Tuesday, 2022-10-04

dcalisteGood morning pvuorela. Thanks for the merge of the placeholder issue.07:05
pvuoreladcaliste: heya07:05
dcalisteI would like to investigate a bit the slow down on week change when the model is instanciating a new week. I guess it's related to the creation of the various QML elements, but I'm a bit surprise because this slow down doesn't exist with the rotating model where the week QML object is reused but the actual content is reset.07:09
dcalisteI'm going to do some timing today or tomorrow and will report on the PR page.07:10
pvuorelai've been experiencing stuttering on all the versions so far when changing the week, though.07:10
dcalisteI think it was more fluid in a 10ii with the rotating model than with the flat one. But compared to a C I guess the slow-down was hidden by a better processor…07:11
pvuorelasure, better performing device in many ways.07:12
pvuorelathe qml profiler could be one option.07:13
dcalisteJust a question : how to you run the profiler ? I was thinking of adding timer in the QML, but it may be simpler with a profiler indeed.07:15
pvuorelafor that i need to check again the docs myself. it's been years since i've even looked at that direction :)07:16
dcalisteOk, never mind, I'll read the doc ;)07:17
dcalisteAbout the initialisation problem of the slide show, I tried yesterday to replace the WeekLayout by a label and removed the various bindings, just to keep the model and the instanciation / destruction of the objects. Strangely it continued to allocate 2 objects, destroy them, then allocate three, delete the last one and create another one with an index -107:17
pvuorelafor the slowness thinking if the plain delegates + layouting would be causing it. it's anyway some repeaters and js code.07:20
flypigdcaliste, sorry to interrupt your meeting, but I was hoping to briefly discuss the repository roundup with you as well; let me know if there's a good moment to talk about that please!07:20
dcalisteYeh, that's what I'm afraid of also. But I would like figures before rewriting everything with C++ hourlayouter ;)07:21
dcalisteflypig, sure I'm available almost all morning.07:21
flypigdcaliste, thank you. I was mostly wondering if you had time to do one for the newsletter this week. Don't worry if not, I can do it (poorly) myself, but I was hoping you might!07:23
dcalisteOh, sure, sorry I forgot about this Friday deadline :/07:24
flypigNo, no apology necessary. But does that mean you're still up for it?07:25
dcalisteI'll do it today and post it to you before the end of the day.07:25
flypigThank you so much as always!07:27
dcalisteflypig, no problem and sorry about forgetting it. Don't hesitate to ping me on Fridays if you don't see anything coming. Normally, I should warn you if I'm enable to finish for Friday or even at all.07:29
flypigUnderstood! I wasn't getting worried anyway... honestly ;)07:29
dcalistepvuorela, about the initialisation, do you think it's worth investigating why it is behaving like that ?07:35
pvuoreladcaliste: the create+delete+create sequence? sure it would be nice to get rid of, though i think it's more important that once changed to the week view the flicks are smooth.07:37
dcalisteYes, I agree. I'll investigate the slow down on week change first.07:37
dcalisteflypig and pvuorela, do you think we can go on on mkcal!30 ? https://github.com/sailfishos/mkcal/pull/30 It's about the removal of ::Ptr part in Notebook API in mKCal, creating an API break that requires to be propagated in nemo-qml-plugin-calendar, jolla-calendar and Buteo plugins. I've done it for all, except EAS one that I cannot access to.08:24
* pvuorela looking at flypig's direction for some final review :)08:31
dcalisteYeh, thanks.08:31
flypigOh, I'm a bit lost. So there are two things needed here? Review and changes to EAS?08:34
pvuorelaeas pr needing review, after that the whole set is ready.08:35
flypigJust looking for it...08:41
dcalisteThanks a lot both of you.08:41
flypigOh yes, I remember it now. I'll take another look.08:41
pvuorelathough tbh, some small part of me is still pondering if all that code churn is worth getting rid of the ::ptr. nice logic changes though, e.g. the writing to read-only notebooks has been strange.08:44
dcalistepvuorela, as you prefer, I can just make a PR from the logical changes not related to the ::Ptr removal. Actually you suggested the idea when discussing another PR ;) I'm pushing a bit on this one because I would like to restart working on separating ExtendedStorage from the actual storage implementation. And doing so need to know if we use ::Ptr for notebooks or not. Both being alright.09:25
pvuorelawell, at least the current version is already there...09:43
dcalisteIndeed, but if you think it's not going in a good direction, or more realisticaly, it's going in the good direction but at a high cost for not so much return on investment, I can withdraw it. I'm not really attached to it going in.10:03
*** fifr[m] is now known as Guest228515:03
*** fifr[m] is now known as Guest235820:22

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!