Repository navigation
tests: minimal test_deps for older standalone tests (batch 4 of 6, libidn–node-lts) - #1407
Conversation
…bidn–node-lts) Replace `base` in test_deps with bash-bootstrap plus the tools each script actually calls, so the test closure no longer masks undeclared runtime libraries. Test commands are unchanged; tests are not in the spec hash. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
No actionable comments were generated in the recent review. Walkthrough
Merge Risk:
|
- maven: runtime_deps += coreutils, sed. bin/mvn computes MAVEN_HOME with `dirname` (and runs `uname`); without coreutils MAVEN_HOME became `/` and `mvn --version` printed no "Apache Maven". With a .mvn/jvm.config it also runs tr and sed. - mono: runtime_deps += bash. /usr/bin/mcs is a #!/bin/sh wrapper, so `/bin/mcs --version` failed ENOENT (no interpreter) once base was gone. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
agent-137
left a comment
There was a problem hiding this comment.
Review on Karl's behalf. This is a comment, not an approval. Nothing was built for it. The findings come from reading the declarations and fetching upstream sources, and each one says how it was checked. The CI rollup on this head is now FAILURE.
CI: minimal build FAILURE, completed 08:36Z. Its details URL is just https://minimal.dev, so GitHub doesn't show which package or test failed. All other checks pass. CodeRabbit was rate-limited and posted no review. The PR body says some failures are expected, but this one has to be identified and fixed on the branch before merge.
Should-fix
packages/mariadb/build.ncl:53(runtime_deps) and:131(dev_files_are_usable,test_deps = [bash-bootstrap]):/bin/mysql_config --cflagsruns MariaDB'sscripts/mysql_config.sh. build.sh keeps it deliberately as a/bin/shscript.- What breaks: that script calls
sedthroughout. It computesbasedirwithecho $me | sed …, and it builds$cflagsthrough afor remove in …; do cflags=$(echo "$cflags" | sed …)loop. With sed missing,cflagscomes out empty and--cflagsprints nothing. The test then fails with "mysql_config --cflags names no directory containing mysql.h". - Why it passed before:
basesupplied sed. mariadb's runtime_deps are bash, glibc, ncurses, openssl, pcre2 and zlib, with no sed. - Fix: add
sedto mariadbruntime_deps, the same way gradle, maven and pnpm got it in this series. Users running mysql_config hit this too. - Confidence: this comes from my knowledge of the upstream script, not from a build. It may be the
minimal buildfailure; check the bot log. - Checked: the mariadb test commands, runtime_deps, and the build.sh comments around line 105 ("They are /bin/sh scripts").
- What breaks: that script calls
Nit
- PR body: "changes no package output" is no longer true. Commit cd95a45 changes runtime_deps for maven (coreutils, sed) and mono (bash), so both rebuild.
Verdict: needs changes (identify the failing build; mariadb sed).
mysql_config / mariadb_config are /bin/sh scripts that derive basedir and build --cflags through sed; without it --cflags prints nothing and dev_files_are_usable fails. base used to supply sed (agent-137 on #1407). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…true` `scmp_sys_resolver --help 2>&1 || true` passed with nothing installed. The test now resolves `read` against two fixed ABIs and requires x86_64=0 and aarch64=63 (libseccomp 2.6.1 src/syscalls.csv), which exercises the library's syscall tables and gives the same answer on either build host. Bash builtins only, so test_deps stays [bash-bootstrap]. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Re-checks older standalone tests with minimal
test_deps(batch 4 of 6,libidn–node-lts): 25 packages, 57 tests.These tests were written with
test_deps = [base].basebrings in bash, coreutils, grep, sed, gawk, diffutils, findutils, tar, gzip, xz, bzip2, file and tzdata, and through them zlib, zstd, libmagic, libstdc++ and more. That masks missing runtime libraries: a binary that needs an undeclared library still loads, because base put the library in the test closure. The smoke-test batches of 11-06/07 (#1364–#1368, #1372, #1373) used minimal deps and found several real missing runtime_deps this way.What changed
Only
test_deps. No test command was changed.baseis removed from every test'stest_deps.bash-bootstrapis added where a command runs/bin/bashor/bin/sh(its only runtime_dep is glibc).coreutils,grep,diffutils,findutils,gawk,tar,gzip. They were found by scanning each script's command positions, including multi-linem%"…"%scripts and#!/usr/bin/envshebangs.toolchain,jdk,ocaml,git, and so on) are kept as they were.let base = import …stays whereverbaseis still a build dep.testsare not part of the spec hash, so the test_deps changes rebuild nothing. The exceptions rebuild because their runtime_deps change: mariadb (+ sed, for mysql_config), maven (+ coreutils, sed) and mono (+ bash, for its /bin/sh wrappers).Expected failures
Some of these tests are expected to fail. That is the point of the PR. A failure with the minimal closure, typically exit 127 with empty stderr (a loader failure) or a missing shared library, shows a package with an UNDECLARED runtime_dep that
basewas hiding. Each one gets fixed in a follow-up commit on this branch by adding the missingruntime_deps(which does rebuild that package). A failure can also come from a script calling a tool the scan missed. Those get fixed by adding that tool totest_deps, never by addingbaseback.The failures were deliberately not predicted ahead of time. This PR exists so the build-bot runs every test under the minimal closure.
Tests
base, glibc-locale-c-utf8bash-bootstrap, glibc-locale-c-utf8base, toolchainbash-bootstrap, toolchain, coreutils, grepbasebash-bootstrapbasebash-bootstrap, grepbasebash-bootstrap, grepbasebash-bootstrap, grepbase, glibc, gccbash-bootstrap, glibc, gcc, grepbasebash-bootstrap, grepbase, glibc, gccbash-bootstrap, glibc, gcc, grepbasebash-bootstrap, grepbase, glibc, gccbash-bootstrap, glibc, gcc, grepbasebash-bootstrap, grepbase, glibc, gccbash-bootstrap, glibc, gcc, grepbasebash-bootstrap, coreutils, grepbasebash-bootstrapbasebash-bootstrap, coreutils, grepbasebash-bootstrap, grepbasebash-bootstrapbasebash-bootstrap, grepbasebash-bootstrap, coreutils, grepbasebash-bootstrap, grepbasebash-bootstrap, grepbasebash-bootstrap, coreutils, grepbase, ocaml, ocamlfind, toolchain, grep, coreutilsbash-bootstrap, ocaml, ocamlfind, toolchain, grep, coreutilsbasebash-bootstrapbase(none)basebash-bootstrap, coreutils, grepbasebash-bootstrap, coreutils, grepbasebash-bootstrap, coreutils, grepbasebash-bootstrap, grepbasebash-bootstrap, coreutils, grepbasebash-bootstrapbasebash-bootstrap, coreutilsbasebash-bootstrap, grepbasebash-bootstrapbasebash-bootstrapbasebash-bootstrap, grepbasebash-bootstrap, coreutils, gawkbasebash-bootstrap, grepbasebash-bootstrap, grepbasebash-bootstrap, grepbasebash-bootstrap, grepbasebash-bootstrap, coreutilsbasebash-bootstrap, coreutilsbase, ca-certificatesbash-bootstrap, ca-certificates, coreutilsbasebash-bootstrap, grepbasebash-bootstrap, grepbasebash-bootstrap, grepbasebash-bootstrap, coreutilsbasebash-bootstrap, coreutilsbasebash-bootstrap, grepbasebash-bootstrap, grepbasebash-bootstrap, grepbasebash-bootstrap, grepbasebash-bootstrap, coreutilsbasebash-bootstrap, coreutilsbase, ca-certificatesbash-bootstrap, ca-certificates, coreutilsAlso: a scmp_sys_resolver test that can fail
scmp_sys_resolver --help 2>&1 || truepassed with nothing installed. It now resolvesreadfor two fixed ABIs and requiresx86_64=0andaarch64=63(libseccomp 2.6.1src/syscalls.csv), the same answer on either build host. Bash builtins only.🤖 Generated with Claude Code
Summary by CodeRabbit