Re: A.R. Index for GDB version 6.1.91_20040719

Re: A.R. Index for GDB version 6.1.91_20040719

[Michael Elizabeth Chastain]

eliz> I fixed the conflicts in bfd and gdb/sim.  As for
eliz> gdb/testsuite/gdb.ada/gnat_ada.gpr.in, I wonder what is that file
eliz> needed for, as I cannot see any reference to it except in Makefile.in
eliz> in the same directory, where the file is removed in the clean: target.
eliz> Can we simply remove this file, instead of remapping it to something
eliz> that doesn't have 2 dots in the name?

This file is actually used.

Googling on the filename turns up an explanation:

  http://sources.redhat.com/ml/gdb-patches/2004-02/msg00783.html

The actual usage point is in the gnatmake command.
Here's an excerpt from gdb.log:

  Executing on host: gnatmake null_record.adb -P/tmp/migbat-testgdb-VZzpSUK2/test/gdb.ada/gnat_ada -gdwarf-2 -g2 -lm -o /tmp/migbat-testgdb-VZzpSUK2/test/gdb.ada/null_record (timeout = 300)

The "-P/blah/gdb.ada/gnat_ada" switch reads the file gnat_ada.gpr
from the build directory.  The "gpr" stands for "gnat project".

$BUILD/gnat_ada.gpr.in is made from $SOURCE/gnat_ada.gpr by
the usual autoconf magic.  Search for "gnat_ada.gpr" in the
gdb/testsuite/configure.

And the name has to be foo.gpr.in with the ".gpr" for the
"gnatmake -P" to work and the ".in" for the autoconfigury.

I'm pretty clueless about Ada but I think that's what's going on;
corrections welcomed.

Michael C

Re: A.R. Index for GDB version 6.1.91_20040719

[Daniel Jacobowitz]

On Mon, Jul 19, 2004 at 05:01:50PM -0400, Michael Chastain wrote:
> eliz> I fixed the conflicts in bfd and gdb/sim.  As for
> eliz> gdb/testsuite/gdb.ada/gnat_ada.gpr.in, I wonder what is that file
> eliz> needed for, as I cannot see any reference to it except in Makefile.in
> eliz> in the same directory, where the file is removed in the clean: target.
> eliz> Can we simply remove this file, instead of remapping it to something
> eliz> that doesn't have 2 dots in the name?
> 
> This file is actually used.
> 
> Googling on the filename turns up an explanation:
> 
>   http://sources.redhat.com/ml/gdb-patches/2004-02/msg00783.html
> 
> The actual usage point is in the gnatmake command.
> Here's an excerpt from gdb.log:
> 
>   Executing on host: gnatmake null_record.adb -P/tmp/migbat-testgdb-VZzpSUK2/test/gdb.ada/gnat_ada -gdwarf-2 -g2 -lm -o /tmp/migbat-testgdb-VZzpSUK2/test/gdb.ada/null_record (timeout = 300)
> 
> The "-P/blah/gdb.ada/gnat_ada" switch reads the file gnat_ada.gpr
> from the build directory.  The "gpr" stands for "gnat project".
> 
> $BUILD/gnat_ada.gpr.in is made from $SOURCE/gnat_ada.gpr by
> the usual autoconf magic.  Search for "gnat_ada.gpr" in the
> gdb/testsuite/configure.
> 
> And the name has to be foo.gpr.in with the ".gpr" for the
> "gnatmake -P" to work and the ".in" for the autoconfigury.
> 
> I'm pretty clueless about Ada but I think that's what's going on;
> corrections welcomed.

If that's a correct analysis then it is not hard to rename the file;
autoconf is flexible in this regard.  See config.h and config.in in gdb/.

-- 
Daniel Jacobowitz

Re: A.R. Index for GDB version 6.1.91_20040719

[Joel Brobecker]

On Mon, Jul 19, 2004 at 05:01:50PM -0400, Michael Chastain wrote:
> eliz> I fixed the conflicts in bfd and gdb/sim.  As for
> eliz> gdb/testsuite/gdb.ada/gnat_ada.gpr.in, I wonder what is that file
> eliz> needed for, as I cannot see any reference to it except in Makefile.in
> eliz> in the same directory, where the file is removed in the clean: target.
> eliz> Can we simply remove this file, instead of remapping it to something
> eliz> that doesn't have 2 dots in the name?

Michael's explainations are acurate. There is a little confusion, though:
gnat_ada.gpr.in is not deleted by make clean.  It is gnat_ada.gpr that's
deleted, and the later is generated from gnat_ada.gpr.in by configure.

> $BUILD/gnat_ada.gpr.in is made from $SOURCE/gnat_ada.gpr by
> the usual autoconf magic.  Search for "gnat_ada.gpr" in the
> gdb/testsuite/configure.

Slight correction (was probably a typo). $BUILD/gnat_ada.gpr is
generated from $SOURCE/gnat_ada.gpr.in.

Back to Eli's problem: We need to have a name such that, after
generation from configure, we end up with "gnat_ada.gpr" (fortunately,
gnat_ada.gpr fits the 8+3 profile). Is that possible?

-- 
Joel

Re: A.R. Index for GDB version 6.1.91_20040719

[Eli Zaretskii]

> Date: Mon, 19 Jul 2004 18:15:09 -0400
> From: Daniel Jacobowitz <drow@false.org>
> 
> > (At first I wanted to use gnat_ada.gpr_in, but then I thought:
> > what if someone configures in the source directory, and what if
> > configuration reaches down into testsuite/ even though testsuite/
> > does not actually run on msdos).
> 
> Correct; I don't know how configuration happens on msdos though.

Almost as it happens on Posix platforms: the configure script is run
(via ported Bash), except that the script and the configure scripts in
subdirectories are patched first to avoid file-naming problems and
problems with drive letters in file names.  See
config/djgpp/djconfig.sh and config/djgpp/config.sed for the gory
details.

The upshot of all this is that any file that is part of the package
configury is actually accessed during the configure phase, and thus
cannot be renamed by fnchange.lst.  If it cannot be renamed in the
repository (which is not the case with gnat_ada.gpr.in), the only
alternative is to rename it as part of the patches done by djconfig.sh
before it runs the configure script.

Re: A.R. Index for GDB version 6.1.91_20040719

[Eli Zaretskii]

> Date: Mon, 19 Jul 2004 18:14:02 -0400 (EDT)
> From: mec.gnu@mindspring.com (Michael Elizabeth Chastain)
> 
> (At first I wanted to use gnat_ada.gpr_in, but then I thought:
> what if someone configures in the source directory, and what if
> configuration reaches down into testsuite/ even though testsuite/
> does not actually run on msdos).

All this does happen on MS-DOS, and therefore gnat_ada.gpr_in is not
good because the generated gnat_ada.gpr will overwrite it.

> Also my understanding is: if the ms-dos port actually uses the file, the
> name has to fit in 8.3.  But if the ms-dos port does not use the file,
> the name just has to be 8.3 unique with truncation, plus some other
> requirements like "no more than one dot".

Not really accurate: even if the file _is_ used, we only care that its
name is unique after 8.3 truncation (plus some other requirements).
This is because all DOS file-related system calls silently truncate
file names to 8.3 limits before they start to actually act on the
file.

> The test suite is not used on ms-dos so null_record.adb and
> null_record.exp are fine.  They get unpacked as null_rec.adb and
> null_rec.exp, so they wouldn't actually work, but they survive the
> process of unpacking the tarball.

And if the test suite worked for the DJGPP port, these two files would
actually work, as explained above, because when the test opens
null_record.adb, say, it will happily find null_rec.adb instead.

Note that files that are part of the configury stuff are actually
accessed on MS-DOS, since the configure script is run in its entirety,
as on Unix.

Re: A.R. Index for GDB version 6.1.91_20040719

[Eli Zaretskii]

> Date: Mon, 19 Jul 2004 17:19:47 -0400 (EDT)
> From: mec.gnu@mindspring.com (Michael Elizabeth Chastain)
> 
> I see.  I could do:
> 
>   AC_OUTPUT([... gdb.ada/gnat_ada.gpr:gdb.ada/gnat_ada.gpr_in ...])
> 
> The autoconf manual even says: "Doing this allows you to keep your
> filenames acceptable to MS-DOS, or to prepend and/or append
> boilerplate to the file."
> 
> Looks good to me.  I'll pursue this.

Thanks!

Re: A.R. Index for GDB version 6.1.91_20040719

[Daniel Jacobowitz]

On Mon, Jul 19, 2004 at 06:14:02PM -0400, Michael Chastain wrote:
> joel> Back to Eli's problem: We need to have a name such that, after
> joel> generation from configure, we end up with "gnat_ada.gpr" (fortunately,
> joel> gnat_ada.gpr fits the 8+3 profile). Is that possible?
> 
> Yeah, I'm testing a patch right now.
> 
> gnat_ada.gpr.in is the default name, but AC_OUTPUT has a syntax for
> specifying an explicit name, so I'm just using that.  The new name will
> be gnat_ada.gin if that's okay with daniel.

That should work fine.

> (At first I wanted to use gnat_ada.gpr_in, but then I thought:
> what if someone configures in the source directory, and what if
> configuration reaches down into testsuite/ even though testsuite/
> does not actually run on msdos).

Correct; I don't know how configuration happens on msdos though.

> Also my understanding is: if the ms-dos port actually uses the file, the
> name has to fit in 8.3.  But if the ms-dos port does not use the file,
> the name just has to be 8.3 unique with truncation, plus some other
> requirements like "no more than one dot".

I think that's right.

-- 
Daniel Jacobowitz

Re: A.R. Index for GDB version 6.1.91_20040719

[Michael Elizabeth Chastain]

joel> Back to Eli's problem: We need to have a name such that, after
joel> generation from configure, we end up with "gnat_ada.gpr" (fortunately,
joel> gnat_ada.gpr fits the 8+3 profile). Is that possible?

Yeah, I'm testing a patch right now.

gnat_ada.gpr.in is the default name, but AC_OUTPUT has a syntax for
specifying an explicit name, so I'm just using that.  The new name will
be gnat_ada.gin if that's okay with daniel.

(At first I wanted to use gnat_ada.gpr_in, but then I thought:
what if someone configures in the source directory, and what if
configuration reaches down into testsuite/ even though testsuite/
does not actually run on msdos).

Also my understanding is: if the ms-dos port actually uses the file, the
name has to fit in 8.3.  But if the ms-dos port does not use the file,
the name just has to be 8.3 unique with truncation, plus some other
requirements like "no more than one dot".

The test suite is not used on ms-dos so null_record.adb and
null_record.exp are fine.  They get unpacked as null_rec.adb and
null_rec.exp, so they wouldn't actually work, but they survive the
process of unpacking the tarball.

Michael C

Re: A.R. Index for GDB version 6.1.91_20040719

[Michael Elizabeth Chastain]

drow> If that's a correct analysis then it is not hard to rename the file;
drow> autoconf is flexible in this regard.  See config.h and config.in in gdb/.

I see.  I could do:

  AC_OUTPUT([... gdb.ada/gnat_ada.gpr:gdb.ada/gnat_ada.gpr_in ...])

The autoconf manual even says: "Doing this allows you to keep your
filenames acceptable to MS-DOS, or to prepend and/or append
boilerplate to the file."

Looks good to me.  I'll pursue this.

Michael C

Re: A.R. Index for GDB version 6.1.91_20040719

[Eli Zaretskii]

> Date: Sun, 18 Jul 2004 23:22:53 -0400
> From: Andrew Cagney <ac131313@redhat.com>
> 
> Eli, fyi, there appear to be a number of 8.3 problems in the 6.2 
> pre-release.
> <http://sources.redhat.com/gdb/ari/>

I fixed the conflicts in bfd and gdb/sim.  As for
gdb/testsuite/gdb.ada/gnat_ada.gpr.in, I wonder what is that file
needed for, as I cannot see any reference to it except in Makefile.in
in the same directory, where the file is removed in the clean: target.
Can we simply remove this file, instead of remapping it to something
that doesn't have 2 dots in the name?

Re: A.R. Index for GDB version 6.1.91_20040719

[Eli Zaretskii]

> Date: Sun, 18 Jul 2004 23:22:53 -0400
> From: Andrew Cagney <ac131313@redhat.com>
> 
> Eli, fyi, there appear to be a number of 8.3 problems in the 6.2 
> pre-release.
> <http://sources.redhat.com/gdb/ari/>

I'll look at this ASAP.  Can this wait until tonight (my time, GMT+3)?

A.R. Index for GDB version 6.1.91_20040719

[Andrew Cagney]

Eli, fyi, there appear to be a number of 8.3 problems in the 6.2 
pre-release.
<http://sources.redhat.com/gdb/ari/>
Andrew