Allow enum values in IDL files to be simple arithmetic expressions.

Allow enums values to be simple arithmetic expressions.
BUG=none
TEST=./generator.py test_parser/enum.idl
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=104588

[noelallen1, 2011-10-05 23:35:51]

You'll also want to modify:
  test_lex.in
  test_paser

To add unit tests for the new expressions.

http://codereview.chromium.org/8161006/diff/1/ppapi/generators/idl_parser.py
File ppapi/generators/idl_parser.py (right):

http://codereview.chromium.org/8161006/diff/1/ppapi/generators/idl_parser.py#...
ppapi/generators/idl_parser.py:449: if p[2] == '+'  : p[0] = p[1] + p[3]
I think what you really want to do, is take a bunch of strings, and concatenate
them. For example, if someone passes a symbol you can't do str(SYMBOL) + int(5).
 
Maybe replace this all with:
 p[0] = "(%s %s %s)" % (str(p[1]), str(p[2]), str(p[3]))

http://codereview.chromium.org/8161006/diff/1/ppapi/generators/idl_parser.py#...
ppapi/generators/idl_parser.py:474: p[0] = p[1]
only allow INT, HEX, OCT?  No float

[bbudge, 2011-10-06 18:29:30]

Added some tests, though not sure if I'm testing the lexer. I added some bitwise
operators.

[noelallen1, 2011-10-06 20:51:43]

Okay minor issue wrt unary ops.
The tests you added will verify that the code was correctly tokenized and
parsed.  To verify it generates correctly you'll want to add to:
  test_cgen/enum_typedef.h and test_cgen/enum_typedef.idl

This is the test that would have caught the extra space in the unary operator.

http://codereview.chromium.org/8161006/diff/5001/ppapi/generators/idl_parser.py
File ppapi/generators/idl_parser.py (right):

http://codereview.chromium.org/8161006/diff/5001/ppapi/generators/idl_parser....
ppapi/generators/idl_parser.py:426: )
prec for |, &, ^?

http://codereview.chromium.org/8161006/diff/5001/ppapi/generators/idl_parser....
ppapi/generators/idl_parser.py:444: p[0] = "%s %s" % (str(p[1]), str(p[2]))
Shouldn't unary ops be  %s%s such as '-' + '5' instead of '-' + ' ' + '5'

http://codereview.chromium.org/8161006/diff/5001/ppapi/generators/test_parser...
File ppapi/generators/test_parser/enum.idl (right):

http://codereview.chromium.org/8161006/diff/5001/ppapi/generators/test_parser...
ppapi/generators/test_parser/enum.idl:68: Es9_1 = 0,
Should mark these as /* OK EnumItem(Es9_1) */ ...

[bbudge, 2011-10-07 00:04:52]

http://codereview.chromium.org/8161006/diff/5001/ppapi/generators/idl_parser.py
File ppapi/generators/idl_parser.py (right):

http://codereview.chromium.org/8161006/diff/5001/ppapi/generators/idl_parser....
ppapi/generators/idl_parser.py:426: )
On 2011/10/06 20:51:44, noelallen1 wrote:
> prec for |, &, ^?

Done.

http://codereview.chromium.org/8161006/diff/5001/ppapi/generators/idl_parser....
ppapi/generators/idl_parser.py:444: p[0] = "%s %s" % (str(p[1]), str(p[2]))
On 2011/10/06 20:51:44, noelallen1 wrote:
> Shouldn't unary ops be  %s%s such as '-' + '5' instead of '-' + ' ' + '5'

Done.

http://codereview.chromium.org/8161006/diff/5001/ppapi/generators/test_parser...
File ppapi/generators/test_parser/enum.idl (right):

http://codereview.chromium.org/8161006/diff/5001/ppapi/generators/test_parser...
ppapi/generators/test_parser/enum.idl:68: Es9_1 = 0,
On 2011/10/06 20:51:44, noelallen1 wrote:
> Should mark these as /* OK EnumItem(Es9_1) */ ...

Done.

[noelallen1, 2011-10-07 22:06:13]

LGTM

[commit-bot: I haz the power, 2011-10-07 22:21:59]

No LGTM from valid reviewers yet.