15:01 <popey> #startmeeting Docviewer meeting
15:01 <meetingology> Meeting started Thu Jan 15 15:01:49 2015 UTC.  The chair is popey. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
15:01 <meetingology> 
15:01 <meetingology> Available commands: action commands idea info link nick
15:01 <popey> How are you?
15:02 <popey> You had exams this week?
15:02 <sverzegnassi> Fine, thanks! I took the exam two days ago, and I think I did it well!
15:02 <popey> Excellent!
15:02 <sverzegnassi> Now I have finally more time to work on code :D
15:02 <popey> \o/
15:02 <popey> Even better ㋛
15:03 <popey> So, I emailed the sailfish-office guy about the license question. No reply yet.
15:03 <popey> Will give it a couple more days and poke again, in case they're on holiday.
15:04 <sverzegnassi> Uh, hope to get an answer soon!
15:04 <popey> We need a backup plan in case the answer is "no".
15:04 <popey> One option is for us to re-license the doc viewer as GPLv2..
15:04 <popey> Any other alternatives? :)
15:05 <sverzegnassi> During the xmas holidays I've worked on some experimental code
15:05 <sverzegnassi> I think we could make things a little bit faster using some cache
15:06 <sverzegnassi> (e.g. testing some relevant Android PDF reader, I've seen they use up to 300MB of RAM on my Nexus 5)
15:06 <popey> ouch.
15:08 <sverzegnassi> We have an huge margin to use cache (I was thinking at 1/10th of the whole RAM - on krillin it would be 100MB for caching and around 80MB for UI and the rest of the app)
15:10 <sverzegnassi> of course, we need to find the best trade-off for our purpose
15:13 <popey> Is this something that we could dynamically configure?
15:14 <sverzegnassi> Do you mean an item in a setting page so user can set the value he prefers?
15:14 <sverzegnassi> or something which is automatically set?
15:15 <popey> Either.
15:17 <sverzegnassi> AFAIK there is nothing in Qt that gives us that information, so we should "manually" get that value (from /proc/meminfo)
15:18 <popey> Ugh.
15:18 <popey> We don't want the docviewer to consume all the ram, and cause other apps to all be killed
15:20 <sverzegnassi> But, assuming that PDF page fetching is an expensive process even on PCs, it might be worth to have some cache somewhere
15:21 <popey> true
15:23 <sverzegnassi> An alternative is to use a fixed value, e.g. 64MB, so that we can still pre-fetch pages and be able to resize pages without asking a new rendering (which takes a lot)
15:23 <popey> I think we need to test this out on each device really, with various PDFs
15:24 <popey> I fully expect someone to find a giant PDF which kills the app/phone at some point
15:24 <popey> But in my mind the use case is "Download pdf of flight ticket and view on device" over "Read entire 8086 instruction set book"
15:25 <popey> We should focus more on the former than the latter in my eyes.
15:25 <sverzegnassi> Ahahah
15:27 <sverzegnassi> I'll try to review the current code we have, re-implementing pages zoom and fixing the last issue we found
15:27 <sverzegnassi> Did you have some news about confinement issue?
15:27 <sverzegnassi> (we talked about it in December)
15:27 <popey> Remind me, sorry.
15:28 <sverzegnassi> Pdf documents are copied in /user/.cache/ and opened from that path
15:28 <popey> ah yes. Sorry. no.
15:28 <popey> I will re-investigate.
15:29 <sverzegnassi> Ok, I look forward for good news :)
15:30 <popey> :)
15:31 <popey> I'm also building a set of test cases for each app
15:31 <popey> for manual testing
15:31 <popey> What's your preferred pdfs you test with?
15:32 <popey> I have a load, but would be good to have some from different sources, and in different languages.
15:33 <sverzegnassi> I use the Ubuntu 14.04 manual, some document (with tables) done with LibreOffice, and some books I need for university
15:34 <sverzegnassi> (there was a big huge document of about 1200-1500 pages, but that was just an experiment)
15:34 <popey> heh
15:34 <popey> I'll grab some various PDFs and collect together for further testing.
15:35 <popey> Anything else?
15:37 <sverzegnassi> The biggest issue are those we already talked. There are a few minor things (e.g. empty-state) but they should be ready in a week.
15:37 <sverzegnassi> So, that's all I think.
15:38 <popey> Ooh! Empty state would be nice to get landed.
15:39 <sverzegnassi> Yup, just need to find out the reason why it brokes everything
15:39 <popey> hah.
15:39 <popey> if you need me to find someone to help, let me know.
15:40 <sverzegnassi> Ok, thanks! I already have an half solution, I need to find the other half...
15:42 <popey> :)
15:42 <popey> Ok, well have a great weekend, you know where I am if you need anything! :)
15:43 <sverzegnassi> Have a great weekend you too! Thank you!
15:43 <sverzegnassi> Cheers!
15:43 <popey> #endmeeting