Page MenuHomeFreeBSD

lua-language-server: update to 3.16.1 and add missing PORTEPOCH.
ClosedPublic

Authored by dave_freedave.net on Wed, Dec 10, 5:12 PM.
Tags
None
Referenced Files
F139419090: D54165.diff
Thu, Dec 11, 9:05 PM
F139349954: D54165.diff
Thu, Dec 11, 2:48 AM
Restricted File
Wed, Dec 10, 6:17 PM
Unknown Object (File)
Wed, Dec 10, 5:49 PM
Subscribers
None

Details

Summary

I missed PORTEPOCH in original port. Adding it back for correct sort behaviour.
And a new version came out.

Test Plan

In stable/14, stable/15, and main:

cd /usr/ports/devel/lua-language-server
make test

And I like to fire up neovim and see if it reports errors bringing up my lua config.

portlint -A, portclippy --strict Makefile, and a poudriere testbuild.

Diff Detail

Repository
R11 FreeBSD ports repository
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

dave_freedave.net created this revision.

{F139313289} log file from poudriere testport -j 15R -p local devel/lua-language-server

Actually I do have portlint -A outpout:

WARN: Makefile: new ports should not set PORTEPOCH.

But that is intentional.

Good test plan! I also find :checkhealth to be helpful, if you bring up a buffer with lua_ls attached.

devel/lua-language-server/Makefile
21

Harmless, but: You sure you want to update this comment on every update? Sticking a placeholder in for the hash may get you what you need...

Just on inspection, this looks good. I'll get it committed this evening.

devel/lua-language-server/Makefile
21

I'm giving submodules2tuple a go instead of me doing by hand. One sec...

Yeah using the submodule2tuple shows I neglected a submodule:
actboy168:ltest:0d2d81f:ltest/3rd/json.lua/test/ltest

It does get tripped up by the fact bee.lua is fetched twice and requires a tiny bit of manual intervention. But long term that will be better than manual and having comments that get out of date.

switched to using ports-mgmt/submodules2tuple to do the initial pass.

Unfortunately I had used the SHA1 from looking at github which has 8 character hashes. The tool used 7 character. This blows up the distfiles cache. But I think its worth the better reliability of the automated tool (already found I missed a unit test for a submodule).

Just to be sure rebuilt and retested.

This revision was not accepted when it landed; it landed in state Needs Review.Thu, Dec 11, 3:12 AM
This revision was automatically updated to reflect the committed changes.