Update of bug #67372 (group groff):
Status: None => In Progress
Summary: [troff] \A and \w behavior change from 1.22.4 ->
1.23.0 on invalid input => [troff] \A and \w changed behavior on invalid input
from 1.22.4 -> 1.23.0
_______________________________________________________
Follow-up Comment #26:
[comment #24 comment #24:]
> [In] _groff_ Git's master branch, [I've] significantly expanded test coverage
> of valid _and_ invalid delimiters.
>
> It's still not complete. I need to:
>
> 1. Deal with a handful of control characters that are valid as GNU _troff_
> delimiters but not in AT&T _troff_. This may be unresolvable; I'm starting
> to suspect that AT&T _troff_ used this handful of character codes as "token"
> values, just as GNU _troff_ does.
Done (documented) in
[https://cgit.git.savannah.gnu.org/cgit/groff.git/commit/?id=6549b7c3d4db4dd0ee65aada318e32b2b0c057f8
6549b7c3d4], committed 14 November.
> 2. Deal with the use of escape sequences themselves as delimiters. Some
> instances of these are the land mines that exploded in _mandoc_'s test suite.
> Support for others _must_ be retained for preprocessor support. (GNU
> _tbl_'s `\[tbl` and GNU _eqn_'s `\(EQ`. In the preprocessors, there is no
> concept of "input level" or "interpolation depth", so these are necessary
> contrivances.)
[https://cgit.git.savannah.gnu.org/cgit/groff.git/commit/?id=18ed47436cf3be79f706af6046bf9789f26c95a7
Commit 18ed47436c], 28 November doesn't resolve this, but does throw some
light on it.
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/bugs/?67372>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/
signature.asc
Description: PGP signature
