From: Keith M. <kei...@us...> - 2009-07-14 16:59:43
|
On Monday 13 July 2009 17:28:38 Andrew Fleenor wrote: > >i had a similar experience recently; try with "-std=gnu++0x"? > > Yes, fixed! Well, you have a work around, but it isn't really fixed, is it? > now that you mention it, I seem to remember seeing > that flag before. Thank you! You are substituting ANSI conformance for conformance to a GNU standard instead. If you really wanted to guarantee that you are using only pure ANSI features, then you have lost that. The issue here, IIRC, is that some inline function definitions, which *are* exposed in __STRICT_ANSI__ mode, gratuitously depend on some other functions which are *not* ANSI, and their prototypes are explicitly *suppressed* by __STRICT_ANSI__. We, (the maintainers), need to identify and implement a sane rationale for resolving this. I can't remember if there was a formal bug report filed for this; there should have been. (I'm answering in offline mode, so can't check the tracker right now). -- Regards, Keith. |