On 2015-06-09 01:25, Linus Torvalds wrote:
On Mon, Jun 8, 2015 at 10:09 AM, Linus Torvalds
<[email protected]> wrote:
I'll get to the actual sample data later today, I hope.
.. and here's that third patch that converts the sample parsing to
dynamically using the actual sample type descriptions.
This includes actually parsing things like the group descriptions (ie
where one "packet type" can be defined as a group of basic types, kind
of like a limited "struct" in C). It doesn't really parse the unit
conversion descriptions, since those seem to be mainly informational,
but it does do things like sanity-check the type sizes etc.
This has so far been tested only with my old dives, but since the code
builds up the parsing tables from the descriptions, that's actually a
pretty good test in itself: instead of "knowing" what the sample
structures are like the old code did, it just worked it out itself.
The end result is identical to the old parser, with the exception that
the new parser also parses the compass heading events.
Maybe I'll remember to take the EON steel to the pool tomorrow and get
this at least partially validated that way - otherwise further testing
will wait until next week.
Please apply
The second and third patch do not apply cleanly against master. I guess
I'll have to sort out some of the differences with the subsurface branch
first. I'll have a look at that tomorrow.
Jef
_______________________________________________
subsurface mailing list
[email protected]
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface
