gEDA-dev: Stack corruption PCB bug in hid_parse_actions????

[Stuart Brorson]

> On Thu, 2006-09-07 at 09:58 -0400, Stuart Brorson wrote:
>> [snip]  Simple (but adequate) code is therefore preferred over
>> tricker code which offers an unnoticible performance improvement since
>> the simple code can be understood and hacked by more developers.
>>
>> Just MHO,
>>
>> Stuart
>
> Oh dear.. you won't like the project manager backend I'm writing for
> scripting gsch2pcb updates to a running copy of PCB. Its in Python and
> (mis)uses generator functions to implement a less than obvious program
> flow. It writes out to PCB, then the generator function yields, until an
> IO callback causes it to re-enter where it left off.
>
> Its quite an elegant way to do it in some respects, and saves a complete
> state-machine implementation for every action I request PCB to do (which
> needs a response). Will require a little documentation.

I may indeed like it!  In any event, I am not the arbitor of
programming style.   Indeed, gattrib may be an example of what not to
do when writing a program.  But at least I put a ton of documentation
and comments in in the code.

Anyway, I am just rather opinionated about certains things. . . .  And
I do worry about code becoming too complicated for volunteer
developers to maintain it.

BTW:  When will you be able to put up an alpha version of your project
manager?  I'd love to see it, play with it, and maybe even offer a few
patches!

Stuart