On Fri, Jan 08, 2010 at 12:57:26AM +1100, Craig Small wrote: > I'm the maintainer of dh-make and need some help with bug 328692. The > problem is that dh_make gets the name of the maintainer and puts it into > various template files. Not a problem for most names and most files. > However some names, such as ones with umlats etc will choke on manual > pages. > > I have heard that Unicode::Normalize may do what I want, but as I don't > really understand what this does, or really much about UTF-8 anyhow I'm > not sure.
I saw your request for help on Planet Debian. When this bug was originally filed, groff required its input to be in the ISO-8859-1 character set, which typically meant doing special encoding of anything outside that. Since last year, though, we've had groff 1.20, which massively improves Unicode input support so that you can typically just type UTF-8 directly into manual pages. As a result, the edict in policy saying that all characters in a manual page must be representable in the legacy encoding for that language was lifted in debian-policy 3.8.3.0. At this point I think that any mangling in dh-make would do more harm than good. If you just make sure that the author's name is encoded using UTF-8 (which you should be doing already, since policy requires it for changelog and control files, although I haven't checked dh-make's code), you shouldn't need to do any other normalisation beyond that. Certainly, the original bug submitter's name works fine in manual pages now. If you get any bugs in future about non-ASCII author names being handled incorrectly in manual pages, feel free to reassign them over to man-db and I'll investigate them from there. Cheers, -- Colin Watson [[email protected]] -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
