hvelarde wrote > > On 14/03/12 12:34, Laurence Rowe wrote: >> I've certainly never observed anything like this. Martin's performance >> tests >> from a few years ago showed that performing the transform in process was >> faster than doing so in a proxy. For Plone, plone.app.theming is the >> recommended deployment mechanism. >> >> There's not really enough information here to know why this might be >> happening. If you can isolate it and turn it into something I can easily >> reproduce in a standard plone.app.theming build I'll take a look. > > I made some additional profiling using a development buildout in my > machine and now I'm more confused: > > ncalls tottime percall cumtime percall filename:lineno(function) > 53 9.218 0.174 12.062 0.228 > transform.py:200(transformIterable) > > so, time spent on a laptop running with low memory in development mode > is half than time on a dedicated server running plenty of memory in > production mode; this doesn't make sense. > > could it be some issue with the lxml installed on the server? some issue > with it's compilation? > > BTW, I'm running Plone 4.1.3, Dexterity 1.1, plone.app.theming 1.0b9 and > lxml 2.3 on my machine. > > I will confirm current versions on the server later. >
plone.app.theming.transform.transformIterable should be called only once per request, I would suggest working out why is being called so many times. Laurence -- View this message in context: http://plone.293351.n2.nabble.com/huge-performance-issues-using-Diazo-tp7372056p7374959.html Sent from the Installation, Setup, Upgrades mailing list archive at Nabble.com. _______________________________________________ Setup mailing list [email protected] https://lists.plone.org/mailman/listinfo/plone-setup
