Saurabh Vyas wrote:
> New webrev : http://cr.opensolaris.org/~saurabhv/fix-15646-rev-4/
publish.py:
- Please be careful not to introduce lines indented by tabs. Use spaces
instead.
- You appear to have broken timestamp handling -- the "else" on line 412
should be matched to the "for" on line 409, not the "if" on line 407.
- line 418: what does this mean?
- line 451: You should be able to just error(e, cmd="import"); there's no
sense in duplicating the message already provided by the bundle code.
- line 452, 456: we should probably abandon the transaction in these
cases, too.
- line 455: error(e) here, too -- the default message is good enough, I
think.
- line 457: Let other EnvironmentErrors get raised, unless you know
specifically what to do with them. "Unknown Error" is not useful in
figuring out what went wrong.
> Here we want not to raise an error & continue in make_bundle() to check
> for the next bundle type also.
Yes, you need to do that.
> v...@opensolaris:~# pkgsend generate hello/MyDir.pkg
> unknown path=pkginfo
> unknown path=pkgmap
> unknown path=install
> unknown path=reloc
> unknown path=reloc/mytest
> pkgsend: generate: Unknown Error
That doesn't look good.
> >>This is where I need more inputs / ideas about how to handle other cases
> >
> >The cases you need to cover are:
> >
> > - pathname is a tarfile
> > - pathname is a generic directory
> > - pathname is an SVr4 package in directory format
> > - pathname is an SVr4 package in datastream format
> > - pathname is something else
> > - pathname doesn't exist
> I tried to cover these cases, please let me know if this attempt
> looks good (or need further improvement / modification)
You only added a "doesn't exist" testcase. That should probably be its own
testcase, rather than another item in an existing one. "Something else"
should be added, too.
Danek
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
