bail out on patch application failure

[kouhaidev]

Melange Pull Request Template

bail out on patch failure.

when building x86_64 images on apple silicon mac, melange may fail because of patch application failure.

silent failure seems like a bad idea! if patch application fails, then the pipeline should then fail, notifying the melange user so that the build can get a proper fix.

2024/09/08 16:05:02 INFO running step "Apply patches"
2024/09/08 16:05:02 WARN qemu-x86_64-static: QEMU internal SIGSEGV {code=MAPERR, addr=0x20}
2024/09/08 16:05:02 WARN Segmentation fault (core dumped)
2024/09/08 16:05:02 INFO running step "Run autoconf make"
2024/09/08 16:05:02 INFO make: Entering directory '/home/build'
2024/09/08 16:05:02 INFO /usr/bin/make -C deps -f Makefile all
2024/09/08 16:05:02 WARN qemu-x86_64-static: QEMU internal SIGSEGV {code=MAPERR, addr=0x20}
2024/09/08 16:05:03 INFO make[1]: Entering directory '/home/build/deps'
2024/09/08 16:05:03 INFO /usr/bin/make -C redis -f Makefile all

Functional Changes

~~- [ ] This change can build all of Wolfi without errors (describe results in notes)~~

Notes:

SCA Changes

~~- [ ] Examining several representative APKs show no regression / the desired effect (details in notes)~~

Notes:

Linter

~~- [ ] The new check is clean across Wolfi~~ ~~- [ ] The new check is opt-in or a warning~~

Notes:

[lf-]

O_o there's not set -e in the shell setup in the first place? nonetheless this should be fixed tbh

There is, but the nested sub-shell probably lacks it.

[xnox]

@smoser should we generate list of patches to apply, and then apply them with the for loop? or should we set -e in the subshell? is there a way to always enforce set -e in the subshells? not sure why we do things like this.