[RFC] MinGW and i386_pe_skip_trampoline_code

[RFC] MinGW and i386_pe_skip_trampoline_code

[Pedro Alves]

Hi all,

Just looking at PE support code, and I noticed that in i386-cygwin-tdep.c,
there is a call to set_gdbarch_skip_trampoline_code that will eventually 
register
i386_pe_skip_trampoline_code, defined in i386-tdep.c.
 From the looks of it, that function will skip 'import lib 
stubs'/trampolines.
I wonder why gdb for MinGW doesn't register it too? Probably an 
oversight, no?

Cheers,
Pedro Alves

Re: [RFC] MinGW and i386_pe_skip_trampoline_code

[Pedro Alves]

Pedro Alves wrote:
> Hi all,
>
> Just looking at PE support code, and I noticed that in 
> i386-cygwin-tdep.c,
> there is a call to set_gdbarch_skip_trampoline_code that will 
> eventually register
> i386_pe_skip_trampoline_code, defined in i386-tdep.c.
> From the looks of it, that function will skip 'import lib 
> stubs'/trampolines.
> I wonder why gdb for MinGW doesn't register it too? Probably an 
> oversight, no?
>
Err, maybe it is the other way around, and find_solib_trampoline_target 
should be enough for cygwin too?
Probably not important, though.

> Cheers,
> Pedro Alves
>
>

Re: [RFC] MinGW and i386_pe_skip_trampoline_code

[Daniel Jacobowitz]

On Wed, Nov 08, 2006 at 12:32:35AM +0000, Pedro Alves wrote:
> Hi all,
> 
> Just looking at PE support code, and I noticed that in i386-cygwin-tdep.c,
> there is a call to set_gdbarch_skip_trampoline_code that will eventually 
> register
> i386_pe_skip_trampoline_code, defined in i386-tdep.c.
> From the looks of it, that function will skip 'import lib 
> stubs'/trampolines.
> I wonder why gdb for MinGW doesn't register it too? Probably an 
> oversight, no?

The FSF tree doesn't even support native GDB on MinGW, to the best of
my knowledge.  The MinGW people have local patches that I don't believe
have ever been submitted (this is hearsay, I've never looked myself).

-- 
Daniel Jacobowitz
CodeSourcery

Re: [RFC] MinGW and i386_pe_skip_trampoline_code

[Pedro Alves]

Daniel Jacobowitz wrote:
> On Wed, Nov 08, 2006 at 12:32:35AM +0000, Pedro Alves wrote:
>   
>> Hi all,
>>
>> Just looking at PE support code, and I noticed that in i386-cygwin-tdep.c,
>> there is a call to set_gdbarch_skip_trampoline_code that will eventually 
>> register
>> i386_pe_skip_trampoline_code, defined in i386-tdep.c.
>> From the looks of it, that function will skip 'import lib 
>> stubs'/trampolines.
>> I wonder why gdb for MinGW doesn't register it too? Probably an 
>> oversight, no?
>>     
>
> The FSF tree doesn't even support native GDB on MinGW, to the best of
> my knowledge.  The MinGW people have local patches that I don't believe
> have ever been submitted (this is hearsay, I've never looked myself).
>
>   

Ah ok. MinGW is handled on gdb/configure.tgt, so I thought it was supported.
I've never looked into MinGW's gdb sources either.

Cheers,
Pedro Alves

Re: [RFC] MinGW and i386_pe_skip_trampoline_code

[Eli Zaretskii]

> Date: Tue, 7 Nov 2006 20:57:52 -0500
> From: Daniel Jacobowitz <drow@false.org>
> Cc: gdb@sourceware.org
> 
> The FSF tree doesn't even support native GDB on MinGW, to the best of
> my knowledge.  The MinGW people have local patches that I don't believe
> have ever been submitted (this is hearsay, I've never looked myself).

Yes, you are right.  I'm not one of the ``MinGW people'' who maintain
the native GDB, but I've seen the patches, and they aren't in the
official sources, for reasons that evade me.

Re: [RFC] MinGW and i386_pe_skip_trampoline_code

[Christopher Faylor]

On Wed, Nov 08, 2006 at 06:19:12AM +0200, Eli Zaretskii wrote:
>>Date: Tue, 7 Nov 2006 20:57:52 -0500
>>From: Daniel Jacobowitz
>>
>>The FSF tree doesn't even support native GDB on MinGW, to the best of
>>my knowledge.  The MinGW people have local patches that I don't believe
>>have ever been submitted (this is hearsay, I've never looked myself).
>
>Yes, you are right.  I'm not one of the ``MinGW people'' who maintain
>the native GDB, but I've seen the patches, and they aren't in the
>official sources, for reasons that evade me.

Isn't the ownership of the patches questionable so there is no clear way
to assign them to the FSF?

cgf

Re: [RFC] MinGW and i386_pe_skip_trampoline_code

[Eli Zaretskii]

> Date: Wed, 8 Nov 2006 13:39:36 -0500
> From: Christopher Faylor <cgf-use-the-mailinglist-please@sourceware.org>
> >
> >Yes, you are right.  I'm not one of the ``MinGW people'' who maintain
> >the native GDB, but I've seen the patches, and they aren't in the
> >official sources, for reasons that evade me.
> 
> Isn't the ownership of the patches questionable so there is no clear way
> to assign them to the FSF?

Maybe, I simply don't know.