ricardo lafuente
2009-06-21 13:19:10 UTC
hai people,
while the new, refactored branch still has quite a way to go regarding
some brokenness (i'm working on it), dividing the code base into
separate files really has its merits. I for one find myself hacking a
lot faster, not having to scroll through a huge file with lots of
classes. And i think that the code structure becomes a lot more evident
this way, be it for newcomers or veterans. (see my previous e-mail on
refactoring for a run-down on the new file structure)
my plan, which i'd leave at your consideration, would be to wrap up the
main repository into a 0.2.1 release, since stuff like key interaction
was introduced after 0.2. I'm thinking of doing this by the end of the
week. from there, i could pull the refactored branch into the main
repository so that we can work on that for 0.3.
also, this way we can also focus on writing the documentation and a new,
spiffy website for Shoebot.
if you happen to be trying out the new branch, there's a caveat to be
aware of: imports changed drastically: no more 'import shoebot'. Though
you can still do that, i've followed the django way of separating
different parts of the code:
* the data structures that used to be in data.py are now in a 'data'
module. So if you need the Color class, the right way would be 'from
shoebot.data import Color'.
* the core engine is inside the 'shoebot.core' module. This includes
Bot, NodeBot, Canvas and CairoCanvas.
* the GUI part is in 'shoebot.gtk'. Here you'll find ShoebotWindow and
ShoebotDrawingArea, the SocketServer and VarWindow, as well as the IDE code.
you'll also find that the shoebot-ide script was very thinned down, as
its innards are now in the gui module (ide.py).
as for the sbot script, it still retains its functions. Maybe we can
look into this later, but for now it works fine.
do you think this sounds hasty? Or would you rather be hacking on the
new file structure already? Any comments on stuff that you think would
be better someplace else?
while the new, refactored branch still has quite a way to go regarding
some brokenness (i'm working on it), dividing the code base into
separate files really has its merits. I for one find myself hacking a
lot faster, not having to scroll through a huge file with lots of
classes. And i think that the code structure becomes a lot more evident
this way, be it for newcomers or veterans. (see my previous e-mail on
refactoring for a run-down on the new file structure)
my plan, which i'd leave at your consideration, would be to wrap up the
main repository into a 0.2.1 release, since stuff like key interaction
was introduced after 0.2. I'm thinking of doing this by the end of the
week. from there, i could pull the refactored branch into the main
repository so that we can work on that for 0.3.
also, this way we can also focus on writing the documentation and a new,
spiffy website for Shoebot.
if you happen to be trying out the new branch, there's a caveat to be
aware of: imports changed drastically: no more 'import shoebot'. Though
you can still do that, i've followed the django way of separating
different parts of the code:
* the data structures that used to be in data.py are now in a 'data'
module. So if you need the Color class, the right way would be 'from
shoebot.data import Color'.
* the core engine is inside the 'shoebot.core' module. This includes
Bot, NodeBot, Canvas and CairoCanvas.
* the GUI part is in 'shoebot.gtk'. Here you'll find ShoebotWindow and
ShoebotDrawingArea, the SocketServer and VarWindow, as well as the IDE code.
you'll also find that the shoebot-ide script was very thinned down, as
its innards are now in the gui module (ide.py).
as for the sbot script, it still retains its functions. Maybe we can
look into this later, but for now it works fine.
do you think this sounds hasty? Or would you rather be hacking on the
new file structure already? Any comments on stuff that you think would
be better someplace else?