[PATCH] remove ada-exp.tab.c

[PATCH] remove ada-exp.tab.c

[Aidan Skinner]

* 2002-06-04  Aidan Skinner <aidan@velvet.net>
	* ada-exp.tab.c: remove, it's a generated file

(no patch, I'd just cvs remove the file)

- Aidan
-- 
aidan@velvet.net  http://www.velvet.net/~aidan/  aim:aidans42
finger for pgp key fingerprint |Unix Programmer/Admin actively seeking work
01AA 1594 2DB0 09E3 B850       |CV at http://www.velvet.net/~aidan/cv.txt
C2D0 9A2C 4CC9 3EC4 75E1       |Gis a job, go on, I can do that, go on, gis it.

Re: [PATCH] remove ada-exp.tab.c

[Joel Brobecker]

> * 2002-06-04  Aidan Skinner <aidan@velvet.net>
> 	* ada-exp.tab.c: remove, it's a generated file
> 
> (no patch, I'd just cvs remove the file)

But if you remove it, you force the dependency on yacc or bison when
building GDB, which may not be available on all machines. At ACT, we
have taken a different approach where we have this filed checked in our
repository, and we regenerate it when we modify ada-exp.y. The same goes
for ada-lex.c, generated from ada-lex.l.

The situation is analogus to configure and configure.in. 
-- 
Joel

Re: [PATCH] remove ada-exp.tab.c

[Aidan Skinner]

On Wed, Jun 05, 2002 at 09:05:51AM -0700, Joel Brobecker wrote

> > * 2002-06-04  Aidan Skinner <aidan@velvet.net>
> > 	* ada-exp.tab.c: remove, it's a generated file
> > 
> > (no patch, I'd just cvs remove the file)
> 
> But if you remove it, you force the dependency on yacc or bison when
> building GDB, which may not be available on all machines. At ACT, we

Java, Pascal and Modula2 all depend on a yacc anyway, and none have
the -exp.tab.c file in CVS.

- Aidan
-- 
aidan@velvet.net  http://www.velvet.net/~aidan/  aim:aidans42
finger for pgp key fingerprint |Unix Programmer/Admin actively seeking work
01AA 1594 2DB0 09E3 B850       |CV at http://www.velvet.net/~aidan/cv.txt
C2D0 9A2C 4CC9 3EC4 75E1       |Gis a job, go on, I can do that, go on, gis it.

Re: [PATCH] remove ada-exp.tab.c

[Andrew Cagney]

> * 2002-06-04  Aidan Skinner <aidan@velvet.net>
>> 	* ada-exp.tab.c: remove, it's a generated file
>> 
>> (no patch, I'd just cvs remove the file)
> 
> 
> But if you remove it, you force the dependency on yacc or bison when
> building GDB, which may not be available on all machines. At ACT, we
> have taken a different approach where we have this filed checked in our
> repository, and we regenerate it when we modify ada-exp.y. The same goes
> for ada-lex.c, generated from ada-lex.l.

The release and snapshot processes include steps to generate these 
files.  Hence the dependency only exists for developers using CVS.

> The situation is analogus to configure and configure.in. 

Similar.

Configure needs to be re-generated with the correct autoconf and we want 
everyone running the same script.  If it isn't we'll end up chasing our 
tail looking for host dependant autoconf bugs :-(

For .l and .y files I think they really should be rebuildable with any 
(and not a specific lex/yacc et.al.).

enjoy,
Andrew

Re: [PATCH] remove ada-exp.tab.c

[Aidan Skinner]

On Wed, Jun 05, 2002 at 12:33:19PM -0400, Andrew Cagney wrote

> For .l and .y files I think they really should be rebuildable with any 
> (and not a specific lex/yacc et.al.).

I have access to FreeBSD and OpenBSD machines, and an Irix box if I
don't scrub it for debian, is there a reasonably brain-dead
implementation available for these that I can test with?

If nobody has any further objects, I'll remove ada-exp.tab.c tommorrow
morning.

- Aidan

-- 
aidan@velvet.net  http://www.velvet.net/~aidan/  aim:aidans42
finger for pgp key fingerprint |Unix Programmer/Admin actively seeking work
01AA 1594 2DB0 09E3 B850       |CV at http://www.velvet.net/~aidan/cv.txt
C2D0 9A2C 4CC9 3EC4 75E1       |Gis a job, go on, I can do that, go on, gis it.

Re: [PATCH] remove ada-exp.tab.c

[Andrew Cagney]

> On Wed, Jun 05, 2002 at 12:33:19PM -0400, Andrew Cagney wrote
> 
> 
>> For .l and .y files I think they really should be rebuildable with any 
>> (and not a specific lex/yacc et.al.).
> 
> 
> I have access to FreeBSD and OpenBSD machines, and an Irix box if I
> don't scrub it for debian, is there a reasonably brain-dead
> implementation available for these that I can test with?

Try it with the boxes you have access to.  However, no need to go out of 
your way.  If developers encounter problems on various hosts they will 
soon let us know :-)

> If nobody has any further objects, I'll remove ada-exp.tab.c tommorrow
> morning.

No need to wait. I believe that ada-lex.c is also generated and, hence, 
should be deleted for similar reasons.

Anyway, now that the code is in CVS I've had a better look.  The 
immediate things are very standard:

- coding style, I noticed things like:
	for () {
	}
The fix is to put the .[hc] files through gdb_indent.sh and commit the 
result.

- See http://sources.redhat.com/gdb/current/ari/ and note the 
regressions (things purged from GDB) related to the ada files.  Things 
like malloc, free, printf("%p") et.al.  You could also look at errors 
and deprecated, however regressions are the big ones(1).

- Can it be built when configured with configured with 
--enable-gdb-build-warnings=,-Werror and fix the results.  If you hit 
problems due to those #if 0's don't be to worried at this stage.

- I noticed that the debug code is using ``fprintf (stderr, ...''.  It 
should use ``fprintf_unfiltered (gdb_stdlog, ...''.  It isn't under 
regressions since the purge of fprintf() is ongoing.  I don't think it 
is fair to give Pierre more work to do though :-)

These are all considered `obvious' so, at your lesure.


Do you want to start working through the other changes?


enjoy,
Andrew

(1) I checked the table last night and the script/table contained a bug. 
  Todays re-generate will be more correct.