https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119256
--- Comment #8 from Simon Sobisch <simonsobisch at gnu dot org> --- Am 13. Oktober 2025 00:52:07 MESZ schrieb "jklowden at gcc dot gnu.org" <[email protected]>: >https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119256 > >--- Comment #7 from James K. Lowden <jklowden at gcc dot gnu.org> --- >This patch presents a problem for us. Well, not us, but for autotest, so us. > >I don't know if it's intentional, but some squiggle underlinings have a >trailing blank. We see this diff from autotest: > >> prog.cob:22:12: error: syntax error, unexpected END-IF > 22 | END-IF >- | ^~~~~~ >+ | ^~~~~~ > >The last line has a trailing blank. The line with the - is the expected >output. The line with the + is the output. > >I have not discovered a way to include trailing whitespace in the expected >autotest output. Says the fine Autoconf manual: > >> Trailing spaces are smashed by ‘autom4te’. This is a feature. > >8.1.6 Quadrigraphs recommends vaguely to use the @&t@ quadrigraph to "to mark >trailing spaces explicitly" but my experiments thus far yield no improvement. > >Is the trailing blank after the `~` a feature or a bug? > I'd consider that a bug in the gcc underlinings. For m4 to not remove trailing blanks use [[ ]] instead of [ ] in the testcases that need that, possibly with a comment in the test allowing to find it easily for adjustment once the gcc bug is solved.
