Patch: Fix .gdbtkinit reading

Patch: Fix .gdbtkinit reading

[Tom Tromey]

Currently Insight tries to encode preference values which include `='.
Unfortunately due to a small bug Insight doesn't properly re-read such
values.

I found this while trying to use a saved session.  I tested the
appended patch on that session.  Ok to commit?

Tom

Index: ChangeLog
from  Tom Tromey  <tromey@redhat.com>

	* library/prefs.tcl (pref_read): Version number appears in
	comment.

Index: library/prefs.tcl
===================================================================
RCS file: /cvs/src/src/gdb/gdbtk/library/prefs.tcl,v
retrieving revision 1.11
diff -u -r1.11 prefs.tcl
--- library/prefs.tcl 2001/08/09 20:43:54 1.11
+++ library/prefs.tcl 2001/08/21 21:39:37
@@ -78,7 +78,8 @@
 	  {^[ \t\n]*#.*} {
 	    # Comment.  We recognize one magic comment that includes
 	    # the version number.
-	    if {[regexp -- "^GDBtkInitVersion: (\[0-9\]+)\$" $line v]} {
+	    if {[regexp -- "^# GDBtkInitVersion: (\[0-9\]+)\$" $line \
+		   dummy v]} {
 	      set version $v
 	    }
 	  }

Re: Patch: Fix .gdbtkinit reading

[Keith Seitz]

On 21 Aug 2001, Tom Tromey wrote:

> Currently Insight tries to encode preference values which include `='.
> Unfortunately due to a small bug Insight doesn't properly re-read such
> values.
>
> I found this while trying to use a saved session.  I tested the
> appended patch on that session.  Ok to commit?

Looks pretty innocent to me. Did this fix your session problem, or do you
still need someone to look at it?

Keith

> Index: ChangeLog
> from  Tom Tromey  <tromey@redhat.com>
>
> 	* library/prefs.tcl (pref_read): Version number appears in
> 	comment.
>
> Index: library/prefs.tcl
> ===================================================================
> RCS file: /cvs/src/src/gdb/gdbtk/library/prefs.tcl,v
> retrieving revision 1.11
> diff -u -r1.11 prefs.tcl
> --- library/prefs.tcl 2001/08/09 20:43:54 1.11
> +++ library/prefs.tcl 2001/08/21 21:39:37
> @@ -78,7 +78,8 @@
>  	  {^[ \t\n]*#.*} {
>  	    # Comment.  We recognize one magic comment that includes
>  	    # the version number.
> -	    if {[regexp -- "^GDBtkInitVersion: (\[0-9\]+)\$" $line v]} {
> +	    if {[regexp -- "^# GDBtkInitVersion: (\[0-9\]+)\$" $line \
> +		   dummy v]} {
>  	      set version $v
>  	    }
>  	  }
>

Re: Patch: Fix .gdbtkinit reading

[Tom Tromey]

Keith> Looks pretty innocent to me. Did this fix your session problem,
Keith> or do you still need someone to look at it?

Nope, this just fixed a reloading problem.
The other problem still exists.
I'd really like the eventual `insight' executable to have a different
command line than gdb.  That way we can make it do more useful things.
What's the current scoop on the insight/gdb exe split?

Tom

Re: Patch: Fix .gdbtkinit reading

[Keith Seitz]

On 22 Aug 2001, Tom Tromey wrote:

> Keith> Looks pretty innocent to me. Did this fix your session problem,
> Keith> or do you still need someone to look at it?
>
> Nope, this just fixed a reloading problem.
> The other problem still exists.
> I'd really like the eventual `insight' executable to have a different
> command line than gdb.  That way we can make it do more useful things.
> What's the current scoop on the insight/gdb exe split?

There was some discussion about how to do this: half the people wanted us
to create a file during the build with "int use_windows=1;" in it. The
other half wanted to tweak top.c to do something with ifdefs. Both are
kinda evil, but...

I'll try to instigate something on the gdb list.
Keith

Re: Patch: Fix .gdbtkinit reading

[Tom Tromey]

Tom> I found this while trying to use a saved session.  I tested the
Tom> appended patch on that session.  Ok to commit?

Keith> Looks pretty innocent to me.

I took this as an `ok' and I've committed the patch.

Tom

Re: Patch: Fix .gdbtkinit reading

[Keith Seitz]

On 22 Aug 2001, Tom Tromey wrote:

> Tom> I found this while trying to use a saved session.  I tested the
> Tom> appended patch on that session.  Ok to commit?
>
> Keith> Looks pretty innocent to me.
>
> I took this as an `ok' and I've committed the patch.

Yes, indeed. I should learn to be more black-and-white on checkins.

Thanks.
Keith