Opened 20 years ago
Closed 15 years ago
#882 closed enhancement (duplicate)
Layer timings in map->debug mode
Reported by: | refractions | Owned by: | refractions |
---|---|---|---|
Priority: | high | Milestone: | |
Component: | MapServer C Library | Version: | 4.3 |
Severity: | minor | Keywords: | |
Cc: | woodbri@…, jmckenna |
Description
By writing out the time required to - open data sources - query data sources - fetch features - draw features - close data sources for each layer processed, users can more easily tune their map files and understand where their performance bottlenecks are. That will lead to faster map files and make mapserver look even better than it does now! This is probably best done for the map->debug flag rather than the layer->debug flag, since timings are really most useful in comparison with each other.
Change History (4)
comment:2 by , 16 years ago
Cc: | added |
---|
wasn't this addressed with ticket#2673 and other 5.x closed tickets?
comment:3 by , 16 years ago
Jeff,
This is partially covered by the changes addressed in #2673, but I do not think that we return any stats on numbers of shapefiles or number of shapes read from each shapefile, which would be useful to have for understanding how well your data is organized on disk. I'm OK, if you want to close this because what we have is way better than what we had in the past when this was written.
comment:4 by , 15 years ago
Resolution: | → duplicate |
---|---|
Status: | new → closed |
Logging the number of features will probably require an addition to the layerinfo so the information can get passed back up to the msDrawMap routine by the specific drivers. Knowing now many features go into a layer is a useful performance tuning metric, but the change would be a bit invasive. This probably calls for a new issue or RFC if it's a high priority.