Summary: | 8.71 language-switch build with CYGWIN broken | ||
---|---|---|---|
Product: | Ghostscript | Reporter: | norbert.janssen |
Component: | Language Switch | Assignee: | Default assignee <ghostpdl-bugs> |
Status: | NOTIFIED FIXED | ||
Severity: | normal | CC: | williambader |
Priority: | P4 | ||
Version: | 8.71 | ||
Hardware: | PC | ||
OS: | Windows XP | ||
Customer: | 661 | Word Size: | --- |
Attachments: | patch |
Description
norbert.janssen
2010-02-19 06:24:39 UTC
I've been able to confirm with cygwin 1.7.1, though what's confusing the header wasn't immediately obvious. Hmm, windows/cygwin doesn't support quite a few POSIX signals, and that particular part of cygwin definitely doesn't inter-op with mingw (i.e. no-cygwin) headers. FWIW, the language switch build can be built with microsoft Windows SDK express (the free version), if that helps. That also looks like FAPI change from Ken - I thought that's optional? The build immediately fails with: cc1: unrecognized option `-Wno-strict-aliasing' fixing this, it fails with ../gs/base/stdint_.h:41:21: stdint.h: No such file or directory next, the build fails with ld: cannot find -ldl and this was the last bug. I cannot reproduce the errors in signal.h I stumbled on this when trying to cmake-ify the ghostpdl (i.e. replace the *.mak with *.cmake so that it can be build with gcc/msvc/... (it fits better in our normal build-environment). Just one set of makefiles to maintain, i.s.o. all flavours now, and the bonus, cmake takes care of the dependencies. Normally I build with msvc, but I wanted to try gcc also, and on Windows that means CYGWIN for me. Created attachment 5975 [details]
patch
Include stat_.h before estack.h, which defines esp macro.
Prevent a spurious macro substitution in Cygwin version of signal.h.
Backward compatibility problems with an ancient version of Cygwin are,
probably, not worth fixing.
The patch has been committed as a rev. 10777. Regression testing shows no differences. *** Bug 691480 has been marked as a duplicate of this bug. *** |