emacs: updated emacsen-common, new emacs28 and emacs30 - #1264
emacs: updated emacsen-common, new emacs28 and emacs30#1264PovlAbrahamsen wants to merge 1 commit into
Conversation
| # will be available inside the loop. | ||
|
|
||
| mkdir -p %p/var/lib/emacsen-common/state/flavor/pending | ||
| for flavor in %p/var/lib/emacsen-common/state/flavor/installed/*[!~]; do |
There was a problem hiding this comment.
For me on 13.7/x86_64 with only emac26 installed, the installed versions are in the physical file %p/var/lib/emacsen-common/installed-flavors. The above .../state/flavor/installed/... path doesn't exist. Is this path something that started with emacs28 and/or emacsen-common-3 ?
There was a problem hiding this comment.
Yes, emacsen-common 3 changes from a single file %p/var/lib/emacsen-common/installed-flavors to a folder %p/var/lib/emacsen-common/state/flavor/installed/, containing a file for each variant. The update should convert from the old format to the new (via %p/var/lib/emacsen-common/state/flavor/pending/).
|
The enums bug is solved by emacs-mirror/emacs@222b597 But still FTBFS on ARM with followed by a core-dump. I definitely support getting the new version moving along and masking out older ones that do not build on newer platforms until someone figures them out. I'm disappointed we won't have an x11 option, but again "better is better, even if not perfect". Added note: leads for possible work-around to the unexec fail: |
|
f7c91fb fixes the emacs26 build on ARM/Tahoe. The x11 variant does not display well for me, but lots of things are not displaying correctly for me even in regular Terminal windows (emacs itself does). Is it worth backporting this to older emacsen? |
|
For the record, XQuartz itself is partially broken on Tahoe. See XQuartz/XQuartz/issues/438 among others. |
This pull request contains several changes:
I have tested on MacOS 15.6, Apple Silicon, XCode 16.4. I've tried various combinations of upgrading and downgrading emacsen-common and emacs packages (though only emacs28-nox and emacs30-nox build on Arm) - and it seems robust! I have not tested on Intel, or with older versions of emacs.
There are various warnings about the scripts calling install-info - but the infofiles seem to work correctly.