On Sun, Apr 21, 2002 at 12:43:18PM -0400, [EMAIL PROTECTED] wrote: > > static /**/const char *const rcsid[] = { (const char *)rcsid, "\100(#)" msg > > } > > I'm not sure what the \100 is for (why not have it a literal @ sign? > something excessively clever is going on there), but I suspect that if > you nuked it, or replaced it with an @-sign, the problem will vanish...
Lots of this code I don't understand, and that line (+ the 4 other identical definitations included in each C file) really puzzles me. I fixed this, and also ignore the return code autoconf returns. It seems to work... Anyway, I have been messing up lots of code. I have changed krb4 to build all libraries again, and not use any from Heimdal. I have changed heimdal to use these libraries. So far there seems to be only two problems (and everything compiles): 1. ftp client and daemon in heimdal require GLOB_LIMIT and GLOB_QUOTE to be defined. However, in krb4 these are defined in glob.h, and this glob.h conflicts with the one in glibc. I have hardcoded these constants. I wonder what version of glob is being used anyway... 2. kafs in krb4 only has krb4 support, in kerberos it has krb5 support. However, krb4 requires kafs to link, and kerberos requires it to link too. I don't know what to do here. At the moment I have two conflicting libraries. As usual, I will put it online <URL:http://snoopy.apana.org.au/~ftp/debian/> Currently only krb4 is online, will put Heimdal there as soon as it builds. Assumming it does build. It built just seconds ago, lets see if dpkg-buildpackage works... rats, still more of this: ../../include/config.h:1283: warning: `RCSID' redefined I thought I fixed that :-( How many times does config.h define RCSID???? If the package doesn't build for any reason (I think it should), I will put the source online instead, so others can at least take a look. After working on this all day, I am getting just a bit sick of looking at this at the moment. Any bugs, please send reports to > /dev/create_patch_to_fix_problem, and your computer will automagically fix the problem. -- Brian May <[EMAIL PROTECTED]> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]