On 05/26/2018 04:42 PM, Lucian Poston wrote:
On Sat, May 26, 2018 at 06:05:00AM -0700, Nikolay Sivov wrote:
How do I test this?I installed this font by manually extracting it from Arch's xorg-fonts-misc 1.0.3-5, but I don't see any failures.
The failure arises only when you apply the subsequent patches (patch 4/6 I believe), which cause that test map characters to the Goha font from the system fonts. So you could apply the whoel patch series except this one to see the failure.
If it doesn't fail with arch's versoin, perhaps mine is corrupted, and it is a non-issue. I isolated this change so that in such a case, you can just drop this patch.
I uploaded the one from my system, in case you wanted to risk trying it. https://transfer.sh/br8lo/GohaTibebZemen.otf
The fact that it's not working on Windows is not very important, we still should try to make it usable.
How would you define the test's expected values if we can't get them from windows?
Depends on the test. Some tests estimate using actual font data, and compare that to what windows returns, so expected values are not defined but computed.