On Thu, Jul 27, 2006 at 04:07:50PM -0700, Brian Dessent wrote: >Christopher Faylor wrote: >>Actually, we have had people who have complained because make became >>confused by certain uses of a ':' in the old version of make. I see >>now that this is because of the attempt to interpret a valid make rule >>as a MS-DOS path. >> >>So, I'm less inclined to want to recommend an upstream patch. I don't >>think I want to break valid makefiles which work on linux. > >Well surely such a patch would include sufficient logic to know that >^[a-z]:[/\\] incidates a DOS path and not to just blindly puke and cry >on a path component that happens to have a : in it somewhere.
I'm saying that I think it is possible for a valid makefile to be interpreted incorrectly in the presence of a ':' when make interprets the ':' as a filename component. I remember a couple of reports in the last few years from people who stumbled across a parsing problem due to colons which wouldn't have been a problem if cygwin's MS-DOS aware make hadn't been trying to treat colons specially. So, I don't think that MinGW's make accepts a superset of linux makefiles and MS-DOS colon-using makefiles. I think it actually accepts a subset of all valid linux makefiles. cgf -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
