Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
* Re: gdb.c++ vs dos names
@ 2003-01-09 17:33 Michael Elizabeth Chastain
  0 siblings, 0 replies; 11+ messages in thread
From: Michael Elizabeth Chastain @ 2003-01-09 17:33 UTC (permalink / raw)
  To: ac131313, carlton; +Cc: gdb-patches

Well, the snapshot tarball has testsuite/gdb.*/Makefile files in it,
thus there are 8.3 conflicts in every testsuite directory.  We're just
lucky that ARI saw a problem with gdb.c++/Makefile because of the
gdb.c++ part.

I can't read the script that makes the snapshot tarballs,
so I can't go any further.

. doc/gdbint.texinfo is out of date

  . "make -f Makefile.in gdb.tar.gz" does not work now that
    Makefile.in is autogen'ed, because it has plenty of @foo@ magic.

  . "configure ; make gdb.tar.gz" fails because the top Makefile
    no longer has any rules to make gdb.tar.gz or gdb.tar.

. gdbadmin is not accessible to me

  . http://sources.redhat.com/gdb/current has instructions for getting
    the administration scripts, but they are not working

    . checkout as anoncvs fails:

      bash-2.05b$ cvs -z9 -d :pserver:anoncvs@sources.redhat.com:/cvs/gdbadmin login
      Logging in to :pserver:anoncvs@sources.redhat.com:2401/cvs/gdbadmin
      CVS password:
      bash-2.05b$ cvs -z9 -d :pserver:anoncvs@sources.redhat.com:/cvs/gdbadmin co .
      cvs server: Updating .
      cvs server: failed to create lock directory for `/cvs/gdbadmin' (/cvs/gdbadmin/#cvs.lock): Permission denied
      cvs server: failed to obtain dir lock in repository `/cvs/gdbadmin'
      cvs [server aborted]: read lock failed - giving up

    . checkout with my gdb developer identity fails:

      bash-2.05b$ cvs -z9 -d :ext:chastain@sources.redhat.com:/cvs/gdbadmin co .
      Enter passphrase for RSA key '/berman/home/mec/.ssh/identity':
      Warning: Remote host denied authentication agent forwarding.
      cvs server: Updating .
      cvs server: failed to create lock directory for `/cvs/gdbadmin' (/cvs/gdbadmin/#cvs.lock): Permission denied
      cvs server: failed to obtain dir lock in repository `/cvs/gdbadmin'
      cvs [server aborted]: read lock failed - giving up

These are doco and web bugs.  Which of them would you like me to
file PR's on?

Michael C


^ permalink raw reply	[flat|nested] 11+ messages in thread
* Re: gdb.c++ vs dos names
@ 2003-01-09 20:49 Michael Elizabeth Chastain
  0 siblings, 0 replies; 11+ messages in thread
From: Michael Elizabeth Chastain @ 2003-01-09 20:49 UTC (permalink / raw)
  To: drow, gdb-patches; +Cc: ac131313, carlton

It seems to work for me.  'make -f src-release gdb.tar' gives
me a tarball with no testsuite/gdb.*/Makefile in it.

Two directories still have both a 'Makefile' and 'Makefile.in'
in the tarball:

  % tar tf gdb.tar | grep Makefile
  ...
  utils/sparclite/Makefile.in
  utils/sparclite/Makefile
  ...
  utils/wince/Makefile.in
  utils/wince/Makefile
  ...

utils/configure.in does not use AC_OUTPUT so there is some other glitch.
This is probably a bug as the generated Makefile has had all its
variables substituted by the release process.

config/djgpp/fnchange.lst is still missing some legitimate files
in gdb.c++.  I will submit a patch for that shortly.

Michael C

===

2003-01-09  Daniel Jacobowitz  <drow@mvista.com>

	* Makefile.in (ALL_SUBDIRS): New variable.
	(subdirs, clean, distclean): Use it.
	* gdb.arch/Makefile.in: Clean up Makefile.  Remove rules for
	regenerating Makefile, since it is generated from the top level.
	* gdb.asm/Makefile.in: Likewise.
	* gdb.base/Makefile.in: Likewise.
	* gdb.c++/Makefile.in: Likewise.
	* gdb.disasm/Makefile.in: Likewise.
	* gdb.java/Makefile.in: Likewise.
	* gdb.mi/Makefile.in: Likewise.
	* gdb.threads/Makefile.in: Likewise.
	* gdb.trace/Makefile.in: Likewise.


^ permalink raw reply	[flat|nested] 11+ messages in thread
* Re: gdb.c++ vs dos names
@ 2003-01-09 17:52 Michael Elizabeth Chastain
  2003-01-09 18:23 ` Daniel Jacobowitz
  0 siblings, 1 reply; 11+ messages in thread
From: Michael Elizabeth Chastain @ 2003-01-09 17:52 UTC (permalink / raw)
  To: ac131313, carlton, mec; +Cc: gdb-patches

Michael C lies:
> Well, the snapshot tarball has testsuite/gdb.*/Makefile files in it,
> thus there are 8.3 conflicts in every testsuite directory.

Whoops, that is a lie, I was sloppy.  David C has a very good point
about gdb.hp/gdb.stabs.

In the snapshot tarball gdb/testsuite has these subdirectories:

  gdb.arch gdb.asm gdb.base gdb.c++ gdb.disasm gdb.hp gdb.java gdb.mi
  gdb.stabs gdb.threads gdb.trace

These subdirectories have 'Makefile' files:

  gdb.arch gdb.asm gdb.base gdb.c++ gdb.disasm gdb.java gdb.mi gdb.threads
  gdb.trace

The difference is gdb.hp and gdb.stabs.  configure.in handles these
as conditionally set 'configdirs'.  All the other testsuite/gdb.*/Makefile
files are produced by an AC_OUTPUT line.

Michael C


^ permalink raw reply	[flat|nested] 11+ messages in thread
* Re: gdb.c++ vs dos names
@ 2003-01-09 16:45 Michael Elizabeth Chastain
  2003-01-10 20:22 ` Eli Zaretskii
  0 siblings, 1 reply; 11+ messages in thread
From: Michael Elizabeth Chastain @ 2003-01-09 16:45 UTC (permalink / raw)
  To: ac131313, carlton; +Cc: gdb-patches

If you look in config/djgpp/fnchange.lst, every file in
testsuite/gdb.c++ is in there.  I presume it's because of the gdb.c++.
All the new files like pr-574.exp need to be registered.

But I'm looking at my source+build directories for gcc_5_3-branch and
HEAD and I don't see any difference with 'Makefile.in' and 'Makefile'.

In both branch and trunk:

  source dir for testsuite/gdb.c++ has a 'Makefile.in'
  build dir for testsuite/gdb.c++ has a 'Makefile'

So I don't know why ari.doschk.bug has this for message for HEAD
and does not have it for gcc_5_3_branch.

Andrew C:
> I suspect that the new makefile stuff is forgetting to clean it up after 
> a configure?  (The script is run over a release, and not the files 
> checked out of CVS).

Ermmm, it looks like an interaction between new makefile stuff and the
release process, neither of which I know very well.

Michael C


^ permalink raw reply	[flat|nested] 11+ messages in thread
* gdb.c++ vs dos names
@ 2003-01-09 16:22 Andrew Cagney
  2003-01-09 17:24 ` David Carlton
  2003-01-09 17:27 ` David Carlton
  0 siblings, 2 replies; 11+ messages in thread
From: Andrew Cagney @ 2003-01-09 16:22 UTC (permalink / raw)
  To: David Carlton, Michael Elizabeth Chastain; +Cc: gdb-patches

David, Michael,

The ARI is tossing up some name problems for GDB's [c++] testsuite 
directory.  One has me puzzled:

testsuite/gdb.c++/Makefile

I suspect that the new makefile stuff is forgetting to clean it up after 
a configure?  (The script is run over a release, and not the files 
checked out of CVS).

Andrew


^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2003-01-10 20:22 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-01-09 17:33 gdb.c++ vs dos names Michael Elizabeth Chastain
  -- strict thread matches above, loose matches on Subject: below --
2003-01-09 20:49 Michael Elizabeth Chastain
2003-01-09 17:52 Michael Elizabeth Chastain
2003-01-09 18:23 ` Daniel Jacobowitz
2003-01-09 18:50   ` Daniel Jacobowitz
2003-01-09 18:55     ` Andrew Cagney
2003-01-09 16:45 Michael Elizabeth Chastain
2003-01-10 20:22 ` Eli Zaretskii
2003-01-09 16:22 Andrew Cagney
2003-01-09 17:24 ` David Carlton
2003-01-09 17:27 ` David Carlton

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox