[PUSHED][3-5] fdo#39415 Writer collapsing table border painting fixes

[Miklos Vajna]

On Thu, Jun 28, 2012 at 02:59:46PM +0200, Michael Stahl <mstahl at redhat.com> wrote:
> hi all, to fix this LO 3.4 regression i've committed 3 patches, which
> should be applied to libreoffice-3-5.
> 
> i've just pushed these to gerrit, which has strangely resulted in 3
> reviews instead of the expected one, but you can look at these here:
> 
> https://gerrit.libreoffice.org/239
> https://gerrit.libreoffice.org/240
> https://gerrit.libreoffice.org/241

Pushed now, thanks.

> the thing i don't understand is how you would add sign-offs to these
> before pushing them, but i haven't actually pushed a gerrit patch yet so
> probably i don't understand it yet :)

As far as I understand, you need to do git commit --amend -s manually or
so -- when gerrit will be the main repo, then just the review will be
enough (without manual push) and gerrit will automatically add the
signed-off-by.

Anyway, this is the first time I tried to use logerrit review, and I
found 3 issues. The generated commandline should be like:

ssh logerrit gerrit review -m \"looks good\" --verified=+1 --code-review=2 1080c131c0c512a9d5273e860a48af13667bc9e8

But logerrit:

- prints " instead of \"
- uses codereview insead of code-review (who tested this? ;) )
- tries to use the changeid instead of the sha1 of the commit

The first two is trivial and I fixed it, the last one is something I'm
not sure about - Norbert/Björn, care to have a look? The error message
is:

fatal: "I7605eda6ef8111612ceb1ded79c0811de9ca7b25" is not a valid patch set

Just getting the sha1 from FETCH_HEAD is what I did manually for now,
but that's not a solution - if I gerrit-cherry-pick two independent
patches, then I have no idea how to get the original sha1 for the first
one, unless we make gerrit-cherry-pick use git cherry-pick -x.