Re: [patch] ipa/ipcp partitioning

[Olga Golovanevsky <OLGA at il dot ibm dot com>]

Steven Bosscher <stevenb@suse.de> wrote on 17/05/2005 21:06:40:

> There are still a few places where you refer to IPCP, you probably want
> to fix up such names/comments, e.g.
>
> /* Creates the worklist for ipcp_iterate_stage().  */
> inline ipa_methodlist_p
> ipa_create_methodlist_node (void)
> {
>
> /* Frees the ipcp_method's IPCP data structures.  */
> void
> ipa_free (void)
> {
>

Fixed.

> You've also been a bit over-zealous with removing static.  Functions
> like ipa_create_methodlist_node are only used in ipa-prop.c (and that
> should probably stay that way) so they can (and should) still be
> static inline.
>

Fixed.

> The comment in the ipa-prop.c file about what the file provides is
> below par:
>
> /* This file contains interfaces common for different
>    interprocedural analyses.  */
>
> This is not very informative.  You should document the interfaces the
> provides here.  Define those interfaces, this is _so_ important for
> other people to understand what is going on!  (I believe we should later
> on also add an ipa.texi to the internals documentation, but... later ;-).

>

Fixed, yet without ipa.texi.

> Nit: The copyright years should also be fixed on ipa-prop.h at least.
>

Fixed.

Attached is a fixed versions of ipcp.c, ipc-prop.c, ipa-rpop.h files.
No changes were made to diff and ChangeLog:

http://gcc.gnu.org/ml/gcc-patches/2005-05/msg01687.html

ok for tree-profiling-branch?

olga

(See attached file: ipcp.c)(See attached file: ipa-prop.h)(See attached
file: ipa-prop.c)

Attachment: ipcp.c
Description: Binary data

Attachment: ipa-prop.h
Description: Binary data

Attachment: ipa-prop.c
Description: Binary data