Re: [PATCH] GDB: Fix detection of ELF support when configuring with -Werror

[Simon Marchi <simark@simark.ca>]

On 2020-11-13 7:09 a.m., Nick Clifton wrote:
> Hi Simon,
>
>>> that's what I originally planned, but it seems like elf-bfd.h (and the
>>> headers it includes) don't include any system headers. Since I'm not
>>> familiar with any of this code I assumed this was intentional.
>>>
>>> Alex
>>>
>>
>> Hi binutils@,
>>
>> Could you check the discussion above?  Is there a reason elf-bfd.h
>> doesn't include the header file it needs to use the functions it uses?
>
> Essentially this is because elf-bfd.h is internal to the binutils
> sources, and not expected to be used elsewhere.  So any code that
> includes it is also expected to include the sysdep.h header which
> does then include the needed system headers.
>
> The idea is that all of the configuration time decisions about which
> system headers to include are confined to one file - sysdep.h - rather
> than having to be copied into all header files.
>
> Cheers
>   Nick
>
>

Ah ok I remember this, we started discussing this in:

https://sourceware.org/pipermail/gdb-patches/2020-September/172041.html

In an ideal world, GDB would stop using BFD's internal functions.  But
for this, BFD would need to expose some ELF-specific bits, such as
program headers.  I don't know if this is against BFD's design
principles to abstract things across various executable formats.

I think Alexander's patch is fine, I don't think the complexity of the
patch linked above is necessary.  Are there even any systems today that
GDB supports for which strncmp isn't found in string.h?

Alexander, may I ask what particular configuration you are using?  It
would be good to document it in the commit log.

Simon