[patch] gnu-v2-abi.c: strchr may return null

[patch] gnu-v2-abi.c: strchr may return null

[msnyder]

[-- Attachment #1: Type: text/plain, Size: 0 bytes --]



[-- Attachment #2: 112.txt --]
[-- Type: text/plain, Size: 1500 bytes --]

2007-08-29  Michael Snyder  <msnyder@access-company.com>

	* gnu-v2-abi.c (gnuv2_value_rtti_type): Guard against null.

Index: gnu-v2-abi.c
===================================================================
RCS file: /cvs/src/src/gdb/gnu-v2-abi.c,v
retrieving revision 1.25
diff -p -r1.25 gnu-v2-abi.c
*** gnu-v2-abi.c	23 Aug 2007 18:08:33 -0000	1.25
--- gnu-v2-abi.c	29 Aug 2007 21:24:09 -0000
*************** gnuv2_value_rtti_type (struct value *v, 
*** 192,198 ****
    CORE_ADDR vtbl;
    struct minimal_symbol *minsym;
    struct symbol *sym;
!   char *demangled_name;
    struct type *btype;
  
    if (full)
--- 192,198 ----
    CORE_ADDR vtbl;
    struct minimal_symbol *minsym;
    struct symbol *sym;
!   char *demangled_name, *p;
    struct type *btype;
  
    if (full)
*************** gnuv2_value_rtti_type (struct value *v, 
*** 252,258 ****
  
    /* If we just skip the prefix, we get screwed by namespaces */
    demangled_name=cplus_demangle(demangled_name,DMGL_PARAMS|DMGL_ANSI);
!   *(strchr(demangled_name,' '))=0;
  
    /* Lookup the type for the name */
    /* FIXME: chastain/2003-11-26: block=NULL is bogus.  See pr gdb/1465. */
--- 252,260 ----
  
    /* If we just skip the prefix, we get screwed by namespaces */
    demangled_name=cplus_demangle(demangled_name,DMGL_PARAMS|DMGL_ANSI);
!   p = strchr (demangled_name, ' ');
!   if (p)
!     *p = '\0';
  
    /* Lookup the type for the name */
    /* FIXME: chastain/2003-11-26: block=NULL is bogus.  See pr gdb/1465. */

Re: [patch] gnu-v2-abi.c: strchr may return null

[Joel Brobecker]

> 2007-08-29  Michael Snyder  <msnyder@access-company.com>
> 
> 	* gnu-v2-abi.c (gnuv2_value_rtti_type): Guard against null.

This looks OK to me, but can you give it a few more days so that Daniel
can provide any comment on it (C++ is not our main focus at AdaCore)?

I'm thinking this is perhaps a case where an assert might be better:
Does the rest of the function make any sense if we don't find our
space? I don't know the encoding so perhaps it's OK to give it a try...


> Index: gnu-v2-abi.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/gnu-v2-abi.c,v
> retrieving revision 1.25
> diff -p -r1.25 gnu-v2-abi.c
> *** gnu-v2-abi.c	23 Aug 2007 18:08:33 -0000	1.25
> --- gnu-v2-abi.c	29 Aug 2007 21:24:09 -0000
> *************** gnuv2_value_rtti_type (struct value *v, 
> *** 192,198 ****
>     CORE_ADDR vtbl;
>     struct minimal_symbol *minsym;
>     struct symbol *sym;
> !   char *demangled_name;
>     struct type *btype;
>   
>     if (full)
> --- 192,198 ----
>     CORE_ADDR vtbl;
>     struct minimal_symbol *minsym;
>     struct symbol *sym;
> !   char *demangled_name, *p;
>     struct type *btype;
>   
>     if (full)
> *************** gnuv2_value_rtti_type (struct value *v, 
> *** 252,258 ****
>   
>     /* If we just skip the prefix, we get screwed by namespaces */
>     demangled_name=cplus_demangle(demangled_name,DMGL_PARAMS|DMGL_ANSI);
> !   *(strchr(demangled_name,' '))=0;
>   
>     /* Lookup the type for the name */
>     /* FIXME: chastain/2003-11-26: block=NULL is bogus.  See pr gdb/1465. */
> --- 252,260 ----
>   
>     /* If we just skip the prefix, we get screwed by namespaces */
>     demangled_name=cplus_demangle(demangled_name,DMGL_PARAMS|DMGL_ANSI);
> !   p = strchr (demangled_name, ' ');
> !   if (p)
> !     *p = '\0';
>   
>     /* Lookup the type for the name */
>     /* FIXME: chastain/2003-11-26: block=NULL is bogus.  See pr gdb/1465. */

-- 
Joel

Re: [patch] gnu-v2-abi.c: strchr may return null

[Daniel Jacobowitz]

On Tue, Sep 04, 2007 at 07:18:30AM -0700, Joel Brobecker wrote:
> > 2007-08-29  Michael Snyder  <msnyder@access-company.com>
> > 
> > 	* gnu-v2-abi.c (gnuv2_value_rtti_type): Guard against null.
> 
> This looks OK to me, but can you give it a few more days so that Daniel
> can provide any comment on it (C++ is not our main focus at AdaCore)?

I've no objection - and I don't really care about this file; it only
affects g++ 2.95.

-- 
Daniel Jacobowitz
CodeSourcery

Re: [patch] gnu-v2-abi.c: strchr may return null

[msnyder]

>> 2007-08-29  Michael Snyder  <msnyder@access-company.com>
>>
>> 	* gnu-v2-abi.c (gnuv2_value_rtti_type): Guard against null.
>
> This looks OK to me, but can you give it a few more days so that Daniel
> can provide any comment on it (C++ is not our main focus at AdaCore)?

On Daniel's subsequent approval (or acquiescence?), committing.

> I'm thinking this is perhaps a case where an assert might be better:
> Does the rest of the function make any sense if we don't find our
> space? I don't know the encoding so perhaps it's OK to give it a try...

Well, it won't make it any worse.  If any reason manifests itself
to go with an assert, we can do that.