[RFA] pei386 dll: auto-import patch

[DJ Delorie]

> +@cindex DLLs, creating
> +@kindex --out-implib
> +@item --out-implib @var{file}
> +
> +@cindex DLLs, creating
> +@kindex --enable-auto-image-base
> +@item --enable-auto-image-base
> +
> +@cindex DLLs, creating
> +@kindex --disable-auto-image-base
> +@item --disable-auto-image-base
> +
> +@cindex DLLs, linking to
> +@kindex --dll-search-prefix
> +@item --dll-search-prefix @var{string}
> +
> +@cindex DLLs, linking to
> +@kindex --enable-auto-import
> +@item --enable-auto-import
> +
> +@cindex DLLs, linking to
> +@kindex --disable-auto-import
> +@item --disable-auto-import

You should only need to do the cindex for the first of each set, yes?

> +typedef struct {
> +  char *name;
> +  int len;
> +} autofilter_entry_type;

I wonder if we don't need the length, because it can be computed on
the fly?  Not a big deal, though - it would only really help
maintainers.

> +#if 0
> +      /* Don't export any 'reserved' symbols */
> +      if (*n && *n == '_' && n[1] == '_')
>  	return 0;
> +#endif

Rather than leave in commented out code, you should simply add a
textual comment explaining what to avoid (if such a comment is
warranted) or simply remove the code.  Not a big deal, though.

> +char *
> +make_import_fixup_mark (rel)

This should be static, or be named pe_*

> +  static char fixup_name[300];

300 is a wonderful number, but dynamically allocating space is better.
Compromise - keep a static pointer to a malloc'd buffer that grows as
needed.

> +  if (pe_dll_extra_pe_debug) 
> +    {
> +      printf (__FUNCTION__"\n");
> +    }

Hmmm... what happens when the function name has a percent in it?  (not
likely in C, but I'm the paranoid type...)

> +char *data_import_dll;

Should this be static?