[arch-dev-public] ncurses rebuilds in testing
Hi all, I have pushed the ncurses wide-character library rebuilds for i686 to the [testing] repo. There are bound to be packages I missed so it would be good if people run lddd and let me know if further rebuilds are needed. Note libncurses.so and libcurses.so might appear to be in the package but they are not real libraries... (run cat on them to see). x86_64 builds are mostly done and will be uploaded tomorrow. Allan
On Sun, 17 Aug 2008, Allan McRae wrote:
Hi all,
I have pushed the ncurses wide-character library rebuilds for i686 to the [testing] repo. There are bound to be packages I missed so it would be good if people run lddd and let me know if further rebuilds are needed. Note libncurses.so and libcurses.so might appear to be in the package but they are not real libraries... (run cat on them to see).
missing rebuilds from extra: lftp xine-ui cmake xemacs dopewars libqalculate libgda gnuplot parted timidity++ abiword-plugins guile I'm forwarding this to the AUR ML as this is a major rebuild. If you use the testing repo and packages in community, be careful. Several community packages (and also packages built from custom/unsupported PKGBUILD) will break as they will need to be rebuild against the ncurses in testing. The community packages will be rebuilt once ncurses will be moved to core. Meanwhile, either stop using the testing repo or rebuild the packages yourself. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
Eric Belanger wrote:
missing rebuilds from extra:
lftp xine-ui cmake xemacs dopewars libqalculate libgda gnuplot parted timidity++ abiword-plugins guile
Done for i686. x86_64 rebuilds for these will occur either tomorrow or the next day (depending on how well I play poker tomorrow night...). Anyone else is free to build them for me :) Allan
Allan McRae wrote:
Hi all,
I have pushed the ncurses wide-character library rebuilds for i686 to the [testing] repo. There are bound to be packages I missed so it would be good if people run lddd and let me know if further rebuilds are needed. Note libncurses.so and libcurses.so might appear to be in the package but they are not real libraries... (run cat on them to see).
x86_64 builds are mostly done and will be uploaded tomorrow.
x86_64 builds are now in [testing]. I had a couple of x86_64 build failures that I didn't find easy fixes for: iptraf-3.0.0-2 yabasic-2.763-3 And there are a few packages which need rebuilt in which the current source line is bad: achessclock giftcurs hexcurse (md5sum failure. Url seems to be a squatter site..) ne And this can be fixed when it gets updated... archboot (currently dependencies can not be satisfied) I will create an internal dev TODO list to keep track of these and the rebuilds Eric pointed out. Allan
On Mon, 18 Aug 2008, Allan McRae wrote:
Allan McRae wrote:
Hi all,
I have pushed the ncurses wide-character library rebuilds for i686 to the [testing] repo. There are bound to be packages I missed so it would be good if people run lddd and let me know if further rebuilds are needed. Note libncurses.so and libcurses.so might appear to be in the package but they are not real libraries... (run cat on them to see).
x86_64 builds are mostly done and will be uploaded tomorrow.
x86_64 builds are now in [testing].
I had a couple of x86_64 build failures that I didn't find easy fixes for: iptraf-3.0.0-2 yabasic-2.763-3
And there are a few packages which need rebuilt in which the current source line is bad: achessclock
Both home page and source have disappeared. The first pages returned by Google only point to the Arch website or Arch mirrors. IMO, we should just remove it.
giftcurs hexcurse (md5sum failure. Url seems to be a squatter site..) Both done.
ne
Google had cached the page a couple of days ago. Probably temporary network/hardware problems. It should be checked again in a few days. BTW, the package in extra is out-of date.
And this can be fixed when it gets updated... archboot (currently dependencies can not be satisfied)
I will create an internal dev TODO list to keep track of these and the rebuilds Eric pointed out.
Allan
-- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
Eric Belanger wrote:
On Mon, 18 Aug 2008, Allan McRae wrote:
And there are a few packages which need rebuilt in which the current source line is bad: achessclock
Both home page and source have disappeared. The first pages returned by Google only point to the Arch website or Arch mirrors. IMO, we should just remove it.
The source couldn't be found when Arch64 was getting built (see http://wiki.archlinux.org/index.php/Arch64_ToDoS). It will definitely be broken with the ncurses change. Any objections to removing this? Allan
On Tue, Aug 19, 2008 at 10:25 AM, Allan McRae <allan@archlinux.org> wrote:
The source couldn't be found when Arch64 was getting built (see http://wiki.archlinux.org/index.php/Arch64_ToDoS). It will definitely be broken with the ncurses change. Any objections to removing this?
Also, the achessclock home page disappeared. I think you can remove it. -- Hugo
Eric Belanger wrote:
On Mon, 18 Aug 2008, Allan McRae wrote:
ne
Google had cached the page a couple of days ago. Probably temporary network/hardware problems. It should be checked again in a few days. BTW, the package in extra is out-of date.
I found a mirrored tarball for the version currently in the repos so have rebuilt that for the time being.
Am Sun, 17 Aug 2008 21:37:18 +1000 schrieb Allan McRae <allan@archlinux.org>:
Hi all,
I have pushed the ncurses wide-character library rebuilds for i686 to the [testing] repo. There are bound to be packages I missed so it would be good if people run lddd and let me know if further rebuilds are needed. Note libncurses.so and libcurses.so might appear to be in the package but they are not real libraries... (run cat on them to see).
x86_64 builds are mostly done and will be uploaded tomorrow.
Allan
While you are at it why don't you add all the hidden ncurses dependencies to the PKGBUILDs in a versionized form ('ncursed>=5.6-7')? Would make the next rebuild session much easier. I still wonder what you have done to my lftp that you need a minor sub pkgrel number for i686... What are your plans for moving packages to core/extra? BTW: congrats for the August commit award :D -Andy
Andreas Radke wrote:
While you are at it why don't you add all the hidden ncurses dependencies to the PKGBUILDs in a versionized form ('ncursed>=5.6-7')? Would make the next rebuild session much easier.
The new packages will run with the old ncurses so it is not really a versioned dependency in that direction. I have made an annotated rebuild list that I will add to the DeveloperWiki for the next ncurses rebuild.
I still wonder what you have done to my lftp that you need a minor sub pkgrel number for i686...
Shhh, I might have built the i686 package in the wrong chroot... :)
What are your plans for moving packages to core/extra?
I will call for sign-offs to core packages once rebuilds are finished. Then, if no-one reports anything gone wrong, this can all move to the main repos. Allan
On Tue, 19 Aug 2008, Allan McRae wrote:
Andreas Radke wrote:
While you are at it why don't you add all the hidden ncurses dependencies to the PKGBUILDs in a versionized form ('ncursed>=5.6-7')? Would make the next rebuild session much easier.
The new packages will run with the old ncurses so it is not really a versioned dependency in that direction. I have made an annotated rebuild list that I will add to the DeveloperWiki for the next ncurses rebuild.
I'm currently doing some rebuild for x86_64. I'm adding an explicit ncurses dependency.
I still wonder what you have done to my lftp that you need a minor sub pkgrel number for i686...
Shhh, I might have built the i686 package in the wrong chroot... :)
What are your plans for moving packages to core/extra?
I will call for sign-offs to core packages once rebuilds are finished. Then, if no-one reports anything gone wrong, this can all move to the main repos.
Allan
-- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
participants (4)
-
Allan McRae
-
Andreas Radke
-
Eric Belanger
-
Hugo Doria