Subzero: Shorten and normalize strings in non-DUMP builds.

Subzero: Shorten and normalize strings in non-DUMP builds.

Jump table labels change from ".Lxxxxx_yy" to "$Jxxxxx_yy".

Pooled float labels change from ".L$float$xxxxyyyy" to "$Fxxxxyyyy".

Pooled double labels change from ".L$double$xxxxxxxxyyyyyyyy" to "$Dxxxxxxxxyyyyyyyy".

All these should in theory not conflict with C/C++ user symbols.

Float labels now likely fit into short strings and don't require extra memory allocation.

Double labels may exceed the limit - encoding in base64 instead of hex would fix that.  Or, directly use the binary bytes as the string value.

BUG= none
R=jpp@chromium.org

Committed: https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-subzero.git;a=commit;h=e922c23632172e4cb5629c702b77959fe18cc686

[Jim Stichnoth, 2016-04-08 00:47:22]

Description was changed from

==========
Subzero: Shorten and normalize strings in non-DUMP builds.

BUG= none
==========

to

==========
Subzero: Shorten and normalize strings in non-DUMP builds.

Jump table labels change from ".Lxxxxx_yy" to "_Jxxxxx_yy".

Pooled float labels change from ".L$float$xxxxyyyy" to "_Fxxxxyyyy".

Pooled double labels change from ".L$double$xxxxxxxxyyyyyyyy" to
"_Dxxxxxxxxyyyyyyyy".

All these should in theory not conflict with C/C++ user symbols, which may not
start with underscore plus uppercase letter.

Float labels now likely fit into short strings and don't require extra memory
allocation.

Double labels may exceed the limit - encoding in base64 instead of hex would fix
that.

BUG= none
==========

[Jim Stichnoth, 2016-04-08 00:47:23]

stichnot@chromium.org changed reviewers:
+ eholk@chromium.org, jpp@chromium.org, kschimpf@google.com, sehr@chromium.org

[Jim Stichnoth, 2016-04-08 00:49:01]

This seems to get a couple percent gain on the sandboxed translator with
threads=4.

[John, 2016-04-08 15:47:46]

lgtm

https://codereview.chromium.org/1868113002/diff/1/src/IceInst.cpp
File src/IceInst.cpp (right):

https://codereview.chromium.org/1868113002/diff/1/src/IceInst.cpp#newcode582
src/IceInst.cpp:582: Ctx, "_J" + std::to_string(FuncName.getID()) + "_" +
std::to_string(Id));
why not $J? that would pretty much guarantee no clashes with C(++) identifiers.

[Jim Stichnoth, 2016-04-09 15:53:56]

Description was changed from

==========
Subzero: Shorten and normalize strings in non-DUMP builds.

Jump table labels change from ".Lxxxxx_yy" to "_Jxxxxx_yy".

Pooled float labels change from ".L$float$xxxxyyyy" to "_Fxxxxyyyy".

Pooled double labels change from ".L$double$xxxxxxxxyyyyyyyy" to
"_Dxxxxxxxxyyyyyyyy".

All these should in theory not conflict with C/C++ user symbols, which may not
start with underscore plus uppercase letter.

Float labels now likely fit into short strings and don't require extra memory
allocation.

Double labels may exceed the limit - encoding in base64 instead of hex would fix
that.

BUG= none
==========

to

==========
Subzero: Shorten and normalize strings in non-DUMP builds.

Jump table labels change from ".Lxxxxx_yy" to "$Jxxxxx_yy".

Pooled float labels change from ".L$float$xxxxyyyy" to "$Fxxxxyyyy".

Pooled double labels change from ".L$double$xxxxxxxxyyyyyyyy" to
"$Dxxxxxxxxyyyyyyyy".

All these should in theory not conflict with C/C++ user symbols.

Float labels now likely fit into short strings and don't require extra memory
allocation.

Double labels may exceed the limit - encoding in base64 instead of hex would fix
that.  Or, directly use the binary bytes as the string value.

BUG= none
==========

[Jim Stichnoth, 2016-04-09 15:54:09]

https://codereview.chromium.org/1868113002/diff/1/src/IceInst.cpp
File src/IceInst.cpp (right):

https://codereview.chromium.org/1868113002/diff/1/src/IceInst.cpp#newcode582
src/IceInst.cpp:582: Ctx, "_J" + std::to_string(FuncName.getID()) + "_" +
std::to_string(Id));
On 2016/04/08 15:47:46, John wrote:
> why not $J? that would pretty much guarantee no clashes with C(++)
identifiers.

Done.

[Jim Stichnoth, 2016-04-09 15:54:23]

Description was changed from

==========
Subzero: Shorten and normalize strings in non-DUMP builds.

Jump table labels change from ".Lxxxxx_yy" to "$Jxxxxx_yy".

Pooled float labels change from ".L$float$xxxxyyyy" to "$Fxxxxyyyy".

Pooled double labels change from ".L$double$xxxxxxxxyyyyyyyy" to
"$Dxxxxxxxxyyyyyyyy".

All these should in theory not conflict with C/C++ user symbols.

Float labels now likely fit into short strings and don't require extra memory
allocation.

Double labels may exceed the limit - encoding in base64 instead of hex would fix
that.  Or, directly use the binary bytes as the string value.

BUG= none
==========

to

==========
Subzero: Shorten and normalize strings in non-DUMP builds.

Jump table labels change from ".Lxxxxx_yy" to "$Jxxxxx_yy".

Pooled float labels change from ".L$float$xxxxyyyy" to "$Fxxxxyyyy".

Pooled double labels change from ".L$double$xxxxxxxxyyyyyyyy" to
"$Dxxxxxxxxyyyyyyyy".

All these should in theory not conflict with C/C++ user symbols.

Float labels now likely fit into short strings and don't require extra memory
allocation.

Double labels may exceed the limit - encoding in base64 instead of hex would fix
that.  Or, directly use the binary bytes as the string value.

BUG= none
R=jpp@chromium.org

Committed:
https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-subzero.git;a...
==========

[Jim Stichnoth, 2016-04-09 15:54:25]

Committed patchset #2 (id:20001) manually as
e922c23632172e4cb5629c702b77959fe18cc686 (presubmit successful).