[PATCH] pe/coff - add support for base64 encoded long section names

[Nick Clifton]

Hi Tristan,

> LLVM has added another convention for very large strtab, using '//xxxxxx' names and base64 encoding of the index in the strtab.

Coincidentally, this has also been reported in PR 30444...

> It has been a while since I haven't submitted a patch, I hope I am correctly following the procedure!

Oh you are. :-)

So - I have a couple of questions on the patch:

> +  name = (char *) bfd_alloc (abfd, (bfd_size_type) strlen (strings) + 1 + 1);
> +  if (name == NULL)
> +    return NULL;
> +  strcpy (name, strings);

Why "+1 +1" in the bfd_alloc ?  I understand one of them, but two ?

> +	  unsigned strindex;
> +	  unsigned i;
> +
> +	  strindex = 0;
> +	  for (i = 2; i < SCNNMLEN; i++)
> +	    {
> +	      char c = hdr->s_name[i];
> +	      unsigned d;
> +
> +	      if (c >= 'A' && c <= 'Z')
> +		d = c - 'A';
> +	      else if (c >= 'a' && c <= 'z')
> +		d = c - 'a' + 26;
> +	      else if (c >= '0' && c <= '9')
> +		d = c - '0' + 52;
> +	      else if (c == '+')
> +		d = 62;
> +	      else if (c == '/')
> +		d = 63;
> +	      else
> +		return false;
> +	      strindex = (strindex << 6) + d;
> +	    }

Is "unsigned" the right type for strindex ?  It seems to me that it might
be possible to encode really large numbers using base64.

Also - it might be nice to move this loop into a stand alone function so
that it can be used from other parts of the BFD library.  (Assuming that
the functionality requested in PR 30444 is implemented).

Cheers
   Nick