Re: [PATCH]: update to df for mainline.

[Bernd Schmidt <bernds_cb1 at t-online dot de>]

Okay, I'm slowly looking through this stuff.  Bear with me, I'm not too 
familiar with this code yet.  (It's not conforming to our coding 
standards, at least as far as comments are concerned.)

Kenneth Zadeck wrote:

> > 2) Clobbers were also not being handled properly.  Clobbers in calls
> > were treated the same as clobbers just in the code.  The two cases are
> > now separated and dealt with properly.
>
> this is now in clobber.diff

Hmm.  There are several places that test DF_REF_CLOBBER, and you're 
modifying only one.

Wouldn't it be cleaner to generate special DF_REF_MAY_CLOBBER references 
 when scanning the insn, rather than having a test for CALL_INSN here? 
 That would show more clearly what's going on.

> > 1) Making more of the code use the abstractions that were defined in df.h.
>
> this is in lex.diff

This also deletes some comments, adds some good and one bad whitespace 
changes, makes a function non-static without adding uses, rearranges 
declarations for no reason, and changes an if into a switch.

The changes that you describe (making use of abstractions) are ok, and 
so is the if->switch change (as it appears to add a consistency check by 
introducing a default: label).

> > 3) Removal of unused parameters from some of the debugging functions.
> this is in debug.diff

This is ok as far as the parts compilable without reginfo.diff are 
concerned.

It doesn't look like you said what setstate.diff is for.

A cleaner split, and a more complete description for each change, would 
ease review.


Bernd