Community
Participate
Working Groups
consider project S dependent on project A, with a file 'S/source.c' including 'A/header.h'. When indexing S the header is either found in PDOM of A or has to be parsed. If it has to be parsed, the question arises whether or not to store the result in the PDOM for A. In other words to we allow indexer to write to the PDOM of dependent projects? If we do not allow that, we can run in situations, where header files get parsed over and over again (just consider that there are many source files that include the header and that S happens to be indexed before A). Furthermore the quality of the parsing result is better in case the header is parsed in context of a source file including it. In case PDOM A cannot be written to by the indexer for S, it has to be parsed stand-alone. I am not sure what's best to do here, to some extent I have the feeling that dependent projects should be handled by a single indexer. This does not look like a strong reason for introducing the complexity of allowing to indexers
