On Monday 25 May 2009 11:51:05 Antoine Pitrou wrote: > On May 25, 11:40 am, "Diez B. Roggisch" <[email protected]> wrote: > > > I'm not sure why all of PIL, genshi, mako, tempita, pygments (!), the > > > whole paste and tw.core subhierarchies, among others, have to be > > > loaded like that. > > > > And the gain of lazyfying that so that it happens when the first request > > hits the system would be what? > > It's not about lazyfying, it's about loading only what's needed. > There's no reason to load three different template engines, the PIL > library, the pygments package, and whatever else, if your application / > never/ makes use of them. > (as for pygments, perhaps it's needed in debug mode for the nice > traceback display? regardless, it shouldn't be loaded unconditionally)
But this *is* lazyfying, aka making imports at the moment you need them. And that's a non-trivial task to do, especially since much of the imports are not directly tg-driven, but instead via 3rd-party-packages which would need to be adapted, too. Then there is the question who's going to decide when it's ok to import what - via configuration? So instead of making usage of a ToscaWidget that happens to use a template-engine you so far didn't use simple, you make it more complicated by needing to somehow propagate a global config setting downwards. > > I still don't see the issue - do you actually have *problems* serving a > > TG app with reasonable speed and scalability? > > Not concretely (the app I'm working on has a couple of requests per > minute only). But I know that if I had a performance-critical app, I'd > be very suspicious about TG's memory requirements. > > Loading lots of modules has a real cost. Besides inflating memory > requirements (and potentially decreasing CPU caches/TLBs efficiency), Please, we are talking about python here, not hand-optimized LAPACK code. And resident modules loaded but not actively used do degrade CPU caches in exact what way? > it also, as I showed, has a dramatic impact on the cost of garbage > collection. > (Python 2.7 will mitigate the latter a bit, but not make it disappear) All of this without a compelling usecase for optimizing - sorry, but you won't get far with this anywhere unless you show a real impact for a real use-case - not some premature optimization guesswork. Diez --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "TurboGears" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/turbogears?hl=en -~----------~----~----~----~------~----~------~--~---
