Reading through some CEDET docs I found that the author herself recommends "to use Semantic only as source of information, and perform names completion using other packages". Does this advice also hold for other tasks? I want to analyze C source and insert e.g. a doxygen documentation for a function which matches the function signature. Which package is the right one for such an attempt? Note that I need to adapt the parser as the source is some awkward embedded we-roll-our-own-syntax C code.