Jesse Keating
d507bed1b2
Use universal newlines in getting process output (Trac #44 )
2007-06-11 14:59:18 -04:00
Jesse Keating
c8e8f436b5
Remove leftover size comparison code snippit.
2007-05-25 18:50:37 -04:00
Jesse Keating
69bf86d8a6
Better handle CD size variable (float vs string and comments)
2007-05-25 14:28:38 -04:00
Jesse Keating
804be3f018
Use a packages checksum to validate cached download
...
Using a package size is not good enough, fails to see
a changed gpg sig.
2007-05-24 02:06:25 -04:00
Jesse Keating
176f7ea5ef
pkgobj.packagesize is now an int, not a str.
2007-05-15 13:29:30 -04:00
Jesse Keating
59ce91800b
- Apply sparc patches (spot)
2007-05-15 13:06:34 -04:00
Jesse Keating
df3900d944
Don't quote ISO label, not running mkisofs in shell
2007-05-15 11:37:37 -04:00
Jesse Keating
0a0bbe6688
Fix source iso building
2007-03-26 10:39:46 -04:00
Jesse Keating
d2c9c76a55
Rescue fix from Prarit
2007-03-22 17:48:41 -04:00
Jesse Keating
ba6c7b1ae8
Specify flavor for buildinstall
2007-03-22 14:01:26 -04:00
Jesse Keating
c0ab3ad59c
- Fix a logic flaw in the depsolving loop
2007-03-16 13:31:31 -04:00
Jesse Keating
aa0f72f352
Remove extra whitespace. (jwilson)
2007-03-15 16:40:01 -04:00
Jesse Keating
003ceaa7be
- Change how depsolving is done
...
- Get all potential matches for a dep, instead of our 'best'
our 'best' may not be the same as install time best.
- Remove anaconda code, use direct yum functions to get deps
- Use a True/False flag to depsolve instead of iterating over
a dict.
- Log what packages are being added for which reasons.
- Also fix a think-o with the false negative unmatched workaround.
2007-03-15 16:25:57 -04:00
Jesse Keating
23e40be9ee
- Work around false positive 'unmatched' packages (globs are fun)
2007-03-15 14:25:24 -04:00
Jesse Keating
997c6bdf15
Get all potential matches for deps, let install time figure
...
the best one to use.
Remove code stolen from anaconda as it isnt needed anymore.
This may bloat the compose, but we need the multiarch stuff
and potential matches so that install time can figure out what
to use.
2007-03-15 14:17:53 -04:00
Jesse Keating
a0562d7c20
- Use yum's built in conditional handling for things from comps
...
- Do excludes before group handling.
2007-03-15 13:35:05 -04:00
Jesse Keating
938113eebe
Use yum's built in exclude handling
2007-03-15 13:29:57 -04:00
Jesse Keating
e2c43c0a93
- Do things faster/smarter if we've only asked for one disc
2007-03-14 17:57:09 -04:00
Jesse Keating
876cb98ba7
- correctly capture errors from subprocess
2007-03-13 22:16:39 -04:00
Jesse Keating
1c45ac7daf
- log the rpm2cpio stuff for release notes
2007-03-13 20:36:48 -04:00
Jesse Keating
dbc2db0a8b
More clean up with using subprocess
...
Have a function we call over and over
Log the output, trace on stderr
2007-03-08 10:58:13 -05:00
Jesse Keating
069f8f98a9
Fix some typos and point at the right comps file
2007-03-07 21:33:45 -05:00
Jesse Keating
9986ab582a
- Call createrepo ourselves for the tree, not buildinstall's job
2007-03-07 17:53:53 -05:00
Jesse Keating
4e8d70bf9f
- Convert from commands to subprocess for things we call out
...
Next I will fix up the logging
2007-03-07 14:42:33 -05:00
Jesse Keating
7e155b84e1
- Make the list we hand off to yum to search for as unique as we can
2007-03-02 11:33:01 -05:00
Jesse Keating
be44e2527d
Fix various typos, work around a yum api bug, strip the manifest lines right, log more output
2007-03-02 10:24:03 -05:00
Jesse Keating
45a4b62308
First attempt at supporting kickstart package syntax in manifest files
2007-03-01 23:06:41 -05:00
Jesse Keating
2f3a891a5c
Fix gathering of srpms (thanks skvidal)
2007-02-26 13:06:31 -05:00
jkeating@reducto.boston.redhat.com
f60b531ee9
Don't use TMPDIR with buildinstall for now
2007-02-21 16:12:31 -05:00
jkeating@reducto.boston.redhat.com
1e9d5392e6
Make cdsize a config option
2007-02-16 13:59:01 -05:00
jkeating@reducto.boston.redhat.com
fba101d3fc
Put yum temp stuff in workdir
2007-02-16 09:37:52 -05:00
jkeating@reducto.boston.redhat.com
07b27a464f
Use TMPDIR when calling anaconda-runtime stuff, get all output in the tree.
2007-02-16 08:45:42 -05:00
jkeating@reducto.boston.redhat.com
6ebb69628f
Fix indent
2007-02-15 16:25:52 -05:00
jkeating@reducto.boston.redhat.com
090eebdcf8
More logging fixes; add pydoc info to functions
2007-02-15 14:17:39 -05:00
jkeating@localhost.localdomain
ddd183386c
Logging patch from jbowes
2007-02-14 17:33:07 -05:00
jkeating@reducto.boston.redhat.com
7883986e1d
Style fixes
2007-02-13 17:20:01 -05:00
wwoods@kraid.test.redhat.com
8bb0495b63
fix up .composeinfo/mkrelative stuff
2007-02-13 13:23:00 -05:00
wwoods@test158.test.redhat.com
e2e81e481c
forgot that relativize is a class method - also change name to avoid spelling mistakes
2007-02-12 16:30:26 -05:00
wwoods@test158.test.redhat.com
bc1a73ace4
oops, forgot to check before removing
2007-02-12 16:29:19 -05:00
wwoods@test158.test.redhat.com
6e9172720a
avoid traceback after interrupted download
2007-02-12 15:57:58 -05:00
wwoods@metroid.rdu.redhat.com
bbe50383dc
oops, left in half-baked code
2007-02-12 13:08:28 -05:00
wwoods@metroid.rdu.redhat.com
289fc2865e
write composeinfo after writing iso/tree
2007-02-09 16:13:52 -05:00
jkeating@reducto.boston.redhat.com
8b1b9b36eb
Add support for globbing in manifest
...
Redo how we find packages making use of a better yum call
Add warning output for unmatched packages in the manifest
2007-02-08 20:25:41 -05:00
jkeating@reducto.boston.redhat.com
5749bcd8da
Minor speedup to short circuit depsolving
2007-02-06 22:53:43 -05:00
jkeating@reducto.boston.redhat.com
2c5642e3d7
- Don't use 'returnSimple' anymore, deprecated in yum api
2007-02-06 13:54:06 -05:00
jkeating@reducto.boston.redhat.com
8a7700a4a1
yum bestPackagesFromList takes an arch argument. Fixes ppc64 bug
2007-02-06 13:51:53 -05:00
jkeating@reducto.boston.redhat.com
78d7d7c613
Think-o.
2007-02-05 14:50:45 -05:00
jkeating@reducto.boston.redhat.com
130e7576fa
Make isodir an object (wwoods)
2007-02-03 16:19:41 -05:00
jkeating@reducto.boston.redhat.com
50bc32037e
Make bugurl optional
2007-02-03 15:53:47 -05:00
jkeating@reducto.boston.redhat.com
cec173adfe
Hey guess what? implantmd5 alters the iso, so sha1sum must be done after that.
2007-01-30 10:26:48 -05:00
