Repository navigation
Conversation
r1w1s1
commented
Oct 8, 2026
- update version
reverse dependencies - development/zigNone |
✅ sbolint - development/zig ✅ |
|
@sbo-bot: build development/zig |
|
@willysr this takes a lot of time to compile. |
|
@aclemons set timeout as 4h, so unless it took more than 4h, it should be safe |
buildpkg i586 development/zigSnipped Build FailureDetails |
buildpkg x86_64 development/zigSnipped Build FailureDetails |
|
@r1w1s1 this failed to build |
|
@willysr I will check. |
2f91a1f to
7f72bec
Compare
|
@sbo-bot: build development/zig |
buildpkg i586 development/zigSnipped Build FailureDetails |
buildpkg x86_64 development/zigSnipped Build LogDetails |
sbopkglint x86_64 development/zig |
7f72bec to
ad44457
Compare
|
This PR contains too many changed packages (more than 10). Skipping normal PR checks. |
|
@sbo-bot: build development/zig |
buildpkg i586 development/zigSnipped Build FailureDetails |
ad44457 to
0018669
Compare
|
@sbo-bot: build development/zig |
buildpkg x86_64 development/zigSnipped Build FailureDetails |
0018669 to
026cad1
Compare
|
@sbo-bot: build development/zig |
buildpkg i586 development/zigSnipped Build FailureDetails |
|
The i586 build now gets past the -Oz/GCC failure, but fails while zig1 generates zig2.c: Could you check whether the i586 sbo-bot container has any memory, ulimit, or address-space limitation? If possible, could this be retried with more memory/swap and/or The failure happens during zig1 generation of zig2.c, before the normal parallel C compilation. There are similar reports upstream for native 32-bit x86/ARM builds, so this may be an inherent 32-bit bootstrap limitation rather than a container problem. |
buildpkg x86_64 development/zigSnipped Build LogDetails |
sbopkglint x86_64 development/zig |
buildpkg i586 development/zigSnipped Build FailureDetails |
buildpkg x86_64 development/zigSnipped Build LogDetails |
There should be no limitation on the container - the host has 64 GB of RAM and there is nothing else running but Jenkins and docker. Could you temporarily add something to the script for 32bit after the find / chmod step to test? maybe something like: if [ "$ARCH" = "i586" ] || [ "$ARCH" = "i686" ]; then
unset MAKEFLAGS
export MAKEFLAGS=-j1
fi |