Add multibyte string support to debugger

Add multibyte string support to debugger

Issue 4110.

Had to extend the JSON protocol to support code points > 0xffff. The debugger client (editor) needs to support this new escape sequence'
format as well.
Committed: https://code.google.com/p/dart/source/detail?r=9751

[siva, 2012-07-18 17:11:31]

lgtm

http://codereview.chromium.org/10801002/diff/3002/lib/json/json.dart
File lib/json/json.dart (right):

http://codereview.chromium.org/10801002/diff/3002/lib/json/json.dart#newcode300
lib/json/json.dart:300: // Non-standard 32 bit characters of format \UHHHHHHHH.
What do you mean by non-standard here? not part of standard JSON?

http://codereview.chromium.org/10801002/diff/3002/runtime/bin/dbg_connection.cc
File runtime/bin/dbg_connection.cc (right):

http://codereview.chromium.org/10801002/diff/3002/runtime/bin/dbg_connection....
runtime/bin/dbg_connection.cc:334: res = Dart_StringGet32(str, codepoints,
&actual_len);
ASSERT_NOT_ERROR(res);

http://codereview.chromium.org/10801002/diff/3002/runtime/platform/json.cc
File runtime/platform/json.cc (right):

http://codereview.chromium.org/10801002/diff/3002/runtime/platform/json.cc#ne...
runtime/platform/json.cc:415: digit0 > 9 ? 'A' + (digit0 - 10) : '0' + digit0,
We have a Utils::IntToHexDigit(..) function for this.
Utils::IntToHexDigit(digit0), ......

[hausner, 2012-07-18 22:20:35]

PTAL.

I eliminated the non-standard \UHHHHHHHH character encoding in JSON. I am
encoding the JSON messages in UTF-8 instead.

http://codereview.chromium.org/10801002/diff/3002/lib/json/json.dart
File lib/json/json.dart (right):

http://codereview.chromium.org/10801002/diff/3002/lib/json/json.dart#newcode300
lib/json/json.dart:300: // Non-standard 32 bit characters of format \UHHHHHHHH.
On 2012/07/18 17:11:31, asiva wrote:
> What do you mean by non-standard here? not part of standard JSON?

I reverted the changes in this file, they are no longer needed.

http://codereview.chromium.org/10801002/diff/3002/runtime/bin/dbg_connection.cc
File runtime/bin/dbg_connection.cc (right):

http://codereview.chromium.org/10801002/diff/3002/runtime/bin/dbg_connection....
runtime/bin/dbg_connection.cc:334: res = Dart_StringGet32(str, codepoints,
&actual_len);
On 2012/07/18 17:11:31, asiva wrote:
> ASSERT_NOT_ERROR(res);

Done.

http://codereview.chromium.org/10801002/diff/3002/runtime/platform/json.cc
File runtime/platform/json.cc (right):

http://codereview.chromium.org/10801002/diff/3002/runtime/platform/json.cc#ne...
runtime/platform/json.cc:415: digit0 > 9 ? 'A' + (digit0 - 10) : '0' + digit0,
On 2012/07/18 17:11:31, asiva wrote:
> We have a Utils::IntToHexDigit(..) function for this.
> Utils::IntToHexDigit(digit0), ......
Ok, but that function contains an assert that the digit value is < 16. I don't
want to execute that assert for each digit when i know that the input is fine.
(We put asserts too low in the call graph...)

[hausner, 2012-07-18 23:28:49]

I turn the PTAL into a TBRA(gain) :-)

Thanks!

[siva, 2012-07-18 23:52:24]

lgtm

http://codereview.chromium.org/10801002/diff/7002/runtime/platform/json.cc
File runtime/platform/json.cc (right):

http://codereview.chromium.org/10801002/diff/7002/runtime/platform/json.cc#ne...
runtime/platform/json.cc:365: EnsureCapacity(sizeof(ch));
why do you have two calls to EnsureCapacity, one here and one inside every if
...., isn't he one inside every if .. sufficient?

http://codereview.chromium.org/10801002/diff/7002/runtime/platform/json.cc#ne...
runtime/platform/json.cc:465: intptr_t new_size = buf_size_ + len +
kBufferSpareCapacity;
you may want to check with Todd on this, he has been working on a set of changes
to guard against overflows in malloc/realloc sizes.

[hausner, 2012-07-19 00:11:03]

Thanks!

http://codereview.chromium.org/10801002/diff/7002/runtime/platform/json.cc
File runtime/platform/json.cc (right):

http://codereview.chromium.org/10801002/diff/7002/runtime/platform/json.cc#ne...
runtime/platform/json.cc:365: EnsureCapacity(sizeof(ch));
On 2012/07/18 23:52:24, asiva wrote:
> why do you have two calls to EnsureCapacity, one here and one inside every if
> ...., isn't he one inside every if .. sufficient?
Oops, good catch. This is a left-over from a previous version.

http://codereview.chromium.org/10801002/diff/7002/runtime/platform/json.cc#ne...
runtime/platform/json.cc:465: intptr_t new_size = buf_size_ + len +
kBufferSpareCapacity;
On 2012/07/18 23:52:24, asiva wrote:
> you may want to check with Todd on this, he has been working on a set of
changes
> to guard against overflows in malloc/realloc sizes.

Filing a bug and adding a todo.