Closed Ferdi265 closed 3 days ago
Just noticed this still has test failures locally, will fix those.
It appears I broke something related to installing built packages, marking this as a draft until that is fixed.
Fixed accidently shadowing the pkgdests
variable in d9e17f9. This caused installing built packages to be broken, which is now fixed.
The test changes in 5064027 only update the list of expected commands for building split packages, since some redundant invocations of makepkg
have been removed by the changes in this PR.
Broken tests as of 5064027: all green :+1:
Fixed another issue in a26822f: Previously, whether a pkgbase was built was considered part of installer.origTargets
was dependent on the order of the packages. This wasn't as big of an issue before this PR since at least one of the buildPkg()
invocations passed the correct value of isTarget
to buildPkg()
. d9e17f9 changed this to only a single invocation of buildPkg()
per pkgbase, so now the isTarget
argument has to be correct for the first (and only) invocation. Fixed this by tracking which pkgbases are part of installer.origTargets
in handleLayer()
and passing that to installAURPackages()
.
:+1: let's get some feedback from yay-git
When building a PKGBUILD pkgbase with multiple pkgnames,
installAURPackages()
invokesbuildPkg()
multiple times for the same pkgbase. This causesprepare()
to be run multiple times for the same pkgbase, since detection of already built packages happens afterprepare()
.Additionally, detection of already built packages can fail if the split debug packages are enabled and the package does not contain any binaries, causing no
-debug
package to be created by makepkg even though it is listed bymakepkg --packagelist
.This PR fixes this by keeping track of the pkgdests built by
buildPkg()
and avoiding rebuilds of the same pkgbase in the same yay invocation.Fixes #2340.