[LLVMbugs] [Bug 18685] weird typo correction to member of class template specialization

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=18685

Kaelyn Takata <rikka at google.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED
           Assignee|unassignedclangbugs at nondot. |rikka at google.com
                   |org                         |

--- Comment #1 from Kaelyn Takata <rikka at google.com> ---
r201056 has been committed to exclude classes that are template specializations
as potential nested-name-specifiers unless the existing nested-name-specifier
is a template specialization.


As a side note, while it probably could use a bit of tuning, correcting both
the name and the nested-name-specifier isn't as low-confidence as your example
might indicate, and has proven quite useful. For example, what if
llvm::BitVector had been declared, but you accidentally typed just "Bitvector"
outside of the llvm namespace (and without a corresponding using statement), or
maybe typed "clang::BitVEctor" because you had just typed in a bunch of stuff
that required "clang::" instead of "llvm::"? Right now both cases would get
corrected to "llvm::BitVector". ;) There are also multiple cases in four of the
unit tests where the combined correction gives the expected correction,
including examples pulled from previous bug reports such as correcting the enum
names in a switch statement to have the right nested-name-qualifier after the
enum was moved into a class (see the correction for PREDICAT in
test/SemaCXX/typo-correction-pt2.cpp).

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20140209/85415552/attachment.html>