Compatibility¶
zig-common |
Zig |
tools built on it |
|---|---|---|
0.8.1 |
0.17.1 |
eb-staged-rebuild, eb-keep-logs and eessi-arch-guard on their next releases (umask read without changing it, std names without deprecated aliases) |
0.8.0 |
0.17.1 |
eb-staged-rebuild 1.0.0 (errno text as Python prints it, from the errno the kernel gave); eb-keep-logs builds against it in CI |
0.7.0 |
0.17.1 |
eb-staged-rebuild (Zig port; exec errors and the ledger group stop of its build-events path); eb-keep-logs builds against it in CI |
0.6.0 |
0.17.1 |
eb-staged-rebuild (Zig port; failing fault points in its -Dtest-ops artifact); eb-keep-logs builds against it in CI |
0.5.0 |
0.17.1 |
eb-staged-rebuild (Zig port; bounded ledger call); eb-keep-logs builds against it in CI |
0.4.0 |
0.17.1 |
eb-staged-rebuild (Zig port); eb-keep-logs builds against it in CI |
0.3.0 |
0.17.1 |
eb-staged-rebuild (Zig port; |
0.2.0 |
0.17.1 |
eb-staged-rebuild (Zig port); eb-keep-logs builds against it in CI |
0.1.0 |
0.17.1 |
eb-keep-logs 0.2.0 |
A Zig upgrade lands here first, as one commit, after zig build test passes
on the new version; each tool then moves to it in a commit of its own. The
scheduled CI job builds every tool against main, so a library change that
breaks a tool shows before the tool moves to it.
Where errno text differs from Python¶
sys.describe prints glibc’s strerror text, as Python’s os.strerror does,
for every errno the Zig standard library’s linux.E names. An errno it does
not name prints [Errno N] Unknown error N; on x86_64 that is errno 134,
which glibc calls “Memory page has hardware error”.