[postgis-devel] PSC Vote PostGIS Committer Guideline changes
lr at pcorp.us
Fri Oct 13 08:59:29 PDT 2017
I didn't say that git pulls and git should be a requirement, just stating that people use them for testing and that's just as good and sometimes better than testing locally.
I just want assurance that you've at least made sure your stuff compiles and have done some rudimentary testing before committing.
I actually don't care if people have a test environment as long as they have some means of doing some preliminary tests before commit.
I agree with Paul, you can't test everything on your machine (I only test PostgreSQL 10 at this point, and only fix if the bots that test lower versions complain).
That's what we have bots for (so we don't have to test all permutations we claim to support) and being able to test with lower or without not having something installed that we said is not required is a good test too we don't often cover.
Our dev squad e.g. me on windows, Paul on Mac, Bborie on Slack, everyone else on Debian and Ubuntu gives extra coverage. We are missing coverage of BSDs and Solaris (but at least Mac is sorta BSDish).
The proof should be when all bots give green lights. Having all bots with red lights is rarely acceptable. Some reds and some greens may be okay e.g. if we are in middle of upgrading travis, dronie, and gittie to what we agreed should be minimal versions for a branch.
Do I need to make any rewording of 14 which reads
"14. When committing new features or significant changes to existing source code, the committer should take reasonable measures to insure that the source code continues to build and work on the most commonly supported platforms (currently Linux, Mac, and Windows), either by testing on those platforms directly, running Buildbot tests, or by getting help from other developers working on those platforms. If new files or library dependencies are added, then the configure.in, Makefile.in and related documentations should be kept up to date."
From: postgis-devel [mailto:postgis-devel-bounces at lists.osgeo.org] On Behalf Of Paul Ramsey
Sent: Friday, October 13, 2017 9:56 AM
To: PostGIS Development Discussion <postgis-devel at lists.osgeo.org>
Subject: Re: [postgis-devel] PSC Vote PostGIS Committer Guideline changes
On Fri, Oct 13, 2017 at 6:52 AM, Sandro Santilli <strk at kbt.io <mailto:strk at kbt.io> > wrote:
I think "git pulls and git" should not be a requirement.
Installing all testing support and passing "make check"
before committing should be.
For some defn's of the word "all". I don't run SFCGAL tests. (Because I'm lazy.) But if I wasn't lazy I wouldn't be testing the "without SFCGAL" case, which is also a valid setup needing coverage. I'm currently working against pgsql 10 on my laptop and 11 on my desktop, but that leaves several other supported versions I am not testing against. At some level I'm just hoping our dev squad has enough overall coverage that we'll hit all our corner cases from time to time.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the postgis-devel