A few updates on the new profile system that I'm (trying) to design.

[Lubomir I. Ivanov]

On 5 January 2014 04:38, Tomaz Canabrava <tcanabrava at kde.org> wrote:
> things are slowly taking shape. the code now alredy takes into consideration
> the dive and dive computer ( but just the first ), there's a new widget
> enabled by default on debug builds that I'm using to test stuff.
> I'm terribly sleepy, but if anyone care to take a look and see if Im not
> doing wrong things, please take a look at github / tcanabrava / subsurface /
> newProfileSystem.
>
> The code is *extremely* initial, so I'm not yet plotting anything, I plan to
> have the first plotting tomorrow.
>

just tried the branch, tomaz.
and reading this really cleared a lot more details for me:
----
    Started the work for a new profile widget that will catch them all.

    The idea of the new profile widget is that it should be something
    dynamic, easy-to-expand, future-proof ( well, kinda, since I'm
    still realying on QGraphcisView and the future is QML, but I
    plan to make it easy to port ._. ).

    - Integration with some QML Items ( mostly the Toolbar / Tooltip
    system )
    - Plugin Integration
    - The return of the Planner
    - Edit / Show dives
    - Fast.
---

i got an assert:
ASSERT: "last >= first" in file kernel/qabstractitemmodel.cpp, line 2413
and these are notoriously difficult to debug on win32, as i can't (or
don't know how to) backtrace.

you mentioned that profile.c has remaining cairo fragments in there,
but i can't find them also 'struct plot_info' and things like
'calculate_max_limits()' are just math, and are fine to remain in the
backend, IMHO.

i think it would be best if you ask the 'struct plot_info' questions
here, because i think IRC (and chat) is a horrible medium for
development.
it's just chit-chat...and also linus doesn't use it and he is the one
who wrote the original profile.c and the math in there, after all.

lubomir
--