|
From: | Paul Eggert |
Subject: | Re: more #include_next patches for iconv_open, netinet_in, sys_stat, etc. |
Date: | Thu, 17 May 2007 08:59:16 -0700 |
User-agent: | Gnus/5.110006 (No Gnus v0.6) Emacs/21.4 (gnu/linux) |
Bruno Haible <address@hidden> writes: > If it triggers the "great confusion" that the gcc doc is talking about, we > can always revert back to the "#pragma GCC system_header" approach. OK, thanks, I applied those extra patches. Simon, if you have an objection to your part please let me know and I'll undo them. My biggest worry wasn't confusion, but compilers that complain about "#include_next" even in code that is ifdeffed out. The C standard says such programs are OK but I vaguely recall some compilers from decades ago that complained about such things. It most likely won't happen in GNU/Linux compilers (since /usr/include/limits.h uses #include_next) but if it happens on other hosts we can easily change the strategy to work around the problem.
[Prev in Thread] | Current Thread | [Next in Thread] |