Hi, While working on a project of ours I have come accross this problem using boost-1.32 on Mac OS X 10.4: It appears to be related to the regex library. <snip> Making all in mantrad source='mantra_request_admin.cpp' object='mantra_request_admin.o' libtool=no \ depfile='.deps/mantra_request_admin.Po' tmpdepfile='.deps/mantra_request_admin.TPo' \ depmode=gcc3 /bin/sh ../../depcomp \ g++ -DHAVE_CONFIG_H -I. -I. -I../.. -I.. -I../libs/sockets -I../libs/sigaware -I../libs/threadpool -Wall -g -D_THREAD_SAFE -c -o mantra_request_admin.o `test -f 'mantra_request_admin.cpp' || echo './'`mantra_request_admin.cpp /opt/local/include/boost/regex/v4/char_regex_traits.hpp: In member function 'uint_fast32_t boost::deprecated::char_regex_traits_i<wchar_t>::lookup_classname(const wchar_t*, const wchar_t*) const': /opt/local/include/boost/regex/v4/char_regex_traits.hpp:68: error: 'boost::re_detail::c_traits_base::<anonymous enum>' is/uses anonymous type /opt/local/include/boost/regex/v4/char_regex_traits.hpp:68: error: trying to instantiate 'template<class T> struct boost::is_integral' make[3]: *** [mantra_request_admin.o] Error 1 make[2]: *** [all-recursive] Error 1 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2 </snip> The thing is I have not had this problem on any other platform, or indeed 2 other Mac's running an (as far as I'm aware) identical setup. Does anyone have any ideas as to the possible cause of this? Many Thanks Chris This message has been checked for viruses but the contents of an attachment may still contain software viruses, which could damage your computer system: you are advised to perform your own checks. Email communications with the University of Nottingham may be monitored as permitted by UK legislation.
Sorry to take so long in getting back to you: I don't understand this error at all, the line in question doesn't instantiate any templates, just refers to the nested member of one that happens to be an enum. That enum type certainly is not used as a template argument (which would be illegal as it's an anonymous enum). Could this be some kind of setup problem at your end? Sorry, I just don't know what else to suggest at present. John.
John Maddock wrote:
I ran into this exact same problem (GCC 4.0.0, Boost 1.32, MacOS X 10.4.2 Tiger) and discovered that the culprits are the template operator== overloaded functions starting on line 560 in boost/function/function_base.hpp. Interestingly, the comment on line 442 for the function_base operator== methods seems to apply to these free-standing functions as well: // GCC 3.3 and newer cannot copy with the global operator==, due to // problems with instantiation of function return types before it // has been verified that the argument types match up. In other words, GCC is blowing up in instantiating the return types of these template functions. I've read ad nauseam on the GCC bugzilla site that anonymous enums are a no-no in template instantiations (they don't even fall under SFINAE) - but surely this is a GCC bug that bombs out on the return type like this alone... So my workaround is to disable these functions completely - it seems to work for me, although I am not building/using the entire library. Brian P.S. I just signed up to the Boost Users list. The only way I could find to reply to an existing thread was through a newsreader (and my normal news client, Entourage X, simply refused to deal with this and only this news server with an error 130 every time - so I had to use Mozilla). Is there an easier way?
participants (3)
-
Brian Dantes -
Chris Coleman -
John Maddock