Re: [rfa] remove dependency of f-exp.tab.c on c-exp.tab.c

Re: [rfa] remove dependency of f-exp.tab.c on c-exp.tab.c

[Michael Elizabeth Chastain]

> f-exp.y wasn't added until 1994, but presumably it also had the
> artificial dependency included for the same reason; and that
> artificial dependency must have gotten accidentally left behind when
> this issue was cleaned up.

Here is one more bit of history:
  
  # top level ChangeLog

  Thu Jul 31 19:49:49 1997  Ian Lance Taylor  <ian@cygnus.com>

        * ylwrap: New file.
        * Makefile.in (DEVO_SUPPORT): Add ylwrap.
 
... so those funky serialization dependencies existed 3-4 years
before ylwrap fixed the problem in a more robust way.

gdb has more history than the Strand bookstore.  :)

Michael C

Re: [rfa] remove dependency of f-exp.tab.c on c-exp.tab.c

[Michael Elizabeth Chastain]

Hi David,

At first I thought that this might be a Makefile hack to prevent Make
from running multiple copies of yacc in parallel if the user runs a
parallel build, "make -j".

Classic yacc writes output into y.tab.c no matter what the input file
name.  Gnu yacc (bison) fixed this brain damage, but gdb builds on lots
of hosts, so it has to accommodate the brain damage.  Adding a bunch
of "foo-exp.tab.c: bar-exp-tab.c" dependencies would serialize the
yacc executions.

But gdb already works around the fixed "y.tab.c" name by using a wrapper
script, ylwrap.  ylwrap creates a unique temporary directory and runs
yacc inside there.

So I don't know what the point of that dependency is!

Michael C

Re: [rfa] remove dependency of f-exp.tab.c on c-exp.tab.c

[David Carlton]

On Mon, 2 Dec 2002 13:33:15 -0600, Michael Elizabeth Chastain <mec@shout.net> said:

> At first I thought that this might be a Makefile hack to prevent
> Make from running multiple copies of yacc in parallel if the user
> runs a parallel build, "make -j".

Ah, good call.  GDB did indeed once do this: the entry

Wed Jun 23 15:04:54 1993  K. Richard Pixley  (rich@sendai.cygnus.com)

does, among other things, the following:

	  (ch-exp.tab.c, m2-exp.tab.c): added artificial dependencies in
	  order to force parallel makes into keeping these rules separate.

f-exp.y wasn't added until 1994, but presumably it also had the
artificial dependency included for the same reason; and that
artificial dependency must have gotten accidentally left behind when
this issue was cleaned up.

With that explanation, then, I think the patch becomes pretty obvious;
I'll commit it in a couple of days unless somebody objects.

David Carlton
carlton@math.stanford.edu

Re: [rfa] remove dependency of f-exp.tab.c on c-exp.tab.c

[David Carlton]

On 02 Dec 2002 11:57:37 -0800, David Carlton <carlton@math.Stanford.EDU> said:

> With that explanation, then, I think the patch becomes pretty
> obvious; I'll commit it in a couple of days unless somebody objects.

Committed.

David Carlton
carlton@math.stanford.edu

[rfa] remove dependency of f-exp.tab.c on c-exp.tab.c

[David Carlton]

Is there any particular reason why f-exp.tab.c depends on c-exp.tab.c?
I'm assuming that it's a bug (I didn't see any reason for it in
f-exp.y), but I wanted to double-check before removing the dependency.

David Carlton
carlton@math.stanford.edu

2002-12-02  David Carlton  <carlton@math.stanford.edu>

	* Makefile.in (f-exp.tab.c): Don't depend on c-exp.tab.c.

Index: Makefile.in
===================================================================
RCS file: /cvs/src/src/gdb/Makefile.in,v
retrieving revision 1.279
diff -u -p -r1.279 Makefile.in
--- Makefile.in	9 Nov 2002 03:14:11 -0000	1.279
+++ Makefile.in	2 Dec 2002 19:10:02 -0000
@@ -1248,7 +1248,7 @@ jv-exp.tab.c: jv-exp.y
 # See comments above ...
 .PRECIOUS: f-exp.tab.c
 f-exp.tab.o: f-exp.tab.c
-f-exp.tab.c: f-exp.y c-exp.tab.c
+f-exp.tab.c: f-exp.y
 	$(SHELL) $(YLWRAP) "$(YACC)" $(srcdir)/f-exp.y  y.tab.c f-exp.tmp -- $(YFLAGS) 
 	-sed -e '/extern.*malloc/d' \
 	     -e '/extern.*realloc/d' \