This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [tree-ssa] COND_EXPR lowering.

In message <>, Zdenek Dvorak wri
 >> > > > Is BIND_EXPR dependant in any way on SWITCH? ie, do we need to get
 >> > > > SWITCH and COND in before a BIND_EXPR patch has any hope of applying?
 >> > > 
 >> > > no, bind_exprs are completely independent issue.  I will try to create
 >> > > the SWITCH_EXPR lowering patch tomorrow (although the fact that
 >> > > COND_EXPR lowering is not in is complicating this; the two patches
 >> > > will neccesarily conflict with each other).
 >> > 
 >> > Hold off for a day or two then, we'll get COND_EXPR in shortly, then you
 >> > can produce the SWITCH patch..
 >> > 
 >> In fact, if your run bootstraps, check in COND_EXPR and produce the
 >> switch patch.
 >it does, except for a problem with tree-ssa-pre (the garbage collector
 >is freeing data allocated for dominance information).  The following
 >patch fixes it (it makes the data non-gc allocated).  This may be
 >reverted once split_critical_edges works; I will commit it for now as
 >	* dominance.c (BB_NODE): Use VARRAY_GENERIC_PTR_NOGC.
 >	(calculate_dominance_info): Use VARRAY_GENERIC_PTR_NOGC_INIT.
 >	* varray.c (element): Add GENERIC_PTR_NOGC entry.
 >	* varray.h (enum varray_data_enum): Add VARRAY_DATA_GENERIC_NOGC.
 >	(union varray_data_tag): Add generic_nogc.
Err, ummm, I thought Daniel fixed this in another way last week!

I certainly wouldn't consider this an "obvious" patch -- just look
at the discussion which led up to it -- anything with that long of
a discussion certainly doesn't qualify as obvious to me.

When you return, please look at what Daniel did and either remove your
patch if it has become redundant or explain why your patch is still


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]