I've been following this thread for a while and there has been much talk about the obvious trials but I have yet to read anything regarding the justification for the added overhead or the acceptance of parallelization (for perhaps some text searching)?
There is definitely great interest in making FPC Lazarus faster which I think will greatly enhance overall experience for developers. Given the requirements and costs associated with complexities of managing Lazarus the modules associated with parallelization will need to be isolated as to not confuse any other group. Merely adding the cthreads unit to Lazarus so we can have faster text searching seems like a trivial task. However, the developers tasked with integrating the improvements should really know the IDE inner workings as to AVOID blowing out the trunked version. I personally have had problems getting trivial changes to Lazarus to propagate through the versions and would only offer an explanation that core developers are looking to get incomplete road map items and bugs resolved rather than accept code from people that are off on their own just excited about bringing their own technologies to the Lazarus community. I have a feeling that all this discussion is for naught because even if you make the changes someone somewhere else will just simply not accept them. So, unless you SVN commit access and are interested in performing any low-level architectural changes to Lazarus - don't bother. And if you are serious the best way to get involved is to be tasked with an aspect of Lazarus development. The time of dedicated Lazarus developers is extremely valuable and limited. It's nice to see ideas but Lazarus is a complicated system and it still has much further to go. It was only until recently did we get Multi-document feature ;-) Getting the FPC team to embrace OLAPish technique to establish unit dependency hierarchy to parallelize FPC would be a better challenge and benefit to us all :-) -- _______________________________________________ Lazarus mailing list [email protected] http://lists.lazarus.freepascal.org/mailman/listinfo/lazarus
