From:	Dave Love <d DOT love AT dl DOT ac DOT uk>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: Building f2c
Followup-To:	comp.lang.fortran
Date:	16 Jun 1997 18:17:33 +0100
Organization:	Daresbury Laboratory, Warrington WA4 4AD, UK
Message-ID:	<rzqbu56sc0i.fsf@djlvig.dl.ac.uk>
References:	<3398C14B DOT 447A AT lps DOT u-psud DOT fr> <19970606202156 DOT 41381 AT gil DOT physik DOT rwth-aachen DOT de>
NNTP-Posting-Host:	djlvig.dl.ac.uk
Mime-Version:	1.0 (generated by tm-edit 7.106)
Lines:	22
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

>>>>> "Christoph" == Christoph Kukulies <kuku AT gilberto DOT physik DOT rwth-aachen DOT de> writes:

 Christoph> On Fri, Jun 06, 1997 at 07:02:51PM -0700, Daniel TAUPIN wrote:
 >> Making and recompiling f2c from the v1appl zips fails for the reason
 >> that no file tokdefs.h is given.
 >> And when I use the previous version's tokdefs.h, the bison finds errors
 >> in it.

 Christoph> Bison? Dangerous. I would suggest to use byacc. (I'm just
 Christoph> afraid that bison could make grammar problems, at least I
 Christoph> can rely on byacc, provided you take care for c \tmp dir
 Christoph> :)

If one must regenerate the f2c parser files, the problem is that the
current version (1.25) of Bison has broken yacc compatibility.  It now
reports errors in the f2c input files (and other sources) which were
previously warnings with `-y'.  This has been reported as a bug but
hadn't changed last I checked.  The previous version of Bison is OK.

I'd use clean, current f2c source from Netlib.

[Followups redirected.]

- Raw text -