[PATH, svnmerge] Fix "svnmerge init" on the GCC repository branches

[Giovanni Bajo <ra...@develer.com>]

Hello,

the regular expression used to parse the XML produced by "svn log" was a
little too optimistic: it basically assumed that no other files but the
branch itself where being added in the commit that creates a branch. This
was actually a sound assumption (as usually branches are created with "svn
mv URL URL", so nothing else is changed in that commit), but it's broken by
the GCC repository (generated by cvs2svn) which contains some spurious
(wrong) trash in the commits that generate the branches.

This fix lets "svnmerge init" succeed also within the GCC repository. Now,
if ElementTree was part of the standard core, this piece of regexp would be
gone already, but I'm too lazy to use SAX when a regexp (somehow) suffices.
For now at least...

Committed as r18341.

[[[
Fix the regular expression used to parse the XML produced by "svn log" to
extract the copy-from information: it was too broad and was allowing
spurious
matches (eg. when used on the GCC repository).

* contrib/client-side/svnmerge.py:
  (get_copyfrom): Make the regexp more strict.
]]]
-- 
Giovanni Bajo

Re: [PATH, svnmerge] Fix "svnmerge init" on the GCC repository branches

[Giovanni Bajo <ra...@develer.com>]

Mattias Engdegård <ma...@virtutech.se> wrote:

>> the regular expression used to parse the XML produced by "svn log"
>> was a little too optimistic: it basically assumed that no other
>> files but the branch itself where being added in the commit that
>> creates a branch. This was actually a sound assumption (as usually
>> branches are created with "svn mv URL URL", so nothing else is
>> changed in that commit), but it's broken by the GCC repository
>> (generated by cvs2svn) which contains some spurious (wrong) trash in
>> the commits that generate the branches.
>
> (s/mv/cp/ I presume)
>
> Branches may also be created by WC->URL copies, which allows changes
> to be done at the same time. It does not have to be trash.

Yes that's right. I was just describing what situation prompted my fix.
svnmerge should now behave fine also in the scenario you describe.

Giovanni Bajo


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: [PATH, svnmerge] Fix "svnmerge init" on the GCC repository branches

[Mattias Engdeg�rd <ma...@virtutech.se>]

"Giovanni Bajo" <ra...@develer.com> writes:
>the regular expression used to parse the XML produced by "svn log" was a
>little too optimistic: it basically assumed that no other files but the
>branch itself where being added in the commit that creates a branch. This
>was actually a sound assumption (as usually branches are created with "svn
>mv URL URL", so nothing else is changed in that commit), but it's broken by
>the GCC repository (generated by cvs2svn) which contains some spurious
>(wrong) trash in the commits that generate the branches.

(s/mv/cp/ I presume)

Branches may also be created by WC->URL copies, which allows changes to be
done at the same time. It does not have to be trash.


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org