Summary: | sys-devel/bison-3.7.1 has parallel build issues in tests | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Paolo Pedroni <paolo.pedroni> |
Component: | Current packages | Assignee: | Gentoo's Team for Core System packages <base-system> |
Status: | RESOLVED INVALID | ||
Severity: | normal | CC: | sam |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
bison-3.7.1:20200901-113125.log.gz - example of failing tests
bison-3.7.1:20200901-124609.log.gz - Successful tests Patch to ebuild to execute tests sequentially testsuite.log of failing 268 test |
Description
Paolo Pedroni
2020-09-01 13:35:10 UTC
Created attachment 657740 [details]
bison-3.7.1:20200901-124609.log.gz - Successful tests
Created attachment 657742 [details, diff]
Patch to ebuild to execute tests sequentially
I cannot reproduce and I am not convinced that this is caused by parallel execution,
> 225: Syntax error in consistent error state: yacc.c ok
> 242: %no-default-prec without %prec ok
> 266: %union syncline ok
> 265: Prologue syncline ok
> 282: Output columns ok
> 136: C++ Output File Prefix Mapping ok
> 268: Postprologue syncline ok
> 267: %union name syncline ok
> 270: Epilogue syncline ok
> 269: Action syncline ok
> 271: %code top syncline ok
> 200: S/R in initial ok
> 284: Invalid CPP guards: --defines=9foo.h ok
> 272: %destructor syncline ok
> 283: Invalid CPP guards: --defines=input/input.h ok
> 274: syncline escapes: yacc.c ok
> 273: %printer syncline ok
> 280: %no-lines: lalr1.cc ok
> 278: %no-lines: yacc.c ok
> 279: %no-lines: glr.c ok
> 226: Syntax error in consistent error state: glr.c ok
> 281: %no-lines: glr.cc ok
> 286: Invalid CPP guards: %glr-parser --defines=9foo.h ok
> 285: Invalid CPP guards: %glr-parser --defines=input/input.h ok
> 305: Add missing %empty ok
> 288: Sane headers: ok
> 227: Syntax error in consistent error state: lalr1.cc ok
> 292: Sane headers: api.pure ok
> 289: Sane headers: %locations %debug ok
> 303: Invalid uses of %empty ok
> 275: syncline escapes: glr.c ok
> 293: Sane headers: api.push-pull=both ok
> 294: Sane headers: api.pure api.push-pull=both ok
> 300: Midrule actions ok
> 304: Valid uses of %empty ok
> 301: Typed midrule actions ok
> 308: Initial location: yacc.c api.pure %parse-param { int x } ok
> 306: Initial location: yacc.c ok
> 307: Initial location: yacc.c api.pure=full ok
> 248: %expect-rr non GLR ok
> 245: Unreachable States After Conflict Resolution ok
> 309: Initial location: yacc.c api.push-pull=both ok
> 310: Initial location: yacc.c api.push-pull=both api.pure=full ok
> 228: Syntax error in consistent error state: glr.cc ok
> 290: Sane headers: %glr-parser ok
> 315: Initial location: yacc.c api.pure=full ok
> 316: Initial location: yacc.c api.pure=full ok
> 312: Initial location: glr.c api.pure ok
> 291: Sane headers: %locations %debug %glr-parser ok
> 276: syncline escapes: lalr1.cc ok
> 311: Initial location: glr.c ok
> 317: Location print: yacc.c ok
> 287: export YYLTYPE ok
> 277: syncline escapes: glr.cc ok
> 318: Location print: glr.c ok
> 336: @$ in %initial-action implies %locations ok
> 337: @$ in %destructor implies %locations ok
> 345: %union vs. api.value.type ok
> 338: @$ in %printer implies %locations ok
Please show detailed log for failed test(s).
Created attachment 658106 [details]
testsuite.log of failing 268 test
You're right, of course :-(
Looking at the extended logs it shows that compilation (with distcc) fails once on the remote host, and then on the local one, giving duplicate error messages on stdout.
Running without distcc gives no errors, of course. Perhaps there is a connection with https://bugs.gentoo.org/722126 - for me, compilation of 3.7.1-r1 did not succeed until "--jobs 1" was specified. (In reply to Chuckery Towers from comment #6) > Perhaps there is a connection with https://bugs.gentoo.org/722126 - for me, > compilation of 3.7.1-r1 did not succeed until "--jobs 1" was specified. Nothing like it. Logs are quite clear: using distcc yields duplicate error messages that make the tests look like they're failing. Closing as INVALID. |