Re: [PATCH 01/30] Changes to machine independent code

[Richard Earnshaw <Richard.Earnshaw@foss.arm.com>]

On 26/06/2019 09:36, Richard Sandiford wrote:
> Jeff Law <law@redhat.com> writes:
>> On 6/25/19 2:22 PM, acsawdey@linux.ibm.com wrote:
>>> From: Aaron Sawdey <acsawdey@linux.ibm.com>
>>>
>>> 	* builtins.c (get_memory_rtx): Fix comment.
>>> 	* optabs.def (movmem_optab): Change to cpymem_optab.
>>> 	* expr.c (emit_block_move_via_cpymem): Change movmem to cpymem.
>>> 	(emit_block_move_hints): Change movmem to cpymem.
>>> 	* defaults.h: Change movmem to cpymem.
>>> 	* targhooks.c (get_move_ratio): Change movmem to cpymem.
>>> 	(default_use_by_pieces_infrastructure_p): Ditto.
>> So I think you're missing an update to the RTL/MD documentation.  This
>> is also likely to cause problems for any out-of-tree ports, so it's
>> probably worth a mention in the gcc-10 changes, which will need to be
>> created (in CVS no less, ugh).
> 
> Mentioning changes like this could give a false sense that the release
> notes are the best place to look.  I can think of quite a few changes
> in this line that don't get mentioned in release notes. :-)
> 
> Diffing the texi files is probably more reliable (but would still miss
> things like Wilco's recent buitin_setjmp/longjmp change, which could
> also be relevant to out-of-tree ports).
> 
> Richard
> 

Simply renaming an API is generally fine.  Out of tree ports will fail
to build and the fixes are generally simple at that point.

Adding a new API with the old name is where the problem usually lies,
since now the OOT port will still build but have changed semantics.

R.