Adding integration with more online dive logs GSoC 2015

Dirk Hohndel dirk at hohndel.org
Tue Mar 10 15:20:42 PDT 2015


On Wed, Mar 11, 2015 at 12:13:17AM +0200, Yosef Hamza wrote:
> On Tue, Mar 10, 2015 at 11:39 PM, Dirk Hohndel <dirk at hohndel.org> wrote:
> 
> > On Tue, Mar 10, 2015 at 11:08:14PM +0200, Yosef Hamza wrote:
> >
> > a) the events are ordered by time, so the moment you have one where
> >    time.seconds > 0 you can stop
> >
> 
> Yea I thought about that, But Gehad told me before that there was an issue
> that can make events be out of order.

There is? That would be bad. Gehad, can you comment on that?

> b) even easier, we have a helper function you could use:
> >         get_next_event(ev, "gaschange");
> >    so you call this with current_dc->events
> >    if you get an event back AND its time stamp is 0, remove that event
> >    try again, again with current_dc->events as pointer
> >    repeat until you don't find a gas change
> >
> 
> That's a better solution indeed I didn't know about, I need to surf the
> helper
> functions more..

We have a lot of them. And not well structured, not consistently named,
not well documented.

There's almost a GSoC project hiding here in code cleanup... but the
problem is that this would have to be done by a rather experienced
developer whom I really trust... hard challenge for GSoC...

> c) I wouldn't call replot() in the middle; while it's unlikely that there
> >    would be two gas changes at t=0, still one replot() would be sufficient,
> >    right? So I'd have a little flag (bool eventRemoved) and then call
> >    mark_divelist_changed() and replot() if eventRemoved is true.
> >
> > Makes sense?
> 
> Yes.

Excellent.

> > I know that this might feel frustrating. You keep sending patches and I
> > keep criticizing the code and not taking them. This really is intended to
> > help you and to make sure you understand the code and the code we have in
> > the end makes sense. Sure, I could just fix this myself (that would be a
> > lot less work for me) - but it's unlikely that you'd learn something if I
> > did that.
>
> No not at all, It give me valuable checkmarks when submitting patches :)

:-)

/D


More information about the subsurface mailing list