[PATCH,take 2,trunk+2.21.1] Re: Fix link order problem with LD plugin API.

[Dave Korn]

On 03/02/2011 05:11, Alan Modra wrote:
> On Mon, Jan 31, 2011 at 02:23:27AM +0000, Dave Korn wrote:
>> 	* ldlang.h (lang_input_statement_type): Add new 'claim_archive' flag.
>> 	* ldmain.c (add_archive_element): Set it if the member is claimed.
>> 	* ldlang.c (new_afile): Initialise claim_archive and claimed members.
>> 	(find_replacements_insert_point): New helper function.
>> 	(lang_process): After adding and opening replacement files passed
>> 	from plugin, splice them into correct place in statement list and
>> 	file chains to preserve critical link order.
>> 	(lang_list_insert_after): New helper function.
>> 	(lang_list_remove_tail): Likewise.
> 
> OK.
> 
>> +  /* Are we adding at the very end of the list?  */
>> +  if (*field == NULL)
>> +    {
>> +      /* (*field == NULL) should imply (destlist->tail == field),
>> +	  if not then the element isn't really in the DESTLIST.  */
>> +      ASSERT (destlist->tail == field);
>> +      /* Yes, append to and update tail pointer.  */
>> +      *(destlist->tail) = srclist->head;
>> +      destlist->tail = srclist->tail;
>> +    }
>> +  else
>> +    {
>> +      /* We're inserting in the middle somewhere.  */
>> +      *(srclist->tail) = *field;
>> +      *field = srclist->head;
>> +    }
> 
> Maybe this instead?
> 
>   *(srclist->tail) = *field;
>   *field = srclist->head;
>   if (destlist->tail == field)
>     destlist->tail = srclist->tail;

  Seems like a good plan.  I notice that I omitted #ifdef ENABLE_PLUGINS
guards around those new functions, which, being static, will cause unused
warnings.  Respin shortly, along with a series of other patches (in
particular, something to get rid of those stray left-over symbols when the IR
symtabs in the input files mention something that ltrans subsequently decides
it can optimise away.)

    cheers,
      DaveK