This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug bootstrap/9968] When gxx-include-dir contains 'include', its removed...
- From: "neroden at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 9 Jul 2003 07:03:34 -0000
- Subject: [Bug bootstrap/9968] When gxx-include-dir contains 'include', its removed...
- References: <20030305174601.9968.raul@pleyades.net>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9968
------- Additional Comments From neroden at gcc dot gnu dot org 2003-07-09 07:03 -------
Try configuring with --prefix=/user/gccfull
rather than --prefix=/user/gccfull/
The code in update_path recognizes the prefix in your directories, strips it,
and passes @g++include/myg++ to translate_name instead of the correct
@g++/include/myg++
One fix is to guarantee that std_prefix as used in prefix.c never ends in a /. A harder fix is to edit update_path to strip trailing slashes from
std_prefix before it does the comparison. Yet another fix is to guarantee that the string constructed by concatenation has a / between the G++ and
the "include".
But I think the best fix :-) is to change the demented scheme used by prefix.c to eliminate that whole @blah/blah and $blah/blah interface to
translate_name. It seems designed to allow repeated substitution (so that @G++/ gets replaced with @GCC/ gets replaced with $FOO/ gets replaced
with $BAR/ ...) but that seems quite unnecessary and even undesirable. (I could be wrong, of course...) If that's eliminated, a single layer of
substitution can be better implemented by having the 'key' as a separate argument to translate_name.
(Or possibly the entire 'key' business could be eliminated outright, as I don't see what it's used for anyway!)