Repository navigation
go's third wall, clisp's configure, and perl narrowed to a half-clean - #630
Merged
Merged
Conversation
go: -DGOROOT_FINAL=\\"...\\" is a Makefile idiom. mk gives double quotes no meaning and passes them through, and rc never re-lexes the contents of an expanded variable -- the same fact that makes every -DFOO=1 safe through $CFLAGS -- so the backslashes reached cpp and the macro was not a string literal. build.c:21 answered `syntax error, last name: goroot', naming the variable two files on from the flag. Three walls in three rounds, every one in this mkfile rather than in go. clisp: both mkfile parses are past, every .d generates, and configure runs -- printing `line NNNN: 5: Bad file descriptor' hundreds of times. Autoconf keeps config.log on fd 5, so the closing `C compiler cannot create executables' is the verdict of a configure that could not write its own log. sys/lib/tests/execfd-test.sh asks it in seven sections, and section 5 is run-redir's own `exec 6<>' line, which CLAUDE.md records as having moved for an unidentified reason -- one question now instead of two. 0 failures on glibc. The plan recorded, not started: hand-answer config.h and write an mkfile rather than run configure and GNU make. perl: the second run cleaned cmd/perl alone, and libperl.a is built in lib/perl. So op.c/perl.c/universal.c were recompiled under the new bit-field rule and fifty other objects were not, and both halves handle OP *. sizeof(OP) is 40 under the old rule and 32 under the new one, so one half writes a 40-byte op into a 32-byte slot at the end of a slab -- which is what a fault at a heap-top address looks like. Next run is lib/perl cleaned first, or a complete distclean now that clisp's clean cannot hang. edg: lib and cmd build and cpfe runs. No functional test yet. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGAwvvTwDg2yknFkmZ3qzs
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
go: -DGOROOT_FINAL=\"...\" is a Makefile idiom. mk gives double quotes no meaning and passes them through, and rc never re-lexes the contents of an expanded variable -- the same fact that makes every -DFOO=1 safe through $CFLAGS -- so the backslashes reached cpp and the macro was not a string literal. build.c:21 answered `syntax error, last name: goroot', naming the variable two files on from the flag. Three walls in three rounds, every one in this mkfile rather than in go.
clisp: both mkfile parses are past, every .d generates, and configure runs -- printing
line NNNN: 5: Bad file descriptor' hundreds of times. Autoconf keeps config.log on fd 5, so the closingC compiler cannot create executables' is the verdict of a configure that could not write its own log. sys/lib/tests/execfd-test.sh asks it in seven sections, and section 5 is run-redir's own `exec 6<>' line, which CLAUDE.md records as having moved for an unidentified reason -- one question now instead of two. 0 failures on glibc. The plan recorded, not started: hand-answer config.h and write an mkfile rather than run configure and GNU make.perl: the second run cleaned cmd/perl alone, and libperl.a is built in lib/perl. So op.c/perl.c/universal.c were recompiled under the new bit-field rule and fifty other objects were not, and both halves handle OP *. sizeof(OP) is 40 under the old rule and 32 under the new one, so one half writes a 40-byte op into a 32-byte slot at the end of a slab -- which is what a fault at a heap-top address looks like. Next run is lib/perl cleaned first, or a complete distclean now that clisp's clean cannot hang.
edg: lib and cmd build and cpfe runs. No functional test yet.
Claude-Session: https://claude.ai/code/session_01WGAwvvTwDg2yknFkmZ3qzs