrong. So there were two options: 1) make a slang1.3 package (and a slang1.4 package etc) even though they were compatible 2) Uploaded a slang 1.2.x package that fixed shlibs.depend and get the slang dependent packages to recompile. After discussing it on the mailing lists I went with 2, and most packages have now been updated. Jim -- Jim Mintha Email: [EMAIL PROTECTED] System Administrator Work: +31 20 525-4919 Informatiseringscentrum Home: +31 20 662-3892 University of Amsterdam Debian GNU/Linux: [EMAIL PROTECTED] _There are always Possibilities_ http://jim.ultralinux.org
Bug#53716: Fails on machines where chars are unsigned.
Package: xpuzzles
Version: 5.4.4-4
There are about 67 occurences of the line...
while ((c = getc(fp)) != EOF && c != SYMBOL);
... where c is a char.
On architectures where char is unsigned (definately ARM and maybe PPC)
this comparison fails because the int from getc has been reduced to
an unsigned char which will never equal EOF.
--
Jim Studt, President
The Federated Software Group, Inc.
Bug#53716: Fails on machines where chars are unsigned.
>Hamish...
> Why is char unsigned on these platforms?
I believe there is a serious performance difference on these platforms.
Its legal ANSI behavior, I guess someone was bound to choose the
alternative.
--
Jim Studt, President
The Federated Software Group, Inc.
