Hi, building Redland on Cygwin works for me.
> 1. testing variables > if variable > --- > endif > Cygwin complains about something near "if". > > Changing "if" to "ifdef" removes complaint. > > 2. setting variables > @variable = ... > Cygwin complains about something near "@" > > Changing "@" to "#", or deleting "@", removes complaint. These look like end-of-line encoding issues. Some Windows tool of yours probably converted them to CRLFs while the Cygwin tools still expect them to be LFs only and complain about the extra CR. > 3. C compilation errors in flex/bison source, for example > raptor_mkr_writer.c: At top level: > raptor_mkr_writer.c:802: error: parse error before '*' token > raptor_mkr_writer.c: In function `raptor_mkr_writer_get_feature': > raptor_mkr_writer.c:807: error: `feature' undeclared (first use in this > function) > raptor_mkr_writer.c:809: error: `mkr_writer' undeclared (first use in this > function) > > I don't yet have a fix for problem 3. Problem seems to be associated with > procedure calls where bison has matched a pattern. It makes me wonder if > gcc > is complaining about procedure argument "declarations" embedded in procedure > call. Hard to say about this without the source. The error message looks like you have a "foo *bar" declaration but have not included the typedef for foo yet. Generally, flex and bison generated source builds fine on cygwin. fix-flex and fix-bison scripts are hacks. They work with the specific flex/bison versions and input files (raptor/rasqal lexers and parsers) but do not necessarily work with other flex/bison versions or other input files. Lauri _______________________________________________ redland-dev mailing list [email protected] http://lists.librdf.org/mailman/listinfo/redland-dev
