Ensure allocated path is long enough for pssolar #8261
Conversation
See GenericMappingTools/pygmt#2959 for background. Ensure we add at least 1 point when going to pole.
|
I am rebuilding the tar balls but no need for new Windows or macOS installers since they were never affected by this strange bug. |
|
Don't know why it was said that it didn't affect Windows. It did. |
|
OK, I just echoed what Yvonne said. So you got the commit included in your new installers then? |
|
No. Saw only now. |
|
Please check out the 6.5.0 tag and build the win64 installers again, then Paul should upload the new installers to the FTP and I will replace the old ones on the release page. |
|
That will probably take me 2 to 3 hours. |
|
Well, true, but we need to move to a system where this is all done by CI so we dont rely on people being alive or on travel or eating or whatever. All this manual activity (from me to) lead to issues. |
|
That is why a freeze should really mean a freeze. New installers and SHAs https://fct-gmt.ualg.pt/tmp/GMT5/gmt-6.5.0-win64.exe |
I am very sorry! I did not expect that this fix will be included in GMT 6.5, and I was busy with other stuff today. |
|
Yvonne, if you didn't see it is very likely because it wasn't there, but you probably used the VCPKG build, that I find often strangely different from my builds (c.f. the different test failures from what I see) |
|
@joa-quim Are you sure the sha256sum values are correct? I have: |
|
@seisman The sha's you posted above are correct and correspond to what I have in my local SHAs.txt file. Either I forgot to update it (which I think I did) or it copied a cache file. Update the SHAs.txt file at site. |
|
OK, I've uploaded the new installers to the release page and also updated the shasum file. |
See GenericMappingTools/pygmt#2959 for background. Ensure we add at least 1 point when going to pole. For whatever reason this is not needed on Windows and macOS but cost us just another allocation point.