By load time i mean the time it takes for TurboGears to process a page request (IE, i click on a link, turbogears churns for five seconds, then it sends me the page i requested). I'm not using any i8n. Here are my versions
Kid 0.9.5 TurboKid 0.9.9 TurboGears 1.0.1 SQLObject 0.7.3 CherryPy 2.2.1 This is sort of a side note, but based on my experience with the 2 sites i have in production right now (specifically the unacceptably long load times for rather simple pages), I'm seriously considering not developing in TG any more in the future. I'll probably take another look when v2 comes out - maybe things will be faster then, but as it stands right now the performance is simply unacceptable. I don't see any reason why i should consistently get 4+sec render times on a mid-level VPS runnin a TG site that does little or nothing complicated or outside the normal use cases. I'll try refactoring the sites with Genshi (possibly Cheetah or Mako if that doesn't help), but i'm a little bitter that the 'out-of-the-box' solution is so slow. Either I'm doing something wrong or TG was designed for sites with access to faster hardware than i can afford for the types of development I do. On May 13, 6:39 pm, Alberto Valverde <[EMAIL PROTECTED]> wrote: > On May 11, 2007, at 3:41 PM, kerinin wrote: > > > > > I did - i'm getting 5 minute load times almost exclusively as a result > > of Kid. The site is almost all widget based, so without refactoring > > to remove the widgets I can't really tell how much is just Kid being > > slow and how much is the widgets being slow. > > I'm not sure what you mean by "load times"... is the time taken to > actually load the app? (import time). If this is the case then moving > to plain templates would probably help since TG widgets compile their > templates when the widget class is built if the template is being set > as a class attribute. In this case I think it might be easier to > change this behaviour in the the widget code so that it lazy-compiles > them, specially if there are many widgets you need to "port". I'll > try to get a go at fixing that, would you mind opening a ticket for > it :) > > However, if you're measuring the actual time it takes to render a > page (request time) then I doubt the overhead introduced by widgets > is significant, specially if profiling says that most time is being > spent inside kid functions. > > Alberto --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
