On Oct 17, 2012, at 6:04 PM, Craig Topper <[email protected]> wrote:
> This patch uses the line number cache to find the start of the line instead
> of searching backwards for it. This helps performance in cases where the line
> is long and have just called getLineNumber for the same position.
This only helps when the FilePos is exactly the same as the FilePos that went
to getLineNumber:
+ // See if we just calculated the line number for this FilePos and can use
+ // that to lookup the start of the line instead of searching for it.
+ if (LastLineNoFileIDQuery == FID && LastLineNoFilePos == (FilePos + 1)) {
Why not simply check that FilePos > SourceLineCache[LastLineNoResult - 1] and <
SourceLineCache[LastLineNoResult], so anything on the most-recently-queried
line can be handled efficiently?
- Doug
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
