This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
| Summary: | Allow modules to bundle preindexed data | ||
|---|---|---|---|
| Product: | editor | Reporter: | Erno Mononen <emononen> |
| Component: | Parsing & Indexing | Assignee: | Tomas Zezula <tzezula> |
| Status: | NEW --- | ||
| Severity: | blocker | CC: | dbalek, dstrupl, johnjullion, pjiricka |
| Priority: | P2 | Keywords: | PERFORMANCE |
| Version: | 6.x | ||
| Hardware: | All | ||
| OS: | All | ||
| Issue Type: | ENHANCEMENT | Exception Reporter: | |
|
Description
Erno Mononen
2009-03-06 09:22:09 UTC
This is a deficiency in the indexing infrastructure. As discussed, modules should have a possibility to preindex libraries and bundle the index data in the IDE installation. When the IDE is installed the preindexed data will be used instead of indexing the same libraries again. This is purely a performance improvement. Considering the high number of other high priority issues related to the indexing infrastructure I've just asked for permission to defer this issue to the next release. Would expect this of all JARs we ship with the IDE, plus major platforms (GF/Java SE/Ruby) etc. When speaking about java: according to Dusan rt.jar (main component of JSE) index is about 3 MB and we could save approx 10 seconds of scanning during the first startup of the IDE. I am not sure it is worth bundling those 3 MB but if we find out that e.g. glassfish jars for web projects would make the initial experience of developing web app better we might consider it for the glassfish jars. Petr? Yes, I think that's a possibility worth looking at. Just an idea: why don't you integrate this process into installer phase, instead of increasing download side? Thanks, Vladimir. The easiest way is to allow modules to provide lucene index which is then copied (extracted) to the ${nbuser}/var/cache/index/**/ folder.
Tor already had this feature in the GSF, probably can be taken from there.
The lucene index was designed to be relocated (it does not contain any absolute path, holds for both parsing.api and java index). So it's safe to copy it or to connect it to the sources located on other absolute URI. We have been living without prebundled indices for several releases now and have had no really strong complaints about it. |