[980425] makeinfo errors

[980425] makeinfo errors

[B. James Phillippe]

Hello,

	I posted this a while ago for 0418 and was asked for more
information.  I know this must be me-specific, or someone else would have
posted it to the list already.  Perhaps I have the stupid bug?

/usr/src/egcs-980425/install-sh -c -m 644 /usr/src/egcs-980425/gcc/cccp.1
/usr/man/man1/cccp.1
                    chmod a-x /usr/man/man1/cccp.1
/usr/src/egcs-980425/objdir/texinfo/makeinfo/makeinfo
-I/usr/src/egcs-980425/gcc /usr/src/egcs-980425/gcc/cpp.texi -o cpp.info
     Making info file `cpp.info' from `/usr/src/egcs-980425/gcc/cpp.texi'. 
-o: No such file or directory Making info file `cpp.info' from `cpp.info'. 
makeinfo: Removing output file `cpp.info' due to errors; use --force to
preserve.  make[1]: *** [cpp.info] Error 2 make[1]: Leaving director y
`/usr/src/egcs-980425/objdir/gcc' make: *** [install-gcc] Error 2

thanks,
-bp
--
B. James Phillippe <bryan@terran.org>
Linux Software Engineer, WGT Inc.
http://earth.terran.org/~bryan

Re: [980425] makeinfo errors

[Jim Wilson]

It appears that the getopt_long function call in texinfo's main function
is failing.  It isn't recognizing -o as an option, so it tries to treat it
as a file, and then we get a 'No such file of directory' message from perror.

getopt_long should be coming from libtxi.a which is built from
texinfo/lib/*.c.

Perhaps the problem here is that your system compiler (gcc?, cc?) is
miscompiling texinfo/lib/getopt1.c.  You could try compiling egcs with
a different compiler than the one you have been using.

Perhaps the problem is that getopt_long is coming from someplace other
than libtxi.a, and this other getopt_long function does not behave the
way that makeinfo expects.

Jim

Re: [980425] makeinfo errors

[B. James Phillippe]

On Wed, 29 Apr 1998, Jim Wilson wrote:

> It appears that the getopt_long function call in texinfo's main function
> is failing.  It isn't recognizing -o as an option, so it tries to treat it
> as a file, and then we get a 'No such file of directory' message from perror.

Exactly.

> Perhaps the problem here is that your system compiler (gcc?, cc?) is
> miscompiling texinfo/lib/getopt1.c.  You could try compiling egcs with
> a different compiler than the one you have been using.

No, that wasn't it.  However, I did find the problem.  I happen to be one
of those very rare people who lives with the environment variable
POSIXLY_CORRECT set. :)  getopt (et al) act differently with this variable
set, and programs which use command-line arguments in a non-POSIXly
conformant fashion will probably break.  When I unset this variable, the
problem vanishes.  I will look at it closer this evening to determine
exactly what is happening.  My guess is that argument processing is
stopping early.

thanks,
-bp
--
B. James Phillippe <bryan@terran.org>
Linux Software Engineer, WGT Inc.
http://earth.terran.org/~bryan

Re: [980425] makeinfo errors

[Andreas Schwab]

B James Phillippe <bryan@terran.org> writes:

|> No, that wasn't it.  However, I did find the problem.  I happen to be one
|> of those very rare people who lives with the environment variable
|> POSIXLY_CORRECT set. :)  getopt (et al) act differently with this variable
|> set, and programs which use command-line arguments in a non-POSIXly
|> conformant fashion will probably break.  When I unset this variable, the
|> problem vanishes.  I will look at it closer this evening to determine
|> exactly what is happening.  My guess is that argument processing is
|> stopping early.

POSIX does not allow options to occur after non-options.  If you put
POSIXLY_CORRECT into the environment then getopt follows POSIX.

-- 
Andreas Schwab                                      "And now for something
schwab@issan.informatik.uni-dortmund.de              completely different"
schwab@gnu.org

Re: [980425] makeinfo errors

[Jim Wilson]

	I happen to be one
	of those very rare people who lives with the environment variable
	POSIXLY_CORRECT set. :)  getopt (et al) act differently with this variable
	set, and programs which use command-line arguments in a non-POSIXly
	conformant fashion will probably break. 

So we fix this problem by making sure that we always put options in front
of input files.

The following patch does this.  I am doing a build to verify that it fixes
the problem.

1998-04-30  Jim Wilson  <wilson@cygnus.com>

	* Makefile.in (cpp.info, gcc.info, INSTALL): Put -o option before
	input file.

	* cp/Makefile.in (g++FAQ.info): Put -o option before input file.

	* f/Make-lang.in (g77.info, g77.dvi, BUGS, INSTALL, NEWS): Put -o
	option before input file.

Index: Makefile.in
===================================================================
RCS file: /egcs/carton/cvsfiles/egcs/gcc/Makefile.in,v
retrieving revision 1.88
diff -p -r1.88 Makefile.in
*** Makefile.in	1998/04/17 17:31:30	1.88
--- Makefile.in	1998/05/01 01:53:16
*************** doc: info
*** 2011,2022 ****
  info: cpp.info gcc.info lang.info
  
  cpp.info: $(srcdir)/cpp.texi
! 	$(MAKEINFO) $(MAKEINFOFLAGS) -I$(srcdir) $(srcdir)/cpp.texi -o cpp.info
  
  gcc.info: $(srcdir)/gcc.texi $(srcdir)/extend.texi $(srcdir)/install.texi \
  	  $(srcdir)/invoke.texi $(srcdir)/md.texi $(srcdir)/rtl.texi \
  	  $(srcdir)/tm.texi $(srcdir)/gcov.texi
! 	$(MAKEINFO) $(MAKEINFOFLAGS) -I$(srcdir) $(srcdir)/gcc.texi -o gcc.info
  
  dvi: gcc.dvi cpp.dvi lang.dvi
  
--- 2011,2022 ----
  info: cpp.info gcc.info lang.info
  
  cpp.info: $(srcdir)/cpp.texi
! 	$(MAKEINFO) $(MAKEINFOFLAGS) -I$(srcdir) -o cpp.info $(srcdir)/cpp.texi
  
  gcc.info: $(srcdir)/gcc.texi $(srcdir)/extend.texi $(srcdir)/install.texi \
  	  $(srcdir)/invoke.texi $(srcdir)/md.texi $(srcdir)/rtl.texi \
  	  $(srcdir)/tm.texi $(srcdir)/gcov.texi
! 	$(MAKEINFO) $(MAKEINFOFLAGS) -I$(srcdir) -o gcc.info $(srcdir)/gcc.texi
  
  dvi: gcc.dvi cpp.dvi lang.dvi
  
*************** cpp.dvi: $(srcdir)/cpp.texi
*** 2035,2042 ****
  
  
  INSTALL: $(srcdir)/install1.texi $(srcdir)/install.texi
! 	$(MAKEINFO) -D INSTALLONLY --no-header --no-split \
! 	$(srcdir)/install1.texi -o INSTALL
  #\f
  # Deletion of files made during compilation.
  # There are four levels of this:
--- 2035,2042 ----
  
  
  INSTALL: $(srcdir)/install1.texi $(srcdir)/install.texi
! 	$(MAKEINFO) -D INSTALLONLY --no-header --no-split -o INSTALL \
! 	$(srcdir)/install1.texi 
  #\f
  # Deletion of files made during compilation.
  # There are four levels of this:
Index: cp/Makefile.in
===================================================================
RCS file: /egcs/carton/cvsfiles/egcs/gcc/cp/Makefile.in,v
retrieving revision 1.11
diff -p -r1.11 Makefile.in
*** Makefile.in	1998/03/31 13:25:10	1.11
--- Makefile.in	1998/05/01 01:53:19
*************** TAGS: force
*** 291,297 ****
  force:
  
  g++FAQ.info:	$(srcdir)/g++FAQ.texi
! 	$(MAKEINFO) --no-split $(srcdir)/g++FAQ.texi -o ./g++FAQ.info
  
  # Preprocess the texi file so that the final document will have
  # hyperlinks.
--- 291,297 ----
  force:
  
  g++FAQ.info:	$(srcdir)/g++FAQ.texi
! 	$(MAKEINFO) --no-split -o ./g++FAQ.info $(srcdir)/g++FAQ.texi
  
  # Preprocess the texi file so that the final document will have
  # hyperlinks.
Index: f/Make-lang.in
===================================================================
RCS file: /egcs/carton/cvsfiles/egcs/gcc/f/Make-lang.in,v
retrieving revision 1.29
diff -p -r1.29 Make-lang.in
*** Make-lang.in	1998/03/28 11:18:05	1.29
--- Make-lang.in	1998/05/01 01:53:19
*************** f77.dvi: f/g77.dvi
*** 249,260 ****
  f/g77.info: $(srcdir)/f/g77.texi $(srcdir)/f/bugs.texi \
  	    $(srcdir)/f/g77install.texi $(srcdir)/f/news.texi \
  	    $(srcdir)/f/intdoc.texi
! 	$(MAKEINFO) -I$(srcdir)/f $(srcdir)/f/g77.texi -o f/g77.info
  
  f/g77.dvi: $(srcdir)/f/g77.texi $(srcdir)/f/bugs.texi \
  		     $(srcdir)/f/g77install.texi $(srcdir)/f/news.texi \
  		     $(srcdir)/f/intdoc.texi
! 	$(TEXI2DVI) -I$(srcdir)/f $(srcdir)/f/g77.texi -o f/g77.dvi
  
  # This dance is all about producing accurate documentation for g77's
  # intrinsics with minimum fuss.  f/ansify appends "\n\" to C strings
--- 249,260 ----
  f/g77.info: $(srcdir)/f/g77.texi $(srcdir)/f/bugs.texi \
  	    $(srcdir)/f/g77install.texi $(srcdir)/f/news.texi \
  	    $(srcdir)/f/intdoc.texi
! 	$(MAKEINFO) -I$(srcdir)/f -o f/g77.info $(srcdir)/f/g77.texi
  
  f/g77.dvi: $(srcdir)/f/g77.texi $(srcdir)/f/bugs.texi \
  		     $(srcdir)/f/g77install.texi $(srcdir)/f/news.texi \
  		     $(srcdir)/f/intdoc.texi
! 	$(TEXI2DVI) -I$(srcdir)/f -o f/g77.dvi $(srcdir)/f/g77.texi
  
  # This dance is all about producing accurate documentation for g77's
  # intrinsics with minimum fuss.  f/ansify appends "\n\" to C strings
*************** $(srcdir)/f/intdoc.texi: f/intdoc.c f/in
*** 286,300 ****
  
  $(srcdir)/f/BUGS: f/bugs0.texi f/bugs.texi
  	cd $(srcdir)/f; $(MAKEINFO) -D BUGSONLY --no-header --no-split \
! 	  --no-validate bugs0.texi -o BUGS
  
  $(srcdir)/f/INSTALL: f/install0.texi f/g77install.texi
  	cd $(srcdir)/f; $(MAKEINFO) -D INSTALLONLY --no-header --no-split \
! 	  --no-validate install0.texi -o INSTALL
  
  $(srcdir)/f/NEWS: f/news0.texi f/news.texi
  	cd $(srcdir)/f; $(MAKEINFO) -D NEWSONLY --no-header --no-split \
! 	  --no-validate news0.texi -o NEWS
  
  f77.rebuilt: $(srcdir)/f/g77.info $(srcdir)/f/BUGS $(srcdir)/f/INSTALL \
    $(srcdir)/f/NEWS 
--- 286,300 ----
  
  $(srcdir)/f/BUGS: f/bugs0.texi f/bugs.texi
  	cd $(srcdir)/f; $(MAKEINFO) -D BUGSONLY --no-header --no-split \
! 	  --no-validate -o BUGS bugs0.texi
  
  $(srcdir)/f/INSTALL: f/install0.texi f/g77install.texi
  	cd $(srcdir)/f; $(MAKEINFO) -D INSTALLONLY --no-header --no-split \
! 	  --no-validate -o INSTALL install0.texi
  
  $(srcdir)/f/NEWS: f/news0.texi f/news.texi
  	cd $(srcdir)/f; $(MAKEINFO) -D NEWSONLY --no-header --no-split \
! 	  --no-validate -o NEWS news0.texi
  
  f77.rebuilt: $(srcdir)/f/g77.info $(srcdir)/f/BUGS $(srcdir)/f/INSTALL \
    $(srcdir)/f/NEWS