-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Gavrichenkov, Artyom wrote: >> But this defines strlen to return the wrong type! (At least, the HP-UX >> machine I have handy has strlen returning int.) > > On several HP-UX machines near me the following is true: > >> printf '%s %s %s\n' `uname -s` `uname -r` `uname -m` > HP-UX B.11.23 ia64 >> grep strlen /usr/include/string.h | grep -v ^#\ \*pragma > extern int strlen(); > extern size_t strlen(char *); > extern size_t strlen(const char *); > extern size_t strlen(); > > where the first declaration is under #ifdef _CLASSIC_ANSI_TYPES -- this macro > isn't used anywhere in Screen code, isn't it? > > I may provide you our string.h if it's needed, but you may look in > documentation as well: http://docs.hp.com/en/B2355-60127/index.html > > Anyway, even in case some (old?) HP-UX versions had strlen() returning int > (guess it converted to size_t), it's better to have a declaration with size_t > than no declaration at all. > > I've attached new patch. You may ignore it if you choose so (that's just > warnings after all).
Thanks for the summary! Actually, after I saw Juergen's conclusion that it was just old cruft, I added checks to the configure script for string.h and strings.h, and just plopped those headers in, expecting they should DTRT. http://git.savannah.gnu.org/gitweb/?p=screen.git;a=commitdiff;h=7f787f3fc96d7f6e365d09e6d2a4f71ecca540b4 - -- Micah J. Cowan Programmer, musician, typesetting enthusiast, gamer. GNU Maintainer: wget, screen, teseq http://micah.cowan.name/ -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkkbChoACgkQ7M8hyUobTrFA3ACeK8bCAzE6ISfUhyZ4gmOd19wu ibQAn2V2crPnuh4pS7+D0M+nc5+huQGE =V3aJ -----END PGP SIGNATURE-----
