Strict mode eval/arguments LHS.

Strict mode eval/arguments LHS.


BUG=
TEST=

[Martin Maly, 2011-01-25 23:19:17]

Next installment of strict mode.

[Lasse Reichstein, 2011-01-26 08:03:59]

LGTM.

http://codereview.chromium.org/6335013/diff/1/src/messages.js
File src/messages.js (right):

http://codereview.chromium.org/6335013/diff/1/src/messages.js#newcode240
src/messages.js:240: strict_lhs_postfix:           "Operand of postfix
expression may not be eval or arguments in strict mode",
Unless this is what Webkit uses as error message (we should generally match
those), it should say that it's only increment/decrement postfix and prefix
operators that must not be eval or arguments (i.e., when the operand must be an
LValue), and not, e.g., the "!" operator.

http://codereview.chromium.org/6335013/diff/1/src/parser.h
File src/parser.h (right):

http://codereview.chromium.org/6335013/diff/1/src/parser.h#newcode617
src/parser.h:617: void CheckStrictModeLHSName(Expression* expression,
I prefer LValue to LHS. We *try*, without being completely successful, to avoid
abbreviations in names, except where the abbreviation is the common way to refer
to the concept (e.g., JSON, ASCII, etc.) 
To me LValue is a canonical name, but LHS is just an abbreviation. That's
obviously purely subjective :)

http://codereview.chromium.org/6335013/diff/1/test/mjsunit/strict-mode.js
File test/mjsunit/strict-mode.js (right):

http://codereview.chromium.org/6335013/diff/1/test/mjsunit/strict-mode.js#new...
test/mjsunit/strict-mode.js:187: 
Try a compound assignment operator as well, e.g., "-=".

http://codereview.chromium.org/6335013/diff/1/test/mjsunit/strict-mode.js#new...
test/mjsunit/strict-mode.js:218: CheckStrictMode("function strict() { var x =
--arguments; }", SyntaxError)
Check that not-lvalue-operand operators doesn't throw, e.g., "!" or "-".

[Martin Maly, 2011-01-26 19:19:33]

incorporated all feedback, thanks for the review, Lasse!

Martin

http://codereview.chromium.org/6335013/diff/1/src/messages.js
File src/messages.js (right):

http://codereview.chromium.org/6335013/diff/1/src/messages.js#newcode240
src/messages.js:240: strict_lhs_postfix:           "Operand of postfix
expression may not be eval or arguments in strict mode",
On 2011/01/26 08:03:59, Lasse Reichstein wrote:
> Unless this is what Webkit uses as error message (we should generally match
> those), it should say that it's only increment/decrement postfix and prefix
> operators that must not be eval or arguments (i.e., when the operand must be
an
> LValue), and not, e.g., the "!" operator.

Webkit only says "parse error". Updated the errors to refer to incr/decr only. I
still have the bug open to eventually go through all the error messages and
reconcile them with jsc and mozilla.

http://codereview.chromium.org/6335013/diff/1/src/parser.h
File src/parser.h (right):

http://codereview.chromium.org/6335013/diff/1/src/parser.h#newcode617
src/parser.h:617: void CheckStrictModeLHSName(Expression* expression,
On 2011/01/26 08:03:59, Lasse Reichstein wrote:
> I prefer LValue to LHS. We *try*, without being completely successful, to
avoid
> abbreviations in names, except where the abbreviation is the common way to
refer
> to the concept (e.g., JSON, ASCII, etc.) 
> To me LValue is a canonical name, but LHS is just an abbreviation. That's
> obviously purely subjective :)

Renamed to CheckStrictModeLValue

http://codereview.chromium.org/6335013/diff/1/test/mjsunit/strict-mode.js
File test/mjsunit/strict-mode.js (right):

http://codereview.chromium.org/6335013/diff/1/test/mjsunit/strict-mode.js#new...
test/mjsunit/strict-mode.js:187: 
On 2011/01/26 08:03:59, Lasse Reichstein wrote:
> Try a compound assignment operator as well, e.g., "-=".

Done.

http://codereview.chromium.org/6335013/diff/1/test/mjsunit/strict-mode.js#new...
test/mjsunit/strict-mode.js:218: CheckStrictMode("function strict() { var x =
--arguments; }", SyntaxError)
On 2011/01/26 08:03:59, Lasse Reichstein wrote:
> Check that not-lvalue-operand operators doesn't throw, e.g., "!" or "-".

Done.