[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: pkg/39879: libX11 canoot build on systems with non-GNU 'cpp' (in this case, IRIX)
The following reply was made to PR pkg/39879; it has been noted by GNATS.
From: Stuart Shelton <srcshelton%gmail.com@localhost>
Joerg Sonnenberger <joerg%britannica.bec.de@localhost>
Subject: Re: pkg/39879: libX11 canoot build on systems with non-GNU 'cpp' (in
this case, IRIX)
Date: Sat, 8 Nov 2008 02:04:19 +0000
> On Fri, Nov 07, 2008 at 05:45:00PM +0000, srcshelton%gmail.com@localhost
>> ... which is because the pkgsrc 'cpp' wrapper actually invokes
>> CC, and this doesn't understand '-undef'. The options '-ansi',
>> '-cckr', '-xansi' all specify different sets of standard definitions
>> which could be present.
> I'm not 100% but it likely wants to have a traditional C
> preprocessor or
> at least something very near to that. What can you offer? :)
'cc -E' or 'CC -E' seem fairly traditional to me...
There is 'cpp', which the man page describes as:
cpp is a K&R C language preprocessor (not an ISO/ANSI C language
preprocessor) designed for standalone use and to be invoked as the
first pass of all Fortran compilations and any K&R mode C
with the -mp option. Thus, the output of cpp is designed to be
form acceptable as input to the next pass of the C compiler.
Standalone use of cpp on C code is not recommended, because the
functionality of cpp has been incorporated into the C compiler.
I don't know whether -traditional and -undef are GNU parameters, but
they smell that way to me. With the two problematic tests removed,
libX11 builds and tests just fine - so I'd have thought that the best
approach would be an IRIX-specific patch which simply removes the
tests (or even just the two resulting 'exit 1's) from the configure
Main Index |
Thread Index |