| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
|
|
I remember now that config.mk's change to ?= was so that xbs Makefiles
could override $(binder). That was stupid.
|
|
|
|
|
|
This test worked before because of work-arounds for systemd bugs, it had
invoked bash inside of the chroot.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
It confuses check() when building the package
|
|
|
|
|
|
|
|
|
|
This should fix a number of bugs in those two scripts, and the summarize
script simplifies aur and libremakepkg:hooks-check.sh
|
|
Because of the work-around for systemd-nspawn's buggy I/O, it was
discarding the exit status if stdout wasn't a TTY. My work-around for
*this* is to set -o pipefail.
|
|
|
|
I keep running it without any arguments, thinking that I am now working
inside of the test environment, when really it just exited right away with
a 0 status.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
find * -type f -not -name '*.patch' -exec sed -i -r 's/\s+$//' {} +
|
|
|
|
|
|
|
|
|
|
- libremakepkg: set BLACKLIST, as it will be used when checking the
PKGBUILD
- test-common.sh: make sure that $HOME isn't the default, so we don't
trash the user's home directory if accidentally run outside of testenv
- testenv: don't worry about adding dest/usr/sbin to the path
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
- set -e didn't work in blacklist-update when followed by ||
- it prompted before overwriting the local blacklist
|
|
|
