FS#55882 - [meson] regression: arch-meson overrides don't work

Attached to Project: Arch Linux
Opened by Chet Gurevitch (chetgurevitch) - Friday, 06 October 2017, 01:56 GMT
Last edited by Levente Polyak (anthraxx) - Monday, 09 October 2017, 22:52 GMT
Task Type Bug Report
Category Packages: Extra
Status Closed
Assigned To Jan Alexander Steffens (heftig)
Levente Polyak (anthraxx)
Architecture All
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description: The change made in 0.42.1-3 breaks overrides. Moving "$@" back to its previous position fixes it.
This task depends upon

Closed by  Levente Polyak (anthraxx)
Monday, 09 October 2017, 22:52 GMT
Reason for closing:  Fixed
Additional comments about closing:  0.43.0-1
Comment by Jan Alexander Steffens (heftig) - Friday, 06 October 2017, 04:41 GMT
Weird; my experience was the opposite.
Comment by Chet Gurevitch (chetgurevitch) - Friday, 06 October 2017, 07:27 GMT
Options not specified in arch-meson get passed correctly, but for example if I add --buildtype=debug to meson_options in the systemd PKGBUILD, run makepkg, enter src/build, and check the options with meson configure buildtype is still set to release. If I move the user arguments to the end of the script everything works b/c the last instance takes precedence presumably. I ran into this while packaging gtk4 locally to test out a patch. For some reason gtk4 fails to build with buildtype release, which was very convenient for teasing out this issue lol.

Anyways removing meson configure was definitely necessary for the script to work from outside of the build directory, but as far as I can tell moving the user arguments wasn't. Meson does some weird stuff with input parsing though which might create some odd edge cases, let me know if there's anything you'd like me to test.
Comment by Jan Alexander Steffens (heftig) - Friday, 06 October 2017, 07:31 GMT
`-Db_pgo=generate -Db_pgo=use` results in `generate` overriding `use`. So the order in which arguments take precedence is not consistent...
Comment by Jan Alexander Steffens (heftig) - Sunday, 08 October 2017, 23:36 GMT
Is it better in 0.43.0-1?
Comment by Chet Gurevitch (chetgurevitch) - Monday, 09 October 2017, 22:51 GMT
Yup, everything seems to work properly now.

Loading...