Various improvements regarding the way we print code code comments.

Various improvements regarding the way we print code code comments.

* All Lithium instructions have an associated Hydrogen instruction now,
  simplifying things.

* Consistently print <Lithium instruction number,Hydrogen value id> prefixes.

* Do not print uninteresting Lithium instructions like empty gaps, jumps to the
  next instruction, etc.

* Removed special handling of HChange-like instructions, it is totally unclear
  why they had this special treatment. If we really want to print more
  information about Lithium instructions, we should do it in a totally way,
  anyway (e.g. by unifying things with the generation of hydrogen*.cfg files).

* Made deferred code and the jump table stand out a little bit more.

* Print info about special blocks like loop headers and OSR entries.

Committed: http://code.google.com/p/v8/source/detail?r=14370

[Jakob Kummerow, 2013-04-22 09:19:59]

DBC (only looked at ia32 changes).

https://codereview.chromium.org/14371005/diff/1/src/ia32/lithium-codegen-ia32.cc
File src/ia32/lithium-codegen-ia32.cc (left):

https://codereview.chromium.org/14371005/diff/1/src/ia32/lithium-codegen-ia32...
src/ia32/lithium-codegen-ia32.cc:357: Comment(";;; @%d: %s. <of #%d %s for #%d
%s>",
I found it very helpful to have these comments. If you come up with a
replacement soon-ish, it's fine to land this as is, otherwise I'd favor
reverting this part of the change for now.

https://codereview.chromium.org/14371005/diff/1/src/ia32/lithium-codegen-ia32.cc
File src/ia32/lithium-codegen-ia32.cc (right):

https://codereview.chromium.org/14371005/diff/1/src/ia32/lithium-codegen-ia32...
src/ia32/lithium-codegen-ia32.cc:347: Comment(";;; <%d,%d> %s",
For grep-ability, I like having a '#' before the hydrogen ID. By the same
reasoning you may want to keep the '@' before the Lithium ID (personally I don't
care much about those), so we'd have ";;; <@%d, #%d> %s".

https://codereview.chromium.org/14371005/diff/1/src/ia32/lithium-ia32.cc
File src/ia32/lithium-ia32.cc (right):

https://codereview.chromium.org/14371005/diff/1/src/ia32/lithium-ia32.cc#newc...
src/ia32/lithium-ia32.cc:214: return !gen->IsNextEmittedBlock(block_id());
As discussed, to make sure this condition is in sync with when to emit an
instruction, there should be a helper method "MustEmitInstruction()" or similar
that's called in both places.

[Toon Verwaest, 2013-04-22 09:33:58]

lgtm

[Sven Panne, 2013-04-22 09:34:25]

Uploading a CL with the changes soon...

https://codereview.chromium.org/14371005/diff/1/src/ia32/lithium-codegen-ia32.cc
File src/ia32/lithium-codegen-ia32.cc (left):

https://codereview.chromium.org/14371005/diff/1/src/ia32/lithium-codegen-ia32...
src/ia32/lithium-codegen-ia32.cc:357: Comment(";;; @%d: %s. <of #%d %s for #%d
%s>",
On 2013/04/22 09:19:59, Jakob wrote:
> I found it very helpful to have these comments. If you come up with a
> replacement soon-ish, it's fine to land this as is, otherwise I'd favor
> reverting this part of the change for now.

As discussed offline, this can and should be handled differently: Sometimes we
need this info, sometimes we need to see the values of some flags, etc. Perhaps
we should just emit what --trace-hydrogen emits. I'll remove the stuff for now,
having the cruel hack here really hurts readability.

https://codereview.chromium.org/14371005/diff/1/src/ia32/lithium-codegen-ia32.cc
File src/ia32/lithium-codegen-ia32.cc (right):

https://codereview.chromium.org/14371005/diff/1/src/ia32/lithium-codegen-ia32...
src/ia32/lithium-codegen-ia32.cc:347: Comment(";;; <%d,%d> %s",
On 2013/04/22 09:19:59, Jakob wrote:
> For grep-ability, I like having a '#' before the hydrogen ID. By the same
> reasoning you may want to keep the '@' before the Lithium ID (personally I
don't
> care much about those), so we'd have ";;; <@%d, #%d> %s". 

Done.

https://codereview.chromium.org/14371005/diff/1/src/ia32/lithium-codegen-ia32...
src/ia32/lithium-codegen-ia32.cc:2196: int destination =
chunk_->LookupDestination(block);
Ooops, this should really use the IsNextEmittedBlock predicate to keep things in
sync, even if this means that we have to look through dead blocks again (which I
probably not a performance killer ;-).

[Sven Panne, 2013-04-22 09:48:48]

Committed patchset #2 manually as r14370 (presubmit successful).