Re: Maarten Bosmans license statement

[Norbert Thiebaud]

>
> (unless it doesn't, and that means someone else has committed changes to the
> same chunk of code in the meantime and your patch no longer applies, but let's
> deal with that only when such situation arises)
>
>> What is the correct way to submit a patch series to gerrit? I'd like
>> to avoid squashing all patches into one commit, but there also has to
>> be a way for reviewers to see that all patches in a series are
>> related.
>
> Yours is the correct way, the reviewers are well aware of false merge 
> conflicts
> of dependent patches. I'm still not sure if this is a gerrit bug or some
> sophisticated feature *grin*

WAD
the second patch cannot be 'merged' until the first one land on
master.. or the second patch is reworked to not depend on the first
one (should the first one be abandonned for some reason)

Jenkins build each patches.. but it pull each patch with it's history.
jenkins check-out the commit related to a given patch, which means for
dependents patches the later pull the former too.

Note it is possible to push patch for master with 'topic' to visually
indicate that they are related see
https://gerrit-review.googlesource.com/Documentation/intro-user.html#topics

Norbert
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: Maarten Bosmans license statement

[Katarina Behrens]

Moin, 

> Both commits have difference Change-Id tags, but I submitted them
> using a single `logerrit submit master` command. Now gerrit says for
> the second patch: 'Cannot Merge'. I could not find what this means,
> but I suppose that the second patch without the first does not apply
> cleanly on master (which is correct, they are dependent). 

Yeah, this is exactly what happens if patch2 depends on patch1 which has not 
been merged to master yet.

Once patch1 gets merged, evil 'Cannot Merge' tag at patch2 automatically 
disappears.
 
(unless it doesn't, and that means someone else has committed changes to the 
same chunk of code in the meantime and your patch no longer applies, but let's 
deal with that only when such situation arises) 

> What is the correct way to submit a patch series to gerrit? I'd like
> to avoid squashing all patches into one commit, but there also has to
> be a way for reviewers to see that all patches in a series are
> related.

Yours is the correct way, the reviewers are well aware of false merge conflicts 
of dependent patches. I'm still not sure if this is a gerrit bug or some 
sophisticated feature *grin*
-- 

Katarina Behrens

Softwareentwicklerin LibreOffice
–––
CIB software GmbH
Geschäftsstelle Hamburg
Flachsland 10
22083 Hamburg
–––
T +49 (40) / 28 48 42 -235
F +49 (40) / 28 48 42 -100

katarina.behr...@cib.de
www.cib.de
–––
Sitz: München
Registergericht München, HRB 123286
Geschäftsführer: Dipl.-Ing. Ulrich Brandner
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: Maarten Bosmans license statement

[Maarten Bosmans]

2016-08-30 16:41 GMT+02:00 Eike Rathke :
> Great, we'll welcome your patches for review on gerrit :-)

I have a question regarding the patch submission on gerrit.
In searching for a faster way to implement
SheetDataBuffer::addColXfStyle (for tdf#100709), I noticed that the
current implementation had some bugs too. To keep it clean, I split up
the work into 2 commits: one for the bugfix (which changes the
imported file) and one for the performance fix (without any output
change).
Both commits have difference Change-Id tags, but I submitted them
using a single `logerrit submit master` command. Now gerrit says for
the second patch: 'Cannot Merge'. I could not find what this means,
but I suppose that the second patch without the first does not apply
cleanly on master (which is correct, they are dependent). But oddly
enough Jenkins can build the second patch, so I gues Jenkins does use
both patches for building.

What is the correct way to submit a patch series to gerrit? I'd like
to avoid squashing all patches into one commit, but there also has to
be a way for reviewers to see that all patches in a series are
related.

Maarten
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: Maarten Bosmans license statement

[Eike Rathke]

Hi Maarten,

On Tuesday, 2016-08-30 12:59:10 +0200, Maarten Bosmans wrote:

> I'm working on some performance issues loading Excel documents in Calc.
> From looking into bugs 100709 and 53698 it seems to me that there's
> ample opportunity to speed things up significantly (at least 2x) with
> relatively minor changes.

Great, we'll welcome your patches for review on gerrit :-)

  Eike

-- 
LibreOffice Calc developer. Number formatter stricken i18n transpositionizer.
GPG key "ID" 0x65632D3A - 2265 D7F3 A7B0 95CC 3918  630B 6A6C D5B7 6563 2D3A
Better use 64-bit 0x6A6CD5B765632D3A here is why: https://evil32.com/
Care about Free Software, support the FSFE https://fsfe.org/support/?erack


signature.asc
Description: PGP signature
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/libreoffice