2 The approach I've been using for a given header is to recursively do each
3 of the "bits" headers which make up the standard header. So, e.g., while
4 there are four headers making up <algorithm>, three of them were already
5 documented in the course of doing other headers.
7 "Untouched" means I've deliberately skipped it for various reasons, or
8 haven't gotten to it yet. It /will/ be done (by somebody, eventually.)
11 Area Still needs to be doxygen-documented
12 -----------------------------------------------------------
14 c17 FINISHED (Nothing in Clause 17 "exists" in terms of code.)
20 c23 See doxygroups.cc and Note B.
22 c25 stl_algo.h (lots of stuff)
23 c26 <complex>, <valarray>, stl_numeric.h[26.4], Note A
26 backward/ Not scanned by doxygen. Should it be?
28 ext/ Some of the SGI algorithm/functional extensions.
29 All of rope/hashing/slist need docs.
33 [1.3.5] "implementation-defined behavior: behavior ... that depends
34 on the implementation *and that each implementation shall
35 document*." [my emphasis] Not all implementation choices
36 have been thus described; doxygen is not necessarily the
37 appropriate place for such descriptions, either.
39 -----------------------------------------------------------
43 A) So far I have not tried to document any of the <c*> headers. So entities
44 such as atexit() are undocumented throughout the library. Since we usually
45 do not have the C code (to which the doxygen comments would be attached),
46 this would need to be done in entirely separate files, a la doxygroups.cc.
48 B) Huge chunks of containers and strings are described in common "Tables"
49 in the standard. How to reproduce this information?
54 stl_deque.h, stl_pair.h, and stl_algobase.h have good examples of what I've
55 been using for class, namespace-scope, and function documentation, respectively.
56 These should serve as starting points. /Please/ maintain the inter-word and
57 inter-sentence spacing, as this might be generated and/or scanned in the