po/POTFILES.in doesn't exist

po/POTFILES.in doesn't exist

[Eli Zaretskii]

When I run the configure script, bfd/configure complains about 
po/POTFILES.in being absent.  That's true: the file isn't there, since we 
now have SRC-POTFILES.in and BLD-POTFILES.in.

Does that mean bfd/configure needs to be updated somehow?  Do others see 
this error message?

Re: po/POTFILES.in doesn't exist

[Andrew Cagney]

> When I run the configure script, bfd/configure complains about 
> po/POTFILES.in being absent.  That's true: the file isn't there, since we 
> now have SRC-POTFILES.in and BLD-POTFILES.in.
> 
> Does that mean bfd/configure needs to be updated somehow?  Do others see 
> this error message?


FYI, I'm not seeing the message but that could be for many reasons. 
Anyone else?

	Andrew

Re: po/POTFILES.in doesn't exist

[Daniel Jacobowitz]

On Tue, Aug 07, 2001 at 07:04:50PM -0400, Andrew Cagney wrote:
> > When I run the configure script, bfd/configure complains about 
> > po/POTFILES.in being absent.  That's true: the file isn't there, since we 
> > now have SRC-POTFILES.in and BLD-POTFILES.in.
> > 
> > Does that mean bfd/configure needs to be updated somehow?  Do others see 
> > this error message?
> 
> 
> FYI, I'm not seeing the message but that could be for many reasons. 
> Anyone else?

Sure, I have been for a while:

checking for xgettext... (cached) /usr/bin/xgettext
./configure: ./po/POTFILES.in: No such file or directory
checking for a BSD compatible install... /usr/bin/install -c

Do you have a po/POTFILES.in lying around?

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: po/POTFILES.in doesn't exist

[Andrew Cagney]

> On Tue, Aug 07, 2001 at 07:04:50PM -0400, Andrew Cagney wrote:
> 
>> > When I run the configure script, bfd/configure complains about 
>> > po/POTFILES.in being absent.  That's true: the file isn't there, since we 
>> > now have SRC-POTFILES.in and BLD-POTFILES.in.
>> > 
>> > Does that mean bfd/configure needs to be updated somehow?  Do others see 
>> > this error message?
> 
>> 
>> 
>> FYI, I'm not seeing the message but that could be for many reasons. 
>> Anyone else?
> 
> 
> Sure, I have been for a while:
> 
> checking for xgettext... (cached) /usr/bin/xgettext
> ./configure: ./po/POTFILES.in: No such file or directory
> checking for a BSD compatible install... /usr/bin/install -c

Ah, yes, sorry.  I'm getting this as well.  Same with the BFD trunk.

	Andrew

Re: po/POTFILES.in doesn't exist

[Alan Modra]

> > checking for xgettext... (cached) /usr/bin/xgettext
> > ./configure: ./po/POTFILES.in: No such file or directory

This tweak should fix it.

	* gettext.m4: Test po/POTFILES.in exists before trying to read.

Anybody have any objections to installing this?

-- 
Alan Modra

Index: gettext.m4
===================================================================
RCS file: /cvs/src/src/gettext.m4,v
retrieving revision 1.1
diff -u -p -r1.1 gettext.m4
--- gettext.m4	2000/08/31 09:25:11	1.1
+++ gettext.m4	2001/10/02 12:43:15
@@ -256,8 +256,8 @@ __argz_count __argz_stringify __argz_nex
 
    dnl Generate list of files to be processed by xgettext which will
    dnl be included in po/Makefile.  But only do this if the po directory
-   dnl exists in srcdir.
-   if test -d $srcdir/po; then
+   dnl exists in srcdir and contains POTFILES.in.
+   if test -d $srcdir/po && test -f $srcdir/po/POTFILES.in; then
       test -d po || mkdir po
       if test "x$srcdir" != "x."; then
 	 if test "x`echo $srcdir | sed 's@/.*@@'`" = "x"; then

Re: po/POTFILES.in doesn't exist

[Andrew Cagney]

[dropped binutils]

Thanks!

BTW, which, if any, configury files need to be re-generated?

Assuming it makes it into the trunk, does anyone have an opinion on 
committing it to the 5.1 branch?  On the up-side is that it eliminates 
the configure warning, on the down side, the warning didn't appear to do 
any harm any way.

enjoy,
	Andrew


>> > checking for xgettext... (cached) /usr/bin/xgettext
>> > ./configure: ./po/POTFILES.in: No such file or directory
> 
> 
> This tweak should fix it.
> 
> 	* gettext.m4: Test po/POTFILES.in exists before trying to read.
> 

Re: po/POTFILES.in doesn't exist

[Eli Zaretskii]

> Date: Tue, 02 Oct 2001 12:05:03 -0400
> From: Andrew Cagney <ac131313@cygnus.com>
> 
> Assuming it makes it into the trunk, does anyone have an opinion on 
> committing it to the 5.1 branch?  On the up-side is that it eliminates 
> the configure warning, on the down side, the warning didn't appear to do 
> any harm any way.

I generally think that something is very wrong when the configure
script prints error messages.

Re: po/POTFILES.in doesn't exist

[Alexandre Oliva]

On Oct  2, 2001, Alan Modra <amodra@bigpond.net.au> wrote:

>> > checking for xgettext... (cached) /usr/bin/xgettext
>> > ./configure: ./po/POTFILES.in: No such file or directory

> This tweak should fix it.

> 	* gettext.m4: Test po/POTFILES.in exists before trying to read.

> Anybody have any objections to installing this?

Erhm...  IIRC, this file is imported from the GNU gettext
distribution, so I'd rather not needlessly diverge.  Anyway, what's
the point of having a `po' directory if it doesn't contain
POTFILES.in?

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me

Re: po/POTFILES.in doesn't exist

[Alan Modra]

On Tue, Oct 02, 2001 at 08:02:27PM -0300, Alexandre Oliva wrote:
> On Oct  2, 2001, Alan Modra <amodra@bigpond.net.au> wrote:
> 
> >> > checking for xgettext... (cached) /usr/bin/xgettext
> >> > ./configure: ./po/POTFILES.in: No such file or directory
> 
> > This tweak should fix it.
> 
> > 	* gettext.m4: Test po/POTFILES.in exists before trying to read.
> 
> > Anybody have any objections to installing this?
> 
> Erhm...  IIRC, this file is imported from the GNU gettext
> distribution, so I'd rather not needlessly diverge.  Anyway, what's
> the point of having a `po' directory if it doesn't contain
> POTFILES.in?

bfd/po is a little unusual in that is has SRC-POTFILES.in and
BLD-POTFILES.in rather than plain POTFILES.in.  The bug has already
been fixed in getttext-0.10.40, and as far as diverging goes, it
looks like our gettext.m4 is very old (serial 3 vs. serial 10).

I'm applying this simpler patch.  I realised the -d test was redundant
10 seconds after sending my last email. :-(

Index: gettext.m4
===================================================================
RCS file: /cvs/src/src/gettext.m4,v
retrieving revision 1.1
diff -u -p -r1.1 gettext.m4
--- gettext.m4	2000/08/31 09:25:11	1.1
+++ gettext.m4	2001/10/03 01:15:56
@@ -256,8 +256,8 @@ __argz_count __argz_stringify __argz_nex
 
    dnl Generate list of files to be processed by xgettext which will
    dnl be included in po/Makefile.  But only do this if the po directory
-   dnl exists in srcdir.
-   if test -d $srcdir/po; then
+   dnl exists in srcdir and contains POTFILES.in.
+   if test -f $srcdir/po/POTFILES.in; then
       test -d po || mkdir po
       if test "x$srcdir" != "x."; then
 	 if test "x`echo $srcdir | sed 's@/.*@@'`" = "x"; then

Re: po/POTFILES.in doesn't exist

[Alexandre Oliva]

On Oct  2, 2001, Alan Modra <amodra@bigpond.net.au> wrote:

> The bug has already been fixed in getttext-0.10.40

> I'm applying this simpler patch.

Great, thanks!

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me

[5.1] Re: po/POTFILES.in doesn't exist

[Andrew Cagney]

>> Date: Tue, 02 Oct 2001 12:05:03 -0400
>> From: Andrew Cagney <ac131313@cygnus.com>
>> 
>> Assuming it makes it into the trunk, does anyone have an opinion on 
>> committing it to the 5.1 branch?  On the up-side is that it eliminates 
>> the configure warning, on the down side, the warning didn't appear to do 
>> any harm any way.
> 
> 
> I generally think that something is very wrong when the configure
> script prints error messages.

Yes.  Er.  If anyone knows the correct tools / process to use to 
re-generate bfd's configure can they let me know.  I tried 
--enable-maintainer-mode using the 000227 auto* tools in the binutils 
ftp site and got the error:

gmake[1]: Entering directory `/home/scratch/51/gdb/native/bfd'
cd /home/scratch/51/gdb/src/bfd && aclocal
cd /home/scratch/51/gdb/src/bfd && automake --cygnus Makefile
cd /home/scratch/51/gdb/src/bfd && autoconf
configure.in:8: AC_TRY_COMPILE was called before AC_ISC_POSIX
configure.in:8: AC_TRY_RUN was called before AC_ISC_POSIX
autoconf: Undefined macros:
***BUG in Autoconf--please report*** AC_FD_MSG
***BUG in Autoconf--please report*** AC_FD_CC
***BUG in Autoconf--please report*** AC_FD_MSG
***BUG in Autoconf--please report*** AC_FD_MSG
***BUG in Autoconf--please report*** AC_FD_MSG
***BUG in Autoconf--please report*** AC_FD_MSG

Andrew

2001-10-03  Alan Modra  <amodra@bigpond.net.au>

     * gettext.m4: Test po/POTFILES.in exists before trying to read.


Index: gettext.m4
===================================================================
RCS file: /cvs/src/src/gettext.m4,v
retrieving revision 1.1
diff -u -p -r1.1 gettext.m4
--- gettext.m4	2000/08/31 09:25:11	1.1
+++ gettext.m4	2001/10/03 01:15:56
@@ -256,8 +256,8 @@ __argz_count __argz_stringify __argz_nex

     dnl Generate list of files to be processed by xgettext which will
     dnl be included in po/Makefile.  But only do this if the po directory
-   dnl exists in srcdir.
-   if test -d $srcdir/po; then
+   dnl exists in srcdir and contains POTFILES.in.
+   if test -f $srcdir/po/POTFILES.in; then
        test -d po || mkdir po
        if test "x$srcdir" != "x."; then
  	 if test "x`echo $srcdir | sed 's@/.*@@'`" = "x"; then

Re: [5.1] Re: po/POTFILES.in doesn't exist

[Andrew Cagney]

> I generally think that something is very wrong when the configure
> script prints error messages.
> 
> Yes.  Er.  If anyone knows the correct tools / process to use to re-generate bfd's configure can they let me know.  I tried --enable-maintainer-mode using the 000227 auto* tools in the binutils ftp site and got the error:
> 
> gmake[1]: Entering directory `/home/scratch/51/gdb/native/bfd'
> cd /home/scratch/51/gdb/src/bfd && aclocal
> cd /home/scratch/51/gdb/src/bfd && automake --cygnus Makefile
> cd /home/scratch/51/gdb/src/bfd && autoconf
> configure.in:8: AC_TRY_COMPILE was called before AC_ISC_POSIX
> configure.in:8: AC_TRY_RUN was called before AC_ISC_POSIX
> autoconf: Undefined macros:
> ***BUG in Autoconf--please report*** AC_FD_MSG
> ***BUG in Autoconf--please report*** AC_FD_CC
> ***BUG in Autoconf--please report*** AC_FD_MSG 

Just FYI, still no luck.

Andrew

Re: [5.1] Re: po/POTFILES.in doesn't exist

[Alexandre Oliva]

Sorry about the long delay, I've been way behind in my GDB e-mailing
folders :-(

On Oct 14, 2001, Andrew Cagney <ac131313@cygnus.com> wrote:

> cd /home/scratch/51/gdb/src/bfd && autoconf
> configure.in:8: AC_TRY_COMPILE was called before AC_ISC_POSIX
> configure.in:8: AC_TRY_RUN was called before AC_ISC_POSIX
> autoconf: Undefined macros:
> ***BUG in Autoconf--please report*** AC_FD_MSG

Which version of autoconf was this?

This last error is generally a symptom of under-quoted macro
invocations in configure.in or aclocal.m4.  Typical scenario is a
construct such as:

AC_FOO(AC_BAR(...))

The correct way to write this is:

AC_FOO([AC_BAR([...])])
       ^       ^   ^ ^

I.e., every argument of every macro should be enclosed in quotation
brackets.  Sometimes, you may get on without so much quoting, in the
hopes that the argument of the macro won't contain any active content
(i.e., no other m4 macro).  But, if any macro argument does, you may
observe this kind of problem.


As for the errors about AC_ISC_POSIX, they can't be helped.  There are
two macros called in bfd's configure that both want to be expanded
before any compilation takes place, but both of them do compiles
themselves.  So, one of them gets to be second and generate these
messages.

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me

Re: [5.1] Re: po/POTFILES.in doesn't exist

[Andrew Cagney]

> Sorry about the long delay, I've been way behind in my GDB e-mailing
> folders :-(
> 
> On Oct 14, 2001, Andrew Cagney <ac131313@cygnus.com> wrote:
> 
> 
>> cd /home/scratch/51/gdb/src/bfd && autoconf
>> configure.in:8: AC_TRY_COMPILE was called before AC_ISC_POSIX
>> configure.in:8: AC_TRY_RUN was called before AC_ISC_POSIX
>> autoconf: Undefined macros:
>> ***BUG in Autoconf--please report*** AC_FD_MSG
> 
> 
> Which version of autoconf was this?


The one in ftp://sources.redhat.com/pub/binutils that everyone is ment 
to use when re-autoconfing BFD.

> This last error is generally a symptom of under-quoted macro
> invocations in configure.in or aclocal.m4.  Typical scenario is a
> construct such as:
> 
> AC_FOO(AC_BAR(...))
> 
> The correct way to write this is:
> 
> AC_FOO([AC_BAR([...])])
>        ^       ^   ^ ^
> 
> I.e., every argument of every macro should be enclosed in quotation
> brackets.  Sometimes, you may get on without so much quoting, in the
> hopes that the argument of the macro won't contain any active content
> (i.e., no other m4 macro).  But, if any macro argument does, you may
> observe this kind of problem.

Ok, thanks.  I suspect the faulty definition is bundled in with the 
gettext that everyone is ment to use when re-building GDB (see ftp 
directory above).  Hmm, perhaps it is time to kill that gettext snapshot.

Andrew

Re: [5.1] Re: po/POTFILES.in doesn't exist

[Andrew Cagney]

> Sorry about the long delay, I've been way behind in my GDB e-mailing
> folders :-(
> 
> On Oct 14, 2001, Andrew Cagney <ac131313@cygnus.com> wrote:
> 
> 
>> cd /home/scratch/51/gdb/src/bfd && autoconf
>> configure.in:8: AC_TRY_COMPILE was called before AC_ISC_POSIX
>> configure.in:8: AC_TRY_RUN was called before AC_ISC_POSIX
>> autoconf: Undefined macros:
>> ***BUG in Autoconf--please report*** AC_FD_MSG
> 
> 
> Which version of autoconf was this?


The one in ftp://sources.redhat.com/pub/binutils that everyone is ment 
to use when re-autoconfing BFD.

> This last error is generally a symptom of under-quoted macro
> invocations in configure.in or aclocal.m4.  Typical scenario is a
> construct such as:
> 
> AC_FOO(AC_BAR(...))
> 
> The correct way to write this is:
> 
> AC_FOO([AC_BAR([...])])
>        ^       ^   ^ ^
> 
> I.e., every argument of every macro should be enclosed in quotation
> brackets.  Sometimes, you may get on without so much quoting, in the
> hopes that the argument of the macro won't contain any active content
> (i.e., no other m4 macro).  But, if any macro argument does, you may
> observe this kind of problem.

Ok, thanks.  I suspect the faulty definition is bundled in with the 
gettext that everyone is ment to use when re-building GDB (see ftp 
directory above).  Hmm, perhaps it is time to kill that gettext snapshot.

Andrew