Re: [arch-general] should gawk be in group base-devel?
Hi, Jan de Groot wrote: For the base/base-devel discussion: when building packages, base-devel and base should be installed. Looking at debian-alike distributions, I don't see packages makedepending on g++ either when c++ code needs to be compiled. The gcc dependency within base-devel would cover the use of g++. Perhaps I muddied the waters by suggesting that a user could choose to not install gawk. Maybe the real question is: What is the purpose of base-devel? 1) To specify the standalone packages required to build base packages or 2) To specify a recommended set of packages required for package building in general If the answer is (1) and something in base requires gawk to build then I suggest gawk should be a dependency of base-devel. If the answer is (2) then I guess I can live with the implied requirement that gawk is installed as part of base _always_. I suppose I'm being a bit pedantic here but, from my engineering background, I don't really like assumptions and would prefer explicit requirement. I don't have the exact quote to hand but it has been said that 90% of programming errors arise from false assumptions. I guess that's why formal design methodologies evolved - but that's another story ;-) Regards, Neil Darlow
[arch-general] should gawk be in group base-devel?
Hi, It is often quoted that, for package building purposes, it is assumed a user has packages in group base-devel installed. gawk is part of base but a user could deselect it within the Arch Installer. Does it make sense to add gawk to group base-devel so that its presence is mandated by the previous assumption? Should I raise a feature request for this? Regards, Neil Darlow
Re: [arch-general] Postgres failure after system upgrade
Hi, richard terry wrote: I wonder if some kind soul could give me some help in tracking down the problem. I've tried looking at the below mentioned files and can't see anything wrong . Perhaps something got overwritten? PostgreSQL requires a dump/restore for an upgrade from 8.2.x to 8.3.x. This also entails saving a copy of pg_hba.conf and postgresql.conf and merging any changes you have made into the new versions. It is normal for the PostgreSQL server to refuse to start if it detects that its database structure is outdated. Regards, Neil Darlow
Re: [arch-general] [arch-dev-public] policy on desktop files?
Hi, Xavier wrote: Hopefully this should reduce the load of arch packagers and move it to arch community and upstream. There is also the situation that upstream may not provide a .desktop file on the basis of: 1) The belief that distributions should decide which menu the item should appear on 2) The desire not to get involved with providing I18N for the .desktop contents If upstream is forced to provide a .desktop file which requires modification of e.g. category or provision of additional I18N content then we're now in the realms of patching .desktop files too. I agree that any .desktop file we create should be submitted upstream, at least for their consideration. Regards, Neil Darlow
Re: [arch-general] alternate dependencies?
Hi, Xavier wrote: That's what provisions are for. Wouln't that require that e.g. tetex and texlive have something like? provides=( tex ) In practice, how many packages include such a generic provides entry? From what I've seen most packages' depends rely solely on the package name. I think there will always be a case where an alternate dependency would better be specified by the package name. Regards, Neil Darlow
[arch-general] graphviz needs to come out of testing for gnome-2.22 upgrade
Hi, graphviz-2.18 is a dependency of anjuta and currently it's stuck in testing. Looking at its repo information it appears to be maintainerless. Could someone please progress it out of testing. Thanks. Regards, Neil Darlow
[arch-general] HP C3180 connected using usb no longer works
Hi, My HP C3180, connected using usb, has recently stopped working. I'm not sure if it's related to recent changes in udev. lsusb -vv gives: Bus 002 Device 007: ID 03f0:5611 Hewlett-Packard PhotoSmart C3180 ls -l /dev/usbdev2.7 gives: crw-rw-rw- 1 root lp 189, 134 2008-04-04 19:40 /dev/usbdev2.7 My groups setting is: sys tty lp wheel network video audio optical storage scanner camera power users neil I am using udev-119-1, foomatic-db-hpijs 20080211-1 and cups-1.3.7-1 on i686. Any suggestions as to where to look for debugging this would be appreciated. I recall having to do this once before. Regards, Neil Darlow
Re: [arch-general] Something screwy with colour terminal output at shutdown
Hi Arvid, Arvid Ephraim Picciani wrote: could be related to that. http://bugs.archlinux.org/task/1186 Not sure it is. Colours at startup are OK. It's just shutdown that's messed-up. Regards, Neil Darlow
Re: [arch-general] cannot mount dvd or camera
Hi Arthur, Arthur wrote: I have hal,dbus fam in my DAEMONS and I have added myself to the optical and storage groups. I have also logged off and on and tried restarting, but I still get this error. Not sure about DVD mounting, storage should handle that. You might want to add yourself to the camera and (for good measure) video groups for the other problem. Regards, Neil Darlow
[arch-general] What should be done with octave?
Hi, Arch has octave-2.1.73 in extra. The current stable version of octave is 3.0.0 but it is not backwards compatible with the 2.1.x series. The gcc-4.2.3 migration has broken octave due to a versioned dependency on gcc-fortran. Also, the PKGBUILD has no license or md5sum entries. The octave-2.1.73 PKGBUILD will need fixing but should it be updated to 3.0.0 and, if so, should a legacy octave2 package be created? I'm looking for opinions before I file any bug report. Regards, Neil Darlow
Re: [arch-general] Name change on crypto modules for kernel26-2.6.24?
Hi, To clarify. Neil Darlow wrote: Too me, it looks like these modules have undergone a name change in 2.6.24. I think the required modules are actually being loaded but cryptsetup complains about missing modules. It appears the following name changes have occurred: * padlock_aes.ko -- padlock-aes.ko * padlock_sha.ko -- padlock-sha.ko Regards, Neil Darlow
[arch-general] Who is confused, pacman or me?
Hi, I did a pacman -Syu this morning and it warned me that: local etracer-0.35-6 is newer than etracer-0.4-1 I'm not sure how that came about. Is this a pacman or pkgbuild problem? My system is i686. Regards, Neil Darlow
Re: [arch-general] Who is confused, pacman or me?
Hi, Xavier wrote: But, when you -Su, does it still want to install etracer 0.4 ? It didn't want to update etracer. I had to manually update the package myself. This versioning confuses pacman (and it isn't something that can be fixed), but we have the force options for these cases. And the force option was used for this package, as you can see there : http://cvs.archlinux.org/cgi-bin/viewcvs.cgi/games/etracer/PKGBUILD?cvsroot=AURonly_with_tag=CURRENT No forced update was done. Just a warning that the local version was newer was issued. Regards, Neil Darlow
[arch-general] Qt3 path setting, is it right?
Hi, I've been building bouml (from AUR) successfully until Qt4 arrived. Examining the build failure it appears that although QTDIR is correctly set for Qt3, the placement of QTDIR/bin at the end of PATH in qt3.sh results in Qt4's qmake taking precedence over that of Qt3. For bouml's build mechanism (no ./configure involved) this leads to qmake-generated files containing resource paths of Qt4. I've managed to get the build to work by placing QTDIR/bin at the head of PATH, using a sed command, but it's seems hackish to me. Is the sourcing of qt3.sh from /etc/profile.d a sensible idea? Would it be better provided as part of the qt3 package and sourced into the PKGBUILD's environment directly? And, surely, placing QTDIR/bin at the end of PATH is not a good idea? Regards, Neil Darlow
