[cairo] Release 1.14.4 was not created from master branch. Is this normal?

Adrian Johnson ajohnson at redneon.com
Wed Nov 4 01:18:59 PST 2015

On 04/11/15 11:48, LRN wrote:
> On 04.11.2015 3:54, qlb1234 wrote:
>> I was trying to build GTK+ (which requires Cairo) since earlier this year
>> and have been waiting for a Cairo release that is "newer than 2015-04-14"
>> (says configure script of GTK+) and now 1.14.4 is released. But with the
>> latest release of Cairo, GTK+ still complains about my Cairo build doesn't
>> have the function cairo_win32_surface_create_with_format(), which is in the
>> header file /src/cairo-win32.h. I went on cloned the project and found out
>> that the release 1.14.4 was branch-ed off master branch back in 2014-10-14,
>> when 1.14.0 was released.
> It saddens me that cairo_win32_surface_create_with_format() was not
> backported to the stable branch (since it is an addition and hardly changes
> anything in the way cairo library behaves). In my defence, i was not aware
> that something like that was maybe needed, and i didn't follow the progress
> of the 1.14.x closely, because i just apply patches to my local version of
> cairo anyway.

New API is not supposed to be added to the stable branch.

I only just noticed we have a 1.14 branch. Since master had a 1.14.x
version I had been committing bug fixes to master. It appears the 1.14
branch is missing some of my most recent bug fixes.

Usually after a new release, stable updates are released from master.
Once the number of bug fixes drops off and we are ready to start adding
new features and API, a stable branch is branched off from master and
the version on master is bumped to an odd number.

> It is normal for a stable branch to be created at some point and to keep it
> going for a while. 1.14.0 was such branch back in 2014, and 1.14.4 is the
> continuation of that branch.

More information about the cairo mailing list