[postgis-devel] PSC Vote: Github move from travis-ci.org to travis-ci.com - big mess now sorry

[Paul Ramsey]

While we stll have travis we have at least at ARM test. I've found the different combos of gcc/clang/C++ std seem to spit out a good swath of usable results. I don't know what I don't know, I guess, but also I do know we just don't have many non-intel architectures in the wild. To the extent that MacOS gives us a proxy to BSD, GH Actions helps. What I currently miss is a valgrind run. Squeezing out the last 500 bytes of leak and then having a valgrind run in CI would be (IMO) more helpful than trying to skin the "weird architecture" cat.

> On Nov 30, 2020, at 9:32 AM, Phillip Ross <phillip.w.g.ross+postgis at gmail.com> wrote:
> 
> That's correct, github actions does not currently provide a native arm
> runner ("runner" is basically a VM environment).  Currently, I believe
> people are using either self-hosted runners (github actions delegates
> to VMs you provision yourself) or using qemu on the standard linux
> runners to emulate other platforms (as docker does).
> 
> On Mon, Nov 30, 2020 at 8:24 AM Greg Troxel <gdt at lexort.com> wrote:
>> 
>> 
>> Do the different CI systems enable different sets of systems that can be
>> tested?  I realize there are no great answers here but it would be
>> really nice to have more operating systems included, and more non-x86
>> CPU types.
>> 
>> In unison, I project I help maintain, github actions appears (other
>> people deal with this) limited to GNU/Linux (Ubuntu), recent macOS, and
>> Windows, x86 only.
>> 
>> _______________________________________________
>> postgis-devel mailing list
>> postgis-devel at lists.osgeo.org
>> https://lists.osgeo.org/mailman/listinfo/postgis-devel
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/postgis-devel