Community
Participate
Working Groups
The source tree and compilation log parsing have performance problem. The use of Eclipse and CDT for big SW project is effectively blocked with the slow parser. To fix the performace problem i propose separate parser(s( with the command line integration to Eclipse CDT IDE. Then optimized parser(s) can be written portable manner with the java or with the ISO C -language. This do not break anything when used in one workstation. It could speedup CDT to quite usable for big projects in one workstation if optimization done well. Any way the best result from separate parser is the possibility to utilize effective compilation farm behind the IDE. The parsing could also be distributed for several machines with the compilation command. Example remote shell command could be " gcc -c file.c | optimized_parser ". That way parsing load is also distributed away from front end machine to compilation farm. The CDT IDE running into front end machine stays usable because it gets ready made results from compilation threads. This works even if parser is not optimized at all. It is enough that the load can be effectively moved to compilation farm side.
This doesn't fit into our 3.0 plans. Feel free to check out the Indexer/DOM requirements documents which have been posted on our web site for the work that we have __committed__ to doing for this release.
Future means you commit to fix it in the Future. Inboxes can't make committments. Moving to '--'.
The performance of the indexer has been improved greatly, I recommend using the fast indexer.