Guillem Jover writes ("Re: Bug#850843: dpkg-source in stretch cannot extract
samba_3.6.5-2.dsc"):
> The problem is that "Not deleting" message. This is coming from a git
> formatted patch which deletes that file, but the sha1sum does not
> match so patch refuses to remove the file.
>
> I'm not sure there's a nice way to handle this. I think refusing to
> remove is probably the correct answer here. But patch didn't use to do
> that, so making this non-fatal could be perhaps fine too. I've not
> tested but an option would be what I've mentioned before, which is to
> handle git formatted patches directly by dpkg-source, which means the
> file would be already removed, and then patch might not fail?
I don't understand all of this, but my position is that dpkg-source in
stretch should extract this package to the same source tree as
dpkg-source in wheezy did.
If the problem is that wheezy's patch accepted this patch and did one
some particular thing with it, but sids patch rejects it, then what is
needed is an option to patch to disable the additional check. Having
dpkg-source spot particular edge cases and work around them is a
second-best solution.
This kind of thing is an inevitable consequence of 1. the way that the
implementation of `3.0 (quilt)' simply invokes patch on whatever
appears in the source package, without any kind of further checking;
combined with 2. patch not having a proper (formal) approach to
protocol compatibility.
Ian.
--
Ian Jackson <ijack...@chiark.greenend.org.uk> These opinions are my own.
If I emailed you from an address @fyvzl.net or @evade.org.uk, that is
a private address which bypasses my fierce spamfilter.
