Commit Graph

6 Commits

Author SHA1 Message Date
Argyrios Kyrtzidis 76dbe8c800 Speed up code-completion by skipping function bodies.
When we are in code-completion mode, skip parsing of all function bodies except the one where the
code-completion point resides.

For big .cpp files like 'SemaExpr.cpp' the improvement makes a huge difference, in some cases cutting down
code-completion time -62% !

We don't get diagnostics for the bodies though, so modify the code-completion tests that check for errors.

See rdar://8814203.

llvm-svn: 122765
2011-01-03 19:44:02 +00:00
Douglas Gregor 21d0e5a06c This test should now work on windows, since we're no longer using temporary files badly
llvm-svn: 116261
2010-10-11 23:18:55 +00:00
Douglas Gregor 023da0d43a Make sure that we have File IDs for all of the unsaved files before we
deserialize diagnostics.

llvm-svn: 99426
2010-03-24 21:04:06 +00:00
Douglas Gregor b3fa919c2d Extend code-completion results with the type of each result
llvm-svn: 91702
2009-12-18 18:53:37 +00:00
Daniel Dunbar c449ed3ebf XFAIL this on Win32 for the time being.
llvm-svn: 90448
2009-12-03 11:12:50 +00:00
Douglas Gregor 9485bf9538 Extend the CIndex code-completion API to perform code completion
involving unsaved files, using the -remap-file= functionality recently
added to clang-cc.

llvm-svn: 90314
2009-12-02 09:21:34 +00:00