[Geany-devel] Ideas on increasing quality of plugins

Colomban Wendling lists.ban at xxxxx
Tue Feb 22 20:40:09 UTC 2011


Le 22/02/2011 20:04, Thomas Martitz a écrit :
> Am 22.02.2011 19:36, schrieb Colomban Wendling:
> 
>>> So my 1st suggestion is to remove all plugins which do have known issues
>>> and don't compile with some -W-flags (needs to be defined) from common
>>> build until these are fixed. Also remove plugins which don't bring
>>> propper documentation as well are unmaintained for some time.
>> Well... I'm puzzled. Disabling unperfect plugins is a good idea in
>> theory for the stability of Geany, but... how would be plugins tested
>> then?
>>
>> However, I agree with making official builds with strong error checking
>> enabled (at least -Werror-implicit-function-declaration comes directly
>> to my mind). This is static checking that will show most obvious
>> problems, and hopefully it'll encourage their maintainer to fix them.
>> However, finding the right flags will not be easy... for example not
>> allowing any implicit cast may have a false-positive effect where
>> unexperimented developers might hide a part of those with a cast where a
>> proper fix would be better.
> 
> I agree. While we may know they have known problems, how would we know
> in the future if they're disabled and receive no testing?
> 
> I think all plugin developers should get more easy about others touching
> there code. I could probably fix a number of problems in other people's
> plugins, but I'm afraid of the surrounded discussion (post or pre commit).
I don't think that touching other one's code without her agreement is a
good idea. There are several reasons for this that comes to mind:
* some people will find this offensive (e.g. "you can't even code right
and you won't even understand if I tell you");
* some people will find it cool, but won't necessarily try to understand
any further (heh, it's already fixed), so would likely reproduce
mistakes later;
* and unfortunately, some people might change things with not enough
care or understanding and break other things -- this is easier than we
might think...
* and some people would not feel confident of somebody else commit --
and this would probably be my case if it's somebody I hardly know.

And AFAICT, people who would need help are generally likely to accept it
when proposed (probably more than if it was imposed).

> What I also wonder; is there a way to prevent plugins from crashing
> geany entirely? I'd rather have a notice "Oops, Plugin Foo crashed".
I doubt. This would need to run plugins in a separate process, and then
doing some IPC to communicate. Probably doable (well, AFAIK recent web
browsers tend to do so with their plugins -- chromium, FF4) but probably
really not easy.


Regards,
Colomban



More information about the Devel mailing list