On Sun, Mar 18, 2012 at 09:07:43AM -0700, Russ Allbery wrote: > Iustin Pop <iustin@debian.org> writes: > > On Sat, Mar 17, 2012 at 06:07:24PM -0700, Russ Allbery wrote: > > >> That's because Debian applies a local override to the *roff macros to > >> force all hyphens to be US-ASCII. But we would really like to be able > >> to remove that override at some point, since hyphens produce nicer > >> output and we'd rather not continue to carry the divergence from > >> upstream. > > > Ah. So, if I understand correctly, lintian doesn't warn for most hyphens > > since it "knows" that Debian's *roff will anyway show them as minus > > signs, but its regex "missed" this particular case due to the font > > change? > > Sort of. Lintian warns for every hyphen that it can find that it thinks > should be a US-ASCII dash, but the only place where it's confident enough > to check for that is where it's pretty sure that the man page is talking > about an option switch. So it looks for leading whitespace before the > hyphen and a pattern that looks like it's an option switch. > > Here, the font change meant there was no leading whitespace, which is why > it missed the first few dashes. > > Note that regular English text hyphens, such as the one in US-ASCII, > should stay unescaped, since those actually should become Unicode hyphens. Thanks for the well-formed explanation. Then this is an invalid bug report, and should be closed (I'm still not quite sure how to close such bugs, just email to nnn-done@ ?). Many thanks, iustin
Attachment:
signature.asc
Description: Digital signature