Bug#698350: [Gpsbabel-code] Bug#698350: HTML entities left undecoded when using -i gpx

[jidanni]

>>>>> "RL" == Robert Lipe <robertl...@gmail.com> writes:
RL> Working as intended.

RL> GPX doesn't have a way to specify if those fields are HTML or not. 
RL> Kml always allows HTML there. We have to entity encode them.

You are not entity encoding them.
You are just not entity decoding them.
Or you are yes just leaving them alone, however still slapping a
CDATA holder around them.

So maybe don't slap the CDATA wrapper, and everything would be fine.
Or if you do, first decode any entities.

I personally don't have much of a real grip on what is wrong actually.
All I know is I was given a GPX file with

    <desc>&lt;ul&gt;&lt;li&gt
    which becomes in KML
    <description><![CDATA[&lt;ul&gt;&lt;li&gt;

So I had do decode it with the programs you see there in
http://radioscanningtw.jidanni.org/images/radioscanningtw/maps/

One might explore
 xmllint --nocdata
    Substitute CDATA section by equivalent text nodes.
But that would just doubly encode things.

Or maybe
htmldoc/fmt_gpx.html or
htmldoc/fmt_kml.html
needs a new option,
because as you say,
the program doesn't know which kind of input it is getting,
so give the user a way to tell it...


--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org