Details
-
Task
-
Resolution: Unresolved
-
P2: Important
-
None
-
None
-
None
-
13
-
Foundation PM Prioritized
Description
Both in tokenize (QStringTokenizer, split(), tokenize()), as well as other functions (count(needle), basically anything that loops over indexOf()), we repeatedly build BM tables for a one-off search and then re-create them for the next loop iteration.
This is clearly wasteful, so we should identify all these issues and fix them.
Attachments
Issue Links
- is cloned by
-
QTBUG-108703 Optimize prepending of UTF-8 to QString
-
- Open
-
Gerrit Reviews
For Gerrit Dashboard: QTBUG-101449 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
640336,1 | Add benchmark for QString::tokenize() | dev | qt/qtbase | Status: NEW | 0 | 0 |