Review Request: Parser support for C99 designated initializers

[Alexandre Courbot]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/103448/
-----------------------------------------------------------

(Updated April 10, 2012, 6:17 a.m.)


Review request for KDevelop and Milian Wolff.


Changes
-------

New version of the patch with the following changes:
- Do not reorder statements in parserInitializerClause (as suggested by Milian)
- Parse the right part as a const instead of a comma expression (cause of infinite loop for some cases)

So now this patch do not change any existing behavior, only introduces a new one. This should make it safer. Parsing is done correctly 99% of the time, but the case of designated arrays initializers still need some work: in the new test introduced, the following array is not correctly parsed:

int ia[10] = {
  [1] = 10,
  [2] = 2,
};

Reordering the statements allows it to pass, but creates an infinite loop issue, so it is safer as-is for now. It should be safe to merge it in its current state, keeping in mind that some more work is necessary for the array case.


Description
-------

Parser support for C99 designated initializers

Support C99 initializers in the C++ parser, e.g:

struct foo_t foo = {
  .has_cake = true,
  .nb_candles = 12,
};

int bar[10] = {
  [1] = 15,
  [9] = 25,
};


Diffs (updated)
-----

  languages/cpp/parser/parser.h 64b5c505866ac2a0ed847aa67e4c7809df3af6b5 
  languages/cpp/parser/parser.cpp a45ee31a1207d56efef292babb2e774b51378ead 
  languages/cpp/parser/tests/test_parser.h 2869d9ddae2632aced2f2e2fa966382da3a0082d 
  languages/cpp/parser/tests/test_parser.cpp 94e3ee06dc9187568c000dcdd8baf8d87d1d06e6 

Diff: http://git.reviewboard.kde.org/r/103448/diff/


Testing
-------

Used it for a couple of weeks, ensured the parser tests all pass.


Thanks,

Alexandre Courbot

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kdevelop-devel/attachments/20120410/5d8dc079/attachment.html>