Jef,
Yes, I think that's better. With the caveats section and the current description ("might interact with the file-system in unexpected ways"), it sounds as if there is a problem. And that's not really true.
I agree: enter version 3! (Minus the Makefile.am bits.) This just has the notice as a separate paragraph.
An unrelated note: I found that sometimes you use dc_datetime_gmtime and sometimes dc_datetime_localtime for the dc_parser_get_datetime value. So is it correct to say that the value always returns the broken-down localtime? (Specifically: the shearwater predator, suunto eon steel, and uwatec smart computers invoke gmtime. The uwatec specifically notes the rationale behind this, however.)
Isn't that obvious? A function to parse date/time will probably use the library's date/time functions :-)
True, but it doesn't help to be specific! I just included a note that it calls to the dc_datetime_gmtime or dc_datetime_localtime functions.
Best,
Kristaps