On Thu, 2004-10-07 at 17:43 -0500, Robert Lipe wrote:
> Patrick Ohly wrote:
> > The date_time.diff needs more testing; I haven't
> > checked yet whether the corner cases (e.g. missing
> > time) are handled correcly. I would do that
> > by extending the test suite after changing the
> > Fugawi style.
> I'll let you make the final call. Your approach looks fine, so whenever
> you're ready to have it committed, send the final version and I'll check
> it in.
It is kind of embarrasing to quote the time when this
discussion took place; my only excuse is that I got
distracted by other stuff (like learning about OpenEmbedded
and packaging gpsbabel for embedded Linux distributions
built with OpenEmbedded as Familiar) and only now started
to use Fugawi again.
Anyway, attached is a patch against the current gpsbabel
head revision which improves the Fugawi import/export
and its documentation.
One quirk that I found when
extending the test suite was that the CSV import always
strips the expected number of header lines at the
start of the input file. The effect was that the first
two waypoints from "old style" Fugawi CSV were skipped
because I had added two more lines to the "new style"
format. To remain backwards compatible I then tacked the
description of the two new fields onto the end of the
last line in the old header without changing the number
I also attach a patch which (I hope) improves the the
documentation with regards to filters and
waypoint/route/track handling/conversion. This is
based on the discussion we had at the same time under
the subject "A novice user's view".
Please note that one sentence in the current documentation
is incomplete and I wasn't quite sure how to complete it:
The presence of "-s" on the command line tends to creat havoc because
tracks and routes. [TODO: insert explanation here]