home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
TeX 1995 July
/
TeX CD-ROM July 1995 (Disc 1)(Walnut Creek)(1995).ISO
/
tex-k
/
tex-k-archive.past
/
1994.12.gz
/
1994.12
/
000080_khc@eece.ksu.edu_Wed Dec 21 11:32:00 1994.msg
< prev
next >
Wrap
Internet Message Format
|
1994-12-30
|
3KB
Received: from wiz.eece.ksu.edu by cs.umb.edu with SMTP id AA20542
(5.65c/IDA-1.4.4 for <tex-k@cs.umb.edu>); Wed, 21 Dec 1994 18:33:20 -0500
Received: from elmo.eece.ksu.edu by wiz.eece.ksu.edu with smtp
(Smail3.1.28.1 #7) id m0rKaX7-000A8wC; Wed, 21 Dec 94 17:33 CST
Received: by elmo.eece.ksu.edu (Smail3.1.28.1 #7)
id m0rKaVz-0001NaC; Wed, 21 Dec 94 17:32 CST
Message-Id: <m0rKaVz-0001NaC@elmo.eece.ksu.edu>
Date: Wed, 21 Dec 94 17:32 CST
From: khc@eece.ksu.edu (Kenneth H. Carpenter)
To: tex-k@cs.umb.edu
Subject: bug in the patch in web2c.kpathsea-2.5.help
Today I tried to build tex from sources from the ctan archives
that are the most recent. I followed the directions in the
file web2c.kpathsea-2.5.help to incorporate kpathsea-2.5 in tex.
All was well except for a bug in the patch provided in web2c.kpathsea-2.5.help.
I have found a work-around for it, and just in case no one else has
hit is yet, I am reporting it here:
The patch to the file web2c/configure in web2c.kpathsea-2.5.help
contains a bug which causes compilation to fail for tex/mf on a
Sun Sparc1 running SunOS4.1.3 using gcc2.6.1 and flex 2.4.7.
The error is due to the configure in the top level setting the
pointer type for the lex yytext, then in the configure in web2c
the test is skipped, so that the LEX_OUTPUT_ROOT symbol is null.
This causes the makefile produced to not have a file name for the
output from (f)lex. The section in error is from line 2810 to line
2830 in web2c.kpathsea-2.5.help:
------------------------------------------------------------
! echo $ac_n "checking for yytext declaration""... $ac_c" 1>&4
! if eval "test \"`echo '${'ac_cv_prog_lex_yytext_pointer'+set}'`\" = set"; then
! echo $ac_n "(cached) $ac_c" 1>&4
! else
! # POSIX says lex can declare yytext either as a pointer or an array; the
! # default is implementation-dependent. Figure out which it is, since
! # not all implementations provide the %pointer and %array declarations.
! #
! # The minimal lex program is just a single line: %%. But some broken lexes
! # (Solaris, I think it was) want two %% lines, so accommodate them.
! ac_cv_prog_lex_yytext_pointer=no
echo '%%
! %%' | $LEX
! if test -f lex.yy.c; then
! LEX_OUTPUT_ROOT=lex.yy
! elif test -f lexyy.c; then
! LEX_OUTPUT_ROOT=lexyy
! else
! { echo "configure: error: cannot find output from $LEX, giving up" 1>&2; exi
t 1; }
fi
------------------------------------------------------------
In this part of the patch, if the first "if" evaluates "true" then
LEX_OUTPUT_ROOT is never defined.
A work-around for this bug is to run ./configure from subdirectory
web2c after ./configure has finished running from the directory above,
and without defining the cache to it.
The fix should be to put LEX_OUTPUT_ROOT in the catch and read it in
to the recursive configures.