[PATCH] -nx-except-gdbtkinit option

[PATCH] -nx-except-gdbtkinit option

[Andrew STUBBS]

[-- Attachment #1: Type: text/plain, Size: 443 bytes --]

Hi,

The attached patch adds a new option named '-nx-except-gdbtkinit'. Not 
exactly a snappy name, but it does what it says on the tin and I only 
intend it for use with a wrapper script we use. I can add a shorter name 
if required.

The -nx option disables .gdbinit but also disables .gdbtkinit, which is 
not what I desire. The -nx-except-gdbtkinit does the same as -nx, but 
without affecting the configuration of Insight.

Andrew Stubbs

[-- Attachment #2: nx-except-gdbtkinit.patch --]
[-- Type: text/plain, Size: 3491 bytes --]

2005-11-16  Andrew Stubbs  <andrew.stubbs@st.com>

	* main.c (captured_main): Add option -nx-except-gdbtkinit
	and adjust -nx to support it.
	(print_gdb_help): Add --nx-except-gdbtkinit.
	* top.c (inhibit_gdbtkinit): New variable.
	* top.h (inhibit_gdbtkinit): New variable.

gdbtk/
	* generic/gdbtk.c (gdbtk_init): Change inhibit_gdbinit to
	inhibit_gdbtkinit.

Index: src/gdb/main.c
===================================================================
--- src.orig/gdb/main.c	2005-11-08 12:39:37.000000000 +0000
+++ src/gdb/main.c	2005-11-08 14:16:37.000000000 +0000
@@ -267,8 +267,8 @@ captured_main (void *data)
       {"quiet", no_argument, &quiet, 1},
       {"q", no_argument, &quiet, 1},
       {"silent", no_argument, &quiet, 1},
-      {"nx", no_argument, &inhibit_gdbinit, 1},
-      {"n", no_argument, &inhibit_gdbinit, 1},
+      {"nx", no_argument, 0, 'n'},
+      {"n", no_argument, 0, 'n'},
       {"batch-silent", no_argument, 0, 'B'},
       {"batch", no_argument, &batch, 1},
       {"epoch", no_argument, &epoch_interface, 1},
@@ -296,6 +296,7 @@ captured_main (void *data)
       {"x", required_argument, 0, 'x'},
       {"ex", required_argument, 0, 'X'},
 #ifdef GDBTK
+      {"nx-except-gdbtkinit", no_argument, &inhibit_gdbinit, 1},
       {"tclcommand", required_argument, 0, 'z'},
       {"enable-external-editor", no_argument, 0, 'y'},
       {"editor-command", required_argument, 0, 'w'},
@@ -418,6 +419,10 @@ captured_main (void *data)
 	    batch = batch_silent = 1;
 	    gdb_stdout = ui_file_new();
 	    break;
+	  case 'n':
+	    inhibit_gdbinit = 1;
+	    inhibit_gdbtkinit = 1;
+	    break;
 #ifdef GDBTK
 	  case 'z':
 	    {
@@ -896,6 +901,8 @@ Options:\n\n\
   --nx               Do not read "), stream);
   fputs_unfiltered (gdbinit, stream);
   fputs_unfiltered (_(" file.\n\
+  --nx-except-gdbtkinit\n\
+                     As for --nx, but do read gdbtkinit for insight.\n\
   --quiet            Do not print version number on startup.\n\
   --readnow          Fully read symbol files on first access.\n\
 "), stream);
Index: src/gdb/top.c
===================================================================
--- src.orig/gdb/top.c	2005-11-08 12:39:37.000000000 +0000
+++ src/gdb/top.c	2005-11-08 14:16:37.000000000 +0000
@@ -86,6 +86,7 @@
 char gdbinit[PATH_MAX + 1] = GDBINIT_FILENAME;
 
 int inhibit_gdbinit = 0;
+int inhibit_gdbtkinit = 0;
 
 /* If nonzero, and GDB has been configured to be able to use windows,
    attempt to open them upon startup.  */
Index: src/gdb/top.h
===================================================================
--- src.orig/gdb/top.h	2005-11-08 12:39:37.000000000 +0000
+++ src/gdb/top.h	2005-11-08 14:16:37.000000000 +0000
@@ -29,6 +29,7 @@ extern int linesize;
 extern FILE *instream;
 extern char gdb_dirbuf[1024];
 extern int inhibit_gdbinit;
+extern int inhibit_gdbtkinit;
 extern int epoch_interface;
 extern char gdbinit[];
 
Index: src/gdb/gdbtk/generic/gdbtk.c
===================================================================
--- src.orig/gdb/gdbtk/generic/gdbtk.c	2005-11-08 12:39:37.000000000 +0000
+++ src/gdb/gdbtk/generic/gdbtk.c	2005-11-08 14:16:37.000000000 +0000
@@ -483,7 +483,7 @@ gdbtk_init (void)
 
   /* Set up some globals used by gdb to pass info to gdbtk
      for start up options and the like */
-  xasprintf (&s, "%d", inhibit_gdbinit);
+  xasprintf (&s, "%d", inhibit_gdbtkinit);
   Tcl_SetVar2 (gdbtk_interp, "GDBStartup", "inhibit_prefs", s, TCL_GLOBAL_ONLY);
   free(s);
    

Re: [PATCH] -nx-except-gdbtkinit option

[Eli Zaretskii]

> Date: Wed, 16 Nov 2005 16:31:15 +0000
> From: Andrew STUBBS <andrew.stubbs@st.com>
> 
> The attached patch adds a new option named '-nx-except-gdbtkinit'. Not 
> exactly a snappy name, but it does what it says on the tin and I only 
> intend it for use with a wrapper script we use. I can add a shorter name 
> if required.
> 
> The -nx option disables .gdbinit but also disables .gdbtkinit, which is 
> not what I desire. The -nx-except-gdbtkinit does the same as -nx, but 
> without affecting the configuration of Insight.

First, you didn't send a documentation patch.

Second, please tell more about the situations when this option could
be useful.

Re: [PATCH] -nx-except-gdbtkinit option

[Andrew STUBBS]

[-- Attachment #1: Type: text/plain, Size: 751 bytes --]

Eli Zaretskii wrote:
> First, you didn't send a documentation patch.

Apologies, I missed this. I have now added one.

> Second, please tell more about the situations when this option could
> be useful.

The specific case I use it for is in a wrapper utility which calls gdb 
as follows (roughly):

   gdb -nx-except-gdbtkinit -d scriptdir -x globalscript \
         -x ~/.gdbinit -x ./.gdbinit .....

The purpose is (among other things) to set up transparently (to the 
user) a selection of user-defined commands (for connecting and 
configuring target boards) which are available to users in the .gdbinit 
files.

If -nx is used it has undesirable side effects when using Insight.

I'm sure others can think of other purposes for it.

Andrew Stubbs

[-- Attachment #2: nx-except-gdbtkinit.patch --]
[-- Type: text/plain, Size: 4254 bytes --]

2005-11-17  Andrew Stubbs  <andrew.stubbs@st.com>

	* main.c (captured_main): Add option -nx-except-gdbtkinit
	and adjust -nx to support it.
	(print_gdb_help): Add --nx-except-gdbtkinit.
	* top.c (inhibit_gdbtkinit): New variable.
	* top.h (inhibit_gdbtkinit): New variable.

doc/
	* gdb.texinfo (Choosing modes): Add -nx-except-gdbtkinit option.

gdbtk/
	* generic/gdbtk.c (gdbtk_init): Change inhibit_gdbinit to
	inhibit_gdbtkinit.

Index: src/gdb/main.c
===================================================================
--- src.orig/gdb/main.c	2005-11-17 15:20:03.000000000 +0000
+++ src/gdb/main.c	2005-11-17 15:22:12.000000000 +0000
@@ -267,8 +267,8 @@ captured_main (void *data)
       {"quiet", no_argument, &quiet, 1},
       {"q", no_argument, &quiet, 1},
       {"silent", no_argument, &quiet, 1},
-      {"nx", no_argument, &inhibit_gdbinit, 1},
-      {"n", no_argument, &inhibit_gdbinit, 1},
+      {"nx", no_argument, 0, 'n'},
+      {"n", no_argument, 0, 'n'},
       {"batch-silent", no_argument, 0, 'B'},
       {"batch", no_argument, &batch, 1},
       {"epoch", no_argument, &epoch_interface, 1},
@@ -296,6 +296,7 @@ captured_main (void *data)
       {"x", required_argument, 0, 'x'},
       {"ex", required_argument, 0, 'X'},
 #ifdef GDBTK
+      {"nx-except-gdbtkinit", no_argument, &inhibit_gdbinit, 1},
       {"tclcommand", required_argument, 0, 'z'},
       {"enable-external-editor", no_argument, 0, 'y'},
       {"editor-command", required_argument, 0, 'w'},
@@ -418,6 +419,10 @@ captured_main (void *data)
 	    batch = batch_silent = 1;
 	    gdb_stdout = ui_file_new();
 	    break;
+	  case 'n':
+	    inhibit_gdbinit = 1;
+	    inhibit_gdbtkinit = 1;
+	    break;
 #ifdef GDBTK
 	  case 'z':
 	    {
@@ -896,6 +901,8 @@ Options:\n\n\
   --nx               Do not read "), stream);
   fputs_unfiltered (gdbinit, stream);
   fputs_unfiltered (_(" file.\n\
+  --nx-except-gdbtkinit\n\
+                     As for --nx, but do read gdbtkinit for insight.\n\
   --quiet            Do not print version number on startup.\n\
   --readnow          Fully read symbol files on first access.\n\
 "), stream);
Index: src/gdb/top.c
===================================================================
--- src.orig/gdb/top.c	2005-11-17 15:20:03.000000000 +0000
+++ src/gdb/top.c	2005-11-17 16:15:44.000000000 +0000
@@ -86,6 +86,7 @@
 char gdbinit[PATH_MAX + 1] = GDBINIT_FILENAME;
 
 int inhibit_gdbinit = 0;
+int inhibit_gdbtkinit = 0;
 
 /* If nonzero, and GDB has been configured to be able to use windows,
    attempt to open them upon startup.  */
Index: src/gdb/top.h
===================================================================
--- src.orig/gdb/top.h	2005-11-17 15:20:03.000000000 +0000
+++ src/gdb/top.h	2005-11-17 15:22:12.000000000 +0000
@@ -29,6 +29,7 @@ extern int linesize;
 extern FILE *instream;
 extern char gdb_dirbuf[1024];
 extern int inhibit_gdbinit;
+extern int inhibit_gdbtkinit;
 extern int epoch_interface;
 extern char gdbinit[];
 
Index: src/gdb/gdbtk/generic/gdbtk.c
===================================================================
--- src.orig/gdb/gdbtk/generic/gdbtk.c	2005-11-17 15:20:03.000000000 +0000
+++ src/gdb/gdbtk/generic/gdbtk.c	2005-11-17 15:22:12.000000000 +0000
@@ -483,7 +483,7 @@ gdbtk_init (void)
 
   /* Set up some globals used by gdb to pass info to gdbtk
      for start up options and the like */
-  xasprintf (&s, "%d", inhibit_gdbinit);
+  xasprintf (&s, "%d", inhibit_gdbtkinit);
   Tcl_SetVar2 (gdbtk_interp, "GDBStartup", "inhibit_prefs", s, TCL_GLOBAL_ONLY);
   free(s);
    
Index: src/gdb/doc/gdb.texinfo
===================================================================
--- src.orig/gdb/doc/gdb.texinfo	2005-11-17 16:15:44.000000000 +0000
+++ src/gdb/doc/gdb.texinfo	2005-11-17 16:22:24.000000000 +0000
@@ -978,6 +978,12 @@ Do not execute commands found in any ini
 options and arguments have been processed.  @xref{Command Files,,Command
 files}.
 
+@item -nx-except-gdbtkinit
+@kindex -nx-except-gdbtkinit
+The same as @samp{-nx} with the difference that it does not prevent
+Insight from loading the settings file @code{.gdbtkinit}.  If you are not
+using Insight then this option is exactly the same as @samp{-nx}.
+
 @item -quiet
 @itemx -silent
 @itemx -q

Re: [PATCH] -nx-except-gdbtkinit option

[Eli Zaretskii]

> Date: Thu, 17 Nov 2005 16:31:05 +0000
> From: Andrew STUBBS <andrew.stubbs@st.com>
> Cc: gdb-patches@sources.redhat.com
> 
> > Second, please tell more about the situations when this option could
> > be useful.
> 
> The specific case I use it for is in a wrapper utility which calls gdb 
> as follows (roughly):
> 
>    gdb -nx-except-gdbtkinit -d scriptdir -x globalscript \
>          -x ~/.gdbinit -x ./.gdbinit .....
> 
> The purpose is (among other things) to set up transparently (to the 
> user) a selection of user-defined commands (for connecting and 
> configuring target boards) which are available to users in the .gdbinit 
> files.

Sorry, I'm still confused.  Doesn't GDB normally read ~/.gdbinit and
./.gdbinit?  If so, why do you need to use -nx and then load these
init files manually?

> If -nx is used it has undesirable side effects when using Insight.

In case it wasn't clear, it's those ``undesirable side effects'' are
what I was asking to describe in some detail.  I don't understand what
kind of problems it causes.  Perhaps those are misfeatures or bugs we
need to fix, instead of introducing a new option.

> +@item -nx-except-gdbtkinit
> +@kindex -nx-except-gdbtkinit
> +The same as @samp{-nx} with the difference that it does not prevent
> +Insight from loading the settings file @code{.gdbtkinit}.  If you are not
                                          ^^^^^
`.gdbtkinit' is a file name, so it should have the @file markup, not
@code.

Other than that, the documentation patch is approved, conditioned on
the code patch approval.  Thanks.

Re: [PATCH] -nx-except-gdbtkinit option

[Daniel Jacobowitz]

On Thu, Nov 17, 2005 at 09:33:17PM +0200, Eli Zaretskii wrote:
> >    gdb -nx-except-gdbtkinit -d scriptdir -x globalscript \
> >          -x ~/.gdbinit -x ./.gdbinit .....
> > 
> > The purpose is (among other things) to set up transparently (to the 
> > user) a selection of user-defined commands (for connecting and 
> > configuring target boards) which are available to users in the .gdbinit 
> > files.
> 
> Sorry, I'm still confused.  Doesn't GDB normally read ~/.gdbinit and
> ./.gdbinit?  If so, why do you need to use -nx and then load these
> init files manually?

In order for the global script to be loaded before .gdbinit, I think.

> > If -nx is used it has undesirable side effects when using Insight.
> 
> In case it wasn't clear, it's those ``undesirable side effects'' are
> what I was asking to describe in some detail.  I don't understand what
> kind of problems it causes.  Perhaps those are misfeatures or bugs we
> need to fix, instead of introducing a new option.

It sounds like there's no GDB command to process .gdbtkinit, which is a
TCL file rather than a GDB CLI script file.  Andrew, is that right?  If
so, have you considered adding such a command to Insight, instead?


-- 
Daniel Jacobowitz
CodeSourcery, LLC

Re: [PATCH] -nx-except-gdbtkinit option

[Andrew STUBBS]

Eli Zaretskii wrote:
>>   gdb -nx-except-gdbtkinit -d scriptdir -x globalscript \
>>         -x ~/.gdbinit -x ./.gdbinit .....
>>
>>The purpose is (among other things) to set up transparently (to the 
>>user) a selection of user-defined commands (for connecting and 
>>configuring target boards) which are available to users in the .gdbinit 
>>files.
> 
> 
> Sorry, I'm still confused.  Doesn't GDB normally read ~/.gdbinit and
> ./.gdbinit?  If so, why do you need to use -nx and then load these
> init files manually?

I want the .gdbinit files loaded after the mystical 'globalscript' file. 
Then they can use commands defined by that script.

>>If -nx is used it has undesirable side effects when using Insight.
> 
> In case it wasn't clear, it's those ``undesirable side effects'' are
> what I was asking to describe in some detail.  I don't understand what
> kind of problems it causes.  Perhaps those are misfeatures or bugs we
> need to fix, instead of introducing a new option.

Without this file Insight does not remember any of your settings, such 
as window size and position, target settings and breakpoints.

> `.gdbtkinit' is a file name, so it should have the @file markup, not
> @code.
> 
> Other than that, the documentation patch is approved, conditioned on
> the code patch approval.  Thanks.

Thanks

Andrew

Re: [PATCH] -nx-except-gdbtkinit option

[Andrew STUBBS]

Daniel Jacobowitz wrote:
> In order for the global script to be loaded before .gdbinit, I think.

Quite so.

> It sounds like there's no GDB command to process .gdbtkinit, which is a
> TCL file rather than a GDB CLI script file.  Andrew, is that right?  If
> so, have you considered adding such a command to Insight, instead?

It isn't my invention. Insight just loads it ... unless you say -nx.

Hmmm, it just occured to me that I probably ought to cross post the 
insight portion of this patch to the insight list. What's the usual form 
on this one?

Thanks

Andrew

Re: [PATCH] -nx-except-gdbtkinit option

[Andrew STUBBS]

Daniel Jacobowitz wrote:
> It sounds like there's no GDB command to process .gdbtkinit, which is a
> TCL file rather than a GDB CLI script file.  Andrew, is that right?  If
> so, have you considered adding such a command to Insight, instead?

Sorry, I think I understand what you are getting at now. Ignore my last 
response to this question.

.gdbtkinit is not a TCL file. It's rather more like a Microsoft .ini 
file (indeed on windows it is known as gdbtk.ini). Does that format have 
a name? I don't know.

I suppose it wouldn't be impossible to screw around with the tcl to get 
it to load on a command, although I don't know how right now. However, 
would Insight then save the file again, as it should, as the end of the 
session? Would this really make sense to a user?

Also, this file is used to define the initial state of the GUI (which 
windows are open and where and what size, as well as with what icons and 
fonts). It probably doesn't make sense to open this much after startup.

I much prefer it the way I have done it.

Andrew Stubbs

Re: [PATCH] -nx-except-gdbtkinit option

[Eli Zaretskii]

> Date: Fri, 18 Nov 2005 11:17:49 +0000
> From: Andrew STUBBS <andrew.stubbs@st.com>
> Cc: Eli Zaretskii <eliz@gnu.org>, gdb-patches@sources.redhat.com
> 
> .gdbtkinit is not a TCL file. It's rather more like a Microsoft .ini 
> file (indeed on windows it is known as gdbtk.ini). Does that format have 
> a name? I don't know.
> 
> I suppose it wouldn't be impossible to screw around with the tcl to get 
> it to load on a command, although I don't know how right now. However, 
> would Insight then save the file again, as it should, as the end of the 
> session? Would this really make sense to a user?
> 
> Also, this file is used to define the initial state of the GUI (which 
> windows are open and where and what size, as well as with what icons and 
> fonts). It probably doesn't make sense to open this much after startup.

Would it make sense to not inhibit its loading under -nx?  After all,
we don't inhibit loading ~/.inputrc under -nx, right?  Why should the
toolkit be different?

> I much prefer it the way I have done it.

Of course.  But you see, I don't like this new feature: I think it
will add to user's confusion.

Re: [PATCH] -nx-except-gdbtkinit option

[Andrew STUBBS]

Eli Zaretskii wrote:
> Would it make sense to not inhibit its loading under -nx?  After all,
> we don't inhibit loading ~/.inputrc under -nx, right?  Why should the
> toolkit be different?

That would work fine for me. I don't know why it is any different, but 
it is different and I didn't feel comfortable changing the default, let 
alone removing the functionality altogether. This is probably a 
discussion for the Insight list. I think I'll test the water.

>>I much prefer it the way I have done it.
> 
> Of course.  But you see, I don't like this new feature: I think it
> will add to user's confusion.

And my point was that I didn't like Daniel's suggestion for the exact 
same reason.

At the end of the day, all _I_ need is to be able to use -nx without 
breaking Insight. I don't really care which way, within reason, but I 
would prefer to keep it simple.

Thanks

Andrew Stubbs

Re: [PATCH] -nx-except-gdbtkinit option

[Daniel Jacobowitz]

On Fri, Nov 18, 2005 at 12:09:31PM +0000, Andrew STUBBS wrote:
> Eli Zaretskii wrote:
> >Would it make sense to not inhibit its loading under -nx?  After all,
> >we don't inhibit loading ~/.inputrc under -nx, right?  Why should the
> >toolkit be different?
> 
> That would work fine for me. I don't know why it is any different, but 
> it is different and I didn't feel comfortable changing the default, let 
> alone removing the functionality altogether. This is probably a 
> discussion for the Insight list. I think I'll test the water.

I like this suggestion better anyway.  If Insight is compiled in, have
it read its configuration.

-- 
Daniel Jacobowitz
CodeSourcery, LLC

Re: [PATCH] -nx-except-gdbtkinit option

[Andrew STUBBS]

Daniel Jacobowitz wrote:
>>>Would it make sense to not inhibit its loading under -nx?  After all,
>>>we don't inhibit loading ~/.inputrc under -nx, right?  Why should the
>>>toolkit be different?
>>
>>That would work fine for me. I don't know why it is any different, but 
>>it is different and I didn't feel comfortable changing the default, let 
>>alone removing the functionality altogether. This is probably a 
>>discussion for the Insight list. I think I'll test the water.
> 
> I like this suggestion better anyway.  If Insight is compiled in, have
> it read its configuration.

But the Insight guys didn't. However, they have accepted my patch here: 
http://sources.redhat.com/ml/insight/2005-q4/msg00037.html.

This solves the problem entirely in Insight by having an environment 
variable override the -nx option.

I don't know how this is for other people, but it solves the problem for me.

Please disregard my -nx-except-gdbtkinit patch from now on (not that you 
weren't going to anyway).

Thanks

Andrew Stubbs