<cruxbot>
[contrib.git/3.7]: remind: update to 04.02.03
<cruxbot>
[contrib.git/3.7]: tig: update to 2.5.8
<cruxbot>
[opt.git/3.7]: mariadb: update to 10.10.3
<cruxbot>
[opt.git/3.7]: dialog: update to 1.3-20230209
<cruxbot>
[opt.git/3.7]: gnuplot: update to 5.4.6
ppetrov^ has joined #crux
farkuhar has joined #crux
elderK has quit [Quit: Connection closed for inactivity]
<farkuhar>
I just noticed, some of my ports installed by an older pkgadd (before the 5.40.10 pkgutils bump) were recorded in /var/lib/pkg/db with a mangled footprint (usr/lib64 instead of usr/lib), despite pkginfo -f showing the desired prefix usr/lib. The timestamps of the latest edits to pkgadd.cc and pkginfo.cc cannot explain the change in behaviour, though.
<farkuhar>
When I 'pkgadd -u -f' the newly-built packages, the correct prefix /usr/lib is written to the database, but packages installed four months ago had those same files listed under /usr/lib64, even though nothing has changed in the build() function.
<farkuhar>
actually it doesn't have to be a newly-built package. Force-reinstalling the old package with the latest pkgadd will correct the database: http://sprunge.us/I2T90u
<ppetrov^>
farkuhar, does it matter on crux whether libs go to /usr/lib or /usr/lib64? Otherwise, I am also curious why the place changed even though the port did not
<farkuhar>
ppetrov^, it doesn't matter during runtime since /usr/lib64 is just a symlink to /usr/lib. But when upgrading the package you'll hit the error "listed file(s) already installed. Use -f to force."
<farkuhar>
and in this case it looks like the port did in fact change, but the timestamp of the relevant commit was nearly a month after the mtime of the built package, so I didn't catch it on a cursory scan of the git log.
<ppetrov^>
farkuhar, "listed file(s) already installed" refers to the lib64 -> lib symlink?
<farkuhar>
ppetrov^: no, it's the pkgadd error message when the new package contains /usr/lib/libX.so in its footprint, which is considered in conflict with /usr/lib64/libX.so in the old package's footprint. pkgadd is not smart enough to recognize that the file location is ultimately the same in both packages.
<ppetrov^>
thanks for the explanation
<farkuhar>
by the way, did you ever get around to trying to build a CRUX iso for i686?
<ppetrov^>
me?
<ppetrov^>
i have never even attempted
<ppetrov^>
and I douybt it's worth it to do that for an old netbook that I do not really use
<farkuhar>
yes, I thought you were asking about that possibility in recent days
<ppetrov^>
yes, i was just curious
<ppetrov^>
if it were available, i would've tried probably