kicad-developers team mailing list archive
-
kicad-developers team
-
Mailing list archive
-
Message #12879
Re: writing a new top level wxPython based project manager
On 04/02/2014 12:39 PM, Dick Hollenbeck wrote:
> Here is more info, which will help you create some low level mock up classes.
>
> Class KIWAY_MGR:
> (temporarily in kicad/kicad.cpp) will get swigged, creating a python class implemented
> fully in C++. It is a collection of PROJECTs/KIWAYs. The constructor of this class may
> take a wxFrame*, so we have the option of hooking child frames into that parent. Possible
> functions:
> createProject( full_path_to_pro_file )
> openProject( full_path_to_pro_file )
> listProjects()
>
>
> Class KIWAY:
> will get swigged, creating a python class implemented fully in C++. Possible functions:
> openSchematic() no args required, schematic's name is in the *.pro file.
> openBoard() no args required, board's name is in the *.pro file.
>
>
> Just with this sketch, you should be able to mock enough both of these classes, enough to
> work magic at in the UI layer. If you want to open dummy child frames, that would be ok
> from openSchematic and openBoard().
>
>
> Feel free to add functions as you find needs for them. This will give us an aspect of top
> down design.
>
>
> HTH,
>
> Dick
>
>
>
>
>
> On 04/02/2014 11:58 AM, Dick Hollenbeck wrote:
>> Hi Jean-Samuel,
>>
>> Thanks very much for your willingness to help!
>>
>> In answer to your question:
>>
>> The wxPython project manager sketch up should concentrate on the user interface, not the
>> actual loading of the sub programs. In fact, if it did not load the subprograms at all,
>> that would be acceptable. The screen real-estate usage, and the icons and whatnot are
>> what's important for now. You could create scaffolding to mark a project and a kiface as
>> being loaded, and test your UI code against that scaffolding, without ever loading any
>> projects or opening any child frames.
>>
>> In milestone B) we are putting a top C++ API into the KIWAY, this is not designed yet.
>> You should not use the KIFACE from python, that was never intended. Python will use the
>> KIWAY API, which in turn will manage the KIFACE's and projects using C++.
>>
>> If you really really want to launch something, then encapsulate that python code into
>> something which will be replaced later, and simply load pcbnew.exe for now from that
>> disposable code. This code will get swapped out later, in favor of the KIWAY API, which
>> will do the actuall loading of the *.kiface, again from C++. This is necessary so that
>> the KIFACE to KIFACE
oops, meant not "KIFACE to KIFACE", but rather
"project specific KIWAY_PLAYER to KIWAY_PLAYER"
comms can work for things like back annotation, and pcbnew<->eeschema
>> cross-probing, per project.
>>
>>
>> There's actually little value in loading executables or kiface's right now. Folks will
>> use the C++ project manager until the user interface in the python project manager is
>> superior. So please focus on the user interface, that's the value added stuff.
>>
>> After milestone B) that's when we start to worry about loading the submodules (kifaces).
>>
>>
>> Thanks,
>>
>> Dick
>>
>>
>>
>> On 04/02/2014 11:23 AM, Jean-Samuel Reynaud wrote:
>>> Hi
>>>
>>> As I understand, one of the feature of this new interface is to load directly from python
>>> the kiface interface (instead of just launch the program aka pcbnew/cvpcb for example).
>>> Is it right ?
>>> As I see, the function KIFACE_1 return a struct with inside all function to use (one of
>>> them is createwindow).
>>> Could you confirm it's the way you think about ?
>>> If yes, ctypes from python will help me and I think I'll fight with the structure
>>> definition inside python...
>>>
>>>
>>> Regards
>>>
>>>
>>>
>>> 2014-04-01 15:47 GMT+02:00 Dick Hollenbeck <dick@xxxxxxxxxxx <mailto:dick@xxxxxxxxxxx>>:
>>>
>>>
>>> In English there is a saying: "Competition breeds success."
>>>
>>> I would consider it a luxury to be able to choose [pieces] from multiple implementations.
>>> Competition in this case is not exactly cruel and unusual punishment. In fact it could
>>> be an *awful lot of fun*, since wxPython is such a high level language and so much can be
>>> done in so few lines of code.
>>>
>>>
>>> You may either compete or collaborate. I think we'll get more ideas if there is
>>> competition. After choices are made, you can continue to compete, improving, innovating
>>> and serving your own needs, even if your code is not merged up to that point in time.
>>> Repeat that last sentence.
>>>
>>>
>>> Any implementation which does not meet the primary objective: "showing which projects are
>>> open, among a larger set of project possibilities" will lose the competition. (Even if
>>> there is only one entry into the competition, *multiple* open projects is a mandate.)
>>> Using screen real estate for a permanently visible directory tree may not be an optimal
>>> pathway to achieving the objective.
>>>
>>> Some tips:
>>>
>>> a) the directory tree does not give sufficient emphasis on what constitutes a project.
>>>
>>> b) the directory tree does not give the ability to show several open projects without a
>>> sea of rows of actual files in between.
>>>
>>> c) we need to be more project centric, less file centric.
>>>
>>> d) once projects have been defined, the user interface procedure to opening the schematic
>>> or layout tools on any of those projects needs to be lean.
>>>
>>> e) I'm thinking the set of files that constitute a project can continue to reside in the
>>> *.pro file.
>>>
>>>
>>>
>>>
>>>
>>> On 04/01/2014 07:24 AM, dileep kushwaha wrote:
>>> > Hi,
>>> > I have so far made the menubar for the wxPython. I now want to integrate the
>>> > mainmenu.py to KiCAD instead of .cpp codes. I little hint and guidence will be great. I
>>> > can send the code. Its ugly right now. i'll modify it according to guidlines after
>>> > completing significant portion...(i see three .cpp files that will have to changed)
>>> >
>>> >
>>> > On Tue, Apr 1, 2014 at 9:14 AM, dileep kushwaha <dilzverykool@xxxxxxxxx
>>> <mailto:dilzverykool@xxxxxxxxx>
>>> > <mailto:dilzverykool@xxxxxxxxx <mailto:dilzverykool@xxxxxxxxx>>> wrote:
>>> >
>>> > Hi,
>>> > Please accept my apology for the delay from my side. I can give you few results
>>> > in maximum one week's time. I am able to make an exact replica of KiCAD window
>>> that we
>>> > have right now. I was carried away and wanted the GUI to look pretty innovative.
>>> I was
>>> > looking into the feasibility of an idea. The idea comes from a member who
>>> proposed it
>>> > to have a game like GUI. I was thinking if right click would give list of options
>>> > surrounding it in circle. I think I will have to wait a little more to get
>>> expertise.
>>> >
>>> > I was shying away from showing my work. Apparently I will start uploading on the
>>> > branch by evening 6PM IST(Indian Standard Time)
>>> >
>>> >
>>> >
>>> > On Tue, Apr 1, 2014 at 2:27 AM, Jean-Samuel Reynaud <js.reynaud@xxxxxxxxx
>>> <mailto:js.reynaud@xxxxxxxxx>
>>> > <mailto:js.reynaud@xxxxxxxxx <mailto:js.reynaud@xxxxxxxxx>>> wrote:
>>> >
>>> > Ok I'm starting to work to meet window 1).
>>> >
>>> >
>>> > 2014-03-31 17:01 GMT+02:00 Dick Hollenbeck <dick@xxxxxxxxxxx
>>> <mailto:dick@xxxxxxxxxxx>
>>> > <mailto:dick@xxxxxxxxxxx <mailto:dick@xxxxxxxxxxx>>>:
>>> >
>>> > On 03/24/2014 10:19 AM, Jean-Samuel Reynaud wrote:
>>> > > Hi,
>>> > >
>>> > > I can help too on this project. Tell me if you need more help.
>>> >
>>> >
>>> >
>>> > https://code.launchpad.net/~kicad-developers/kicad/prj-mgr
>>> >
>>> > is still empty, so therefore I need help.
>>> >
>>> >
>>> > There are two windows of opportunity here, with a middle section where the
>>> > window is not open:
>>> >
>>> >
>>> > 1) Now, before I get to milestone C)
>>> >
>>> >
>>> > 2) Me doing milestone C), where collisions of effort are so likely its most
>>> > sensible to
>>> > consider python contributions in this time un-helpful. Plus I don't
>>> know what
>>> > I'll have
>>> > to work with at this point. And I won't be waiting for it.
>>> >
>>> >
>>> > 3) After the C++ layer and framework are in place, then collisions of effort
>>> > will not
>>> > concern me, the python project manager can and should be expanded into a
>>> number of
>>> > directions, and I don't expect to play a role in that.
>>> >
>>> >
>>> > I want to be clear about these two windows of opportunity, 1) & 3), and the
>>> > approximate
>>> > timing, at least the sequencing.
>>> >
>>> > The contributions that I would value most are in the window described by 1).
>>> >
>>> >
>>> >
>>> >
>>> > _______________________________________________
>>> > Mailing list: https://launchpad.net/~kicad-developers
>>> > Post to : kicad-developers@xxxxxxxxxxxxxxxxxxx
>>> <mailto:kicad-developers@xxxxxxxxxxxxxxxxxxx>
>>> > <mailto:kicad-developers@xxxxxxxxxxxxxxxxxxx
>>> <mailto:kicad-developers@xxxxxxxxxxxxxxxxxxx>>
>>> > Unsubscribe: https://launchpad.net/~kicad-developers
>>> > More help : https://help.launchpad.net/ListHelp
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> > --
>>> > Dileep Kumar
>>> > M.Tech(VLSI Design)
>>> > Mob:9891455965
>>>
>>>
>>
>>
>> _______________________________________________
>> Mailing list: https://launchpad.net/~kicad-developers
>> Post to : kicad-developers@xxxxxxxxxxxxxxxxxxx
>> Unsubscribe : https://launchpad.net/~kicad-developers
>> More help : https://help.launchpad.net/ListHelp
>>
>
>
> _______________________________________________
> Mailing list: https://launchpad.net/~kicad-developers
> Post to : kicad-developers@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~kicad-developers
> More help : https://help.launchpad.net/ListHelp
>
Follow ups
References
-
writing a new top level wxPython based project manager
From: dileep kushwaha, 2014-03-12
-
Re: writing a new top level wxPython based project manager
From: Dick Hollenbeck, 2014-03-12
-
Re: writing a new top level wxPython based project manager
From: Dick Hollenbeck, 2014-03-13
-
Re: writing a new top level wxPython based project manager
From: dileep kushwaha, 2014-03-24
-
Re: writing a new top level wxPython based project manager
From: dileep kushwaha, 2014-03-24
-
Re: writing a new top level wxPython based project manager
From: Brian Sidebotham, 2014-03-24
-
Re: writing a new top level wxPython based project manager
From: Dick Hollenbeck, 2014-03-24
-
Re: writing a new top level wxPython based project manager
From: Dick Hollenbeck, 2014-03-24
-
Re: writing a new top level wxPython based project manager
From: Jean-Samuel Reynaud, 2014-03-24
-
Re: writing a new top level wxPython based project manager
From: Dick Hollenbeck, 2014-03-31
-
Re: writing a new top level wxPython based project manager
From: Jean-Samuel Reynaud, 2014-03-31
-
Re: writing a new top level wxPython based project manager
From: Dick Hollenbeck, 2014-04-01
-
Re: writing a new top level wxPython based project manager
From: Jean-Samuel Reynaud, 2014-04-02
-
Re: writing a new top level wxPython based project manager
From: Dick Hollenbeck, 2014-04-02
-
Re: writing a new top level wxPython based project manager
From: Dick Hollenbeck, 2014-04-02