[PATCH] Fix PR pascal/2283

[PATCH] Fix PR pascal/2283

[Pierre Muller]

  This patch fixes the display of char arrays
for pascal language as reported by Jonas Maebe
in the GDB bugs database.

ChangeLog entry:

2007-09-26  Pierre Muller  <muller@ics.u-strasbg.fr>

	* Fix PR pascal/2283
	p-valprint.c (pascal_val_print): correct
	current language check.
	Also print array of char as strings.


$ cvs diff -up p-valprint.c
Index: p-valprint.c
===================================================================
RCS file: /cvs/src/src/gdb/p-valprint.c,v
retrieving revision 1.51
diff -u -p -r1.51 p-valprint.c
--- p-valprint.c        7 Sep 2007 12:34:55 -0000       1.51
+++ p-valprint.c        26 Sep 2007 12:14:19 -0000
@@ -87,7 +87,7 @@ pascal_val_print (struct type *type, con
          /* For an array of chars, print with string syntax.  */
          if (eltlen == 1
              && ((TYPE_CODE (elttype) == TYPE_CODE_INT)
-              || ((current_language->la_language == language_m2)
+              || ((current_language->la_language == language_pascal)
                   && (TYPE_CODE (elttype) == TYPE_CODE_CHAR)))
              && (format == 0 || format == 's'))
            {
@@ -170,7 +170,8 @@ pascal_val_print (struct type *type, con
          /* For a pointer to char or unsigned char, also print the string
             pointed to, unless pointer is null.  */
          if (TYPE_LENGTH (elttype) == 1
-             && TYPE_CODE (elttype) == TYPE_CODE_INT
+             && (TYPE_CODE (elttype) == TYPE_CODE_INT
+                 || TYPE_CODE(elttype) == TYPE_CODE_CHAR)
              && (format == 0 || format == 's')
              && addr != 0)
            {

Pierre Muller
Chargé de recherches
Institut Charles Sadron
6, rue Boussingault
F 67083 Strasbourg Cedex
Tél. : +(33)3-88-41-40-07
Email : pierre.muller@ics.u-strasbg.fr

Re: [PATCH] Fix PR pascal/2283

[Eli Zaretskii]

> From: "Pierre Muller" <muller@ics.u-strasbg.fr>
> Cc: "'Jonas Maebe'" <jonas@freepascal.org>
> Date: Wed, 26 Sep 2007 14:25:38 +0200
> 
> -              || ((current_language->la_language == language_m2)
> +              || ((current_language->la_language == language_pascal)

Why is it TRT to remove Modula-2 here?

Re: [PATCH] Fix PR pascal/2283

[Jonas Maebe]

On 26 Sep 2007, at 15:49, Eli Zaretskii wrote:

>> From: "Pierre Muller" <muller@ics.u-strasbg.fr>
>> Cc: "'Jonas Maebe'" <jonas@freepascal.org>
>> Date: Wed, 26 Sep 2007 14:25:38 +0200
>>
>> -              || ((current_language->la_language == language_m2)
>> +              || ((current_language->la_language == language_pascal)
>
> Why is it TRT to remove Modula-2 here?

Because the that code resides in pascal_val_print() in p-valprint.c,  
and not in m2_val_print() in m2-valprint.c (the former was based on  
the latter, and the leftover language_m2 was simply a copy/paste  
error afaics).

In fact, I have no idea why that code (in both the Pascal and  
Modula-2 cases) even checks the current language, as those routines  
are obviously language-specific. Maybe the Modula-2 code originally  
came from a generic routine, and the check was never removed (and  
then copied over to the Pascal version when that one was written)?


Jonas

Re: [PATCH] Fix PR pascal/2283

[Eli Zaretskii]

> Cc: "Pierre Muller" <muller@ics.u-strasbg.fr>,
> 	gdb-patches@sourceware.org
> From: Jonas Maebe <jonas@freepascal.org>
> Date: Wed, 26 Sep 2007 16:03:13 +0200
> 
> In fact, I have no idea why that code (in both the Pascal and  
> Modula-2 cases) even checks the current language, as those routines  
> are obviously language-specific.

Exactly!  If this code is Pascal-specific, you shouldn't be checking
the language at all.

Re: [PATCH] Fix PR pascal/2283

[Jonas Maebe]

On 26 Sep 2007, at 16:07, Eli Zaretskii wrote:

>> Cc: "Pierre Muller" <muller@ics.u-strasbg.fr>,
>> 	gdb-patches@sourceware.org
>> From: Jonas Maebe <jonas@freepascal.org>
>> Date: Wed, 26 Sep 2007 16:03:13 +0200
>>
>> In fact, I have no idea why that code (in both the Pascal and
>> Modula-2 cases) even checks the current language, as those routines
>> are obviously language-specific.
>
> Exactly!  If this code is Pascal-specific, you shouldn't be checking
> the language at all.

In that case, the Modula-2 version shouldn't either. But I'm far too  
unfamiliar with the gdb source base to be able to assess whether it  
is in fact correct and safe to remove those checks.


Jonas

RE: [PATCH] Fix PR pascal/2283

[Pierre Muller]

 I think that the idea is that these functions
are exported and might be used by
other similar languages, but that this tiny
part is really specific and should not be
used in the case it is called by some
other language code.

  c_val_print is used in several other 
language specific XXX-lang.c files.
  If pascal_val_print gets used 
one day by some other language, the
behavior would be different for pascal and for XXX language.

  But I agree that this is all pretty
hypothetical for now.

  Should we remove it altogether?
It seems that it is the only location
inside pascal specific code
where such a test is performed.

  There are three references to
another language (language_cplus)
inside p-valprint, but this is because
pascal code has no demangler and the C++
demangler is used there.


Pierre
 


> -----Original Message-----
> From: gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] On Behalf Of Eli Zaretskii
> Sent: Wednesday, September 26, 2007 4:08 PM
> To: Jonas Maebe
> Cc: muller@ics.u-strasbg.fr; gdb-patches@sourceware.org
> Subject: Re: [PATCH] Fix PR pascal/2283
> 
> > Cc: "Pierre Muller" <muller@ics.u-strasbg.fr>,
> > 	gdb-patches@sourceware.org
> > From: Jonas Maebe <jonas@freepascal.org>
> > Date: Wed, 26 Sep 2007 16:03:13 +0200
> >
> > In fact, I have no idea why that code (in both the Pascal and
> > Modula-2 cases) even checks the current language, as those routines
> > are obviously language-specific.
> 
> Exactly!  If this code is Pascal-specific, you shouldn't be checking
> the language at all.