On 03/25/11 08:41 AM, Bart Smaalders wrote:
On 03/24/11 14:38, Shawn Walker wrote:
src/util/publish/pkgmerge.py:
line 648: This does mean that if they say
'pkg.merge.blend=variant.arch' things won't work. Do
we care about 'variant.arch' vs 'arch'? Any way to
make that clearer in man page perhaps?
I'll fix this to work either way; easier than explaining :-).
Thanks.
src/tests/cli/t_pkgmerge.py:
line 275-282: I thought you didn't need the sleep() anymore?
I don't.. but depending on where the timestamp changes, the
order of the variants in the output switches :-(. Until I
convert this testcase to use pkgdiff, it needs to be byte-for-byte
compatible from run to run.
Doh; sounds reasonable.
line 601: what does it mean if an action doesn't have a source? seems
like this should assert; we know the action has a payload, so it
should have a source
Note that this code replaces the duplicate check in the previous
version; we will only record a retrieval for the first action that uses
a particular hash. If (due to logic error) we fail to retrieve all
the hashes, the publish step _will_ tip over, as I discovered when
I got it wrong the first time.
Ah, ok.
-Shawn
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
