| *** ecloud is now known as ecloud_wfh | 06:26 | |
| dcaliste | Hello chriadam and jpetrell, could we move the meeting 20 minutes later? | 06:47 |
|---|---|---|
| chriadam | dcaliste: sure | 06:48 |
| dcaliste | Thank you chriadam, I hope you're well. | 06:48 |
| chriadam | not 100% recovered - injured my back/neck. but recovering | 06:49 |
| chriadam | but thanks, hopefully I continue to recover swiftly | 06:50 |
| chriadam | (not a bad injury, just some compressed nerves causing pain / numbness / loss of movement in my arm/hand) | 06:50 |
| chriadam | made typing impossible last week, but I can already type again now, so that's something ;-) | 06:50 |
| r0kk3rz | chriadam: mountain biking? | 06:51 |
| chriadam | r0kk3rz: slept in a bad position and then played squash. apparently I'm getting old. | 06:53 |
| r0kk3rz | lame, hope you get better | 06:53 |
| dcaliste | me too whish you quick recover, as quick as it can possibly be. | 06:54 |
| chriadam | cheers :-) | 06:54 |
| dcaliste | Hello chriadam, and jpetrell, sorry for the delay, I'm ready! | 07:29 |
| chriadam | dcaliste: hi, no problem :-) I hope you had a good week | 07:29 |
| chriadam | due to being away, I didn't manage to merge the gpg plugin etc | 07:30 |
| dcaliste | yes, thanks, discovering delta wing fly last Sunday, but it's another story. I've a question about installation plan: as far as I understood, secret will not be shipped by default immediately. | 07:31 |
| chriadam | I will do that this week. IIRC the license of that plugin was GPLv2 - does that pose a problem for us, jpetrell, or is it ok to merge that one? note that it's packaged separately. | 07:31 |
| chriadam | delta wing fly? as in, a new species of fly? or do you mean, you flew in a delta-wing aircraft? | 07:32 |
| dcaliste | Second part, I don't know how it's called in English, but yeh, flight, not fly ;) | 07:33 |
| chriadam | dcaliste: yes, my understanding is that sailfish-secrets is currently planned to be installable from store (similarly to the way exchange activesync is currently installable from store, except that sailfish-secrets would be available to anyone, not just paid customers) | 07:33 |
| chriadam | dcaliste: sounds intense, I know they can pull some pretty high-G-force maneuvers | 07:33 |
| dcaliste | So my question, currently in account settings, there is a new section about digital signature, with a combobox. | 07:34 |
| dcaliste | jpetrell suggested in bitbucket not to show the full section if secret is not installed (correct me if I misunderstood). | 07:35 |
| dcaliste | Should it show the section instead and a placeholder suggesting going to store ? | 07:35 |
| dcaliste | Last week, I've moved everything for account setting behind loaders, but I was undecided how to handle the case when both secrets and GPG plugin are not installed. | 07:36 |
| chriadam | this sort of design question probably needs feedback from jpetrell / mschuele. I guess there are two options: | 07:36 |
| chriadam | 1) show placeholder as you describe | 07:36 |
| chriadam | 2) show nothing, for now, until we want the feature to be more discoverable (i.e. until we have had more time to test and polish, etc) | 07:37 |
| jpetrell | I would not show anything in settings, there are variants of OS without Jolla Store | 07:37 |
| chriadam | true | 07:38 |
| dcaliste | jpetrell : ok, now I see the full reasons (no store, more testing…), sounds reasonable. | 07:38 |
| dcaliste | So, I'll create a new DigitalSection.qml containing the section header and the combo and a JS function to test if Secrets is available. | 07:39 |
| jpetrell | just checking existence of top-level package that requires all the rest, secrets, pgp, etc. is enough | 07:42 |
| chriadam | my brain is still catching up with current state - what are the action points for me? is it just review+merge+tag the sailfish-secrets gpg plugin? | 07:43 |
| *** zbenjamin_ is now known as zbenjamin | 07:44 | |
| dcaliste | jpetrell: I was thinking about using JS to actually import the Sailfish.Secrets QML module, like for Feedback in office: https://github.com/sailfishos/sailfish-office/blob/00e1bac3ca90a59a46a7a7e18f462421d1124b69/plugin/PDFView.qml#L216 | 07:45 |
| chriadam | oh, I guess I also need to review https://codereview.qt-project.org/240384 and poke Matt about that one | 07:46 |
| dcaliste | chriadam: you're right, and I need review for https://codereview.qt-project.org/240384, without it nothing will work ! | 07:46 |
| dcaliste | you're a bit quicker ;) | 07:46 |
| chriadam | I'm not sure how much time Matt has spare at the moment for such reviews, unfortunately, so might need pvuorela to also take a look at that one | 07:47 |
| dcaliste | chriadam: yes, thank you. Besides, this PR in upstream messaging framework is a bit to bit rebase from the Mer one that Matthew Vogt already looked at and improved last year. | 07:49 |
| dcaliste | That being said, it's not hurting to have other reviewers and comments ! | 07:50 |
| chriadam | great :-) | 07:50 |
| dcaliste | jpetrell: I've put every request in account settings behind loaders, it's cleaner now and only required objects are loaded. Give me your opinion when you have time. | 07:52 |
| dcaliste | Nevertheless the page is still slow to load. Looking at the code, it seems to me that the new animatorPush() is not used because there is a wait before the push() for the whole page to be loaded. | 07:53 |
| jpetrell | dcaliste: great! yeah will try to find time. we are a bit pressed with sailfish 3 schedule | 07:53 |
| dcaliste | Do you plan to ugrade this part to animatorPush() ? | 07:53 |
| jpetrell | dcaliste: account pages are pre-created before pushing to get some account-specific info, would need to separate the loading of the model/data and UI | 07:55 |
| jpetrell | would make sense performance-wise | 07:55 |
| dcaliste | It's just an open question, I was curious to see that this page is still slow to open while others are now very reactive. Besides, opening the page immediately and having the content to appear gradually may have some effect on the part I'm dealing with. | 07:57 |
| chriadam | I wouldn't spend too much effort on accounts settings pages until we know for sure what the plan is with the accounts framework rewrite | 08:00 |
| chriadam | a lot of the "hoops" which are jumped through are related to the fact that with current a&sso we have to create an identity separately (and prior) to the account creation, and roll back both on error, etc | 08:01 |
| chriadam | and much of the complexity will vanish immediately if we choose to go with the new accounts framework... | 08:01 |
| dcaliste | chriadam: ok, cool. Will it be related to QtAccount ? I'm wondering because I may rewrite the QMF patch to make it more clean for upstream inclusion. | 08:03 |
| dcaliste | As discussed with pvuorela yesterday, I'm evaluating now the work to create an interface in QMF for account management, implemented by two plugins, the current full built-in SQlite backend or the patched QtAccount and SSO framework. | 08:04 |
| chriadam | dcaliste: unrelated to the qmf accounts per-se, although we will need to patch qmf to support the new accounts framework | 08:05 |
| chriadam | btw you have access to the rewrite prototype, in sailfish-components-accounts:rewrite branch. it doesn't belong there, but that's where it lives for now until we decide to properly productise it... | 08:06 |
| chriadam | but no decision has been made there yet AFAIK | 08:06 |
| dcaliste | Ok thanks, I'll give a look later. I'll put you, chriadam, in the loop also about discussion for an interface in QMF for account related storage. Maybe this plugin thing for account management makes even more sense if we have a third way to actually store account related settings. | 08:08 |
| chriadam | true, that may indeed be the case | 08:09 |
| dcaliste | This would make the patch much less invasive in QMF core source by just providing an new interface implementation. | 08:09 |
| dcaliste | chriadam: I've encountered a bug in user input handling from secrets. While testing repeatedly the "generate a new key pair" button in account settings, the system window asking for a new passphrase was appearing at each demand, up to a certain point. | 08:11 |
| chriadam | up to a certain point? i.e. did the lipstick-security-ui process crash and then fail to be restarted until you restarted lipstick? | 08:12 |
| dcaliste | When the window was not appearing anymore and the request time out (I guess), but still returning a request.status == Suceeded with an empty passphrase. | 08:12 |
| chriadam | well that's technically two bugs :-P | 08:12 |
| dcaliste | Yes, to recover, I need to restart lipstick. | 08:12 |
| chriadam | damn, I thought we fixed that issue | 08:12 |
| chriadam | if you can get a backtrace from lipstick-security-ui that would be greatly appreciated | 08:13 |
| dcaliste | Otherwise I don't have the system window anymore for whatever request I do. | 08:13 |
| dcaliste | Ok, thanks for the process name to look at. | 08:14 |
| chriadam | do you ahve acccess to lipstick-jolla-home? | 08:14 |
| chriadam | it lives in that repo | 08:14 |
| dcaliste | Can it be run from command-line or I should attach a GDB to it after launch ? | 08:14 |
| chriadam | currently, it connects to the secrets process... in the future, I think we need to make it a service, and secrets should be a client of it, and we should use systemd to automatically restart it etc when required (and somehow force secrets to reconnect to the "new" one) | 08:15 |
| chriadam | I don't recall how it's run currently, presumably a systemd unit | 08:15 |
| chriadam | attaching gdb is probably the best way | 08:15 |
| dcaliste | No, I don't have access to lipstick-jolla-home. I will try to see how it's launch, I'm far from understanding properly how systemd is running, but it's a good occasion to improve this (lack of) knowledge. | 08:17 |
| dcaliste | Besides, Attaching will easily produce a backtrace if indeed the issue is a crash at one point. | 08:17 |
| dcaliste | From secret side, is it the passwordagentauthplugin that I should look at ? | 08:18 |
| chriadam | yes | 08:18 |
| chriadam | check /usr/lib/systemd/user/user-session.target.wants/lipstick-security-ui.service | 08:19 |
| chriadam | for the systemd unit | 08:19 |
| dcaliste | Ok, if I can reproduce it easily, I'll try to gather as much info as possible from the pointers you gave me. | 08:19 |
| chriadam | that'd be great, thanks. hopefully the debuginfo is available in the obs repo | 08:20 |
| chriadam | ooi what version of lipstick-jolla-home do you have? | 08:21 |
| chriadam | I fixed a crash bug in version 0.37.8 | 08:21 |
| dcaliste | It's 0.38.1, the one of 2.2.2.2. | 08:21 |
| chriadam | ok, a new one then | 08:22 |
| chriadam | thanks for that, hopefully it's not hard to repro | 08:23 |
| chriadam | did you have anything else to discuss? if not, I'll follow up on those two action points for me (review the qmf cr, review+merge+tag the gpg plugin) this week. | 08:23 |
| dcaliste | Just one (related) question: is there a UserInputRequest() option to ask for new passphrase input, so the dialog is shown twice to confirm that passphrase is the one we want to actually enter ? | 08:24 |
| dcaliste | I know there is the show password in clear option, that can help to ensure that we define the right passphrase though. | 08:25 |
| dcaliste | So it's more a design related question ;) Not in a hurry to answer though if it's late for you. | 08:26 |
| chriadam | ISTR that the lipstick-security-ui supports such an option | 08:26 |
| chriadam | not 100% sure if we expose that one in the client API | 08:26 |
| chriadam | dcaliste: passwordagentplugin.cpp has "enterNewPasswordText" and "repeatNewPasswordText" prompt data | 08:28 |
| chriadam | so it supports it. but as mentioned, not sure that we expose such an operation in the API | 08:28 |
| dcaliste | Ok, I'll see if I put the repeatNewPasswordText to not empty data, I get a second confirming window. | 08:29 |
| chriadam | could indeed work | 08:29 |
| chriadam | ok, I have to head home now. thanks again for your work. sorry for not emailing you to notify that I was unavailable last week, but I couldn't type :-/ | 08:31 |
| chriadam | I hope you have a great week :-) | 08:31 |
| dcaliste | chriadam, no problem at all, I understand and wish you to fully recover. Have a nice week. | 08:31 |
| chriadam | :-) | 08:31 |
| dcaliste | jpetrell: maybe it's not the right moment and you would prefer to discuss this after SFOS3 is out, tell me if so. As said, the performance regression in Calligra 3.1.0 packaging has been found so I think that Calligra latest can widely tested and schedule for upgrade. What do you think ? | 08:41 |
| *** frinring_ is now known as frinring | 09:35 | |
| Tekk_ | Is the Xperia X still using this for its base: https://developer.sony.com/file/download/software-binaries-for-aosp-marshmallow-android-6-0-1-kernel-3-10-loire/ ? | 12:53 |
| Tekk_ | An update borked my camera, so I'm going to try a fresh reflash. | 12:53 |
| Tekk_ | (Already tried factory reset) | 12:53 |
| r0kk3rz | abranson: ^^ | 13:01 |
| Tekk_ | ...we also have the problem of putting the device into fastboot mode. Forgot how much of a pain this is. | 13:04 |
| r0kk3rz | it should be easy to put into fastboot mode | 13:05 |
| Tekk_ | Should be, but it just insists on booting anyway :) | 13:06 |
| r0kk3rz | what are you doing? | 13:07 |
| Tekk_ | Turn off 15 seconds, hold down volup, plug in USB. | 13:07 |
| Mister_Magister | Tekk_: xperia always had pretty bad ways of doing things | 13:07 |
| r0kk3rz | iirc its volup + pwr | 13:09 |
| Tekk_ | volup+pwr+plug in? | 13:10 |
| Tekk_ | I can try it. | 13:10 |
| r0kk3rz | no need to plug | 13:11 |
| Mister_Magister | r0kk3rz: its usually that simple | 13:13 |
| Tekk_ | That gets me an extra couple vibrations. | 13:15 |
| Tekk_ | Still boots :( | 13:15 |
| Mister_Magister | sony's state of art | 13:17 |
| mal | https://developer.sony.com/develop/open-devices/get-started/flash-tool/useful-key-combinations/ | 13:17 |
| mal | it's simply volume up and then plug the usb cable | 13:18 |
| Tekk_ | Should be. | 13:18 |
| Tekk_ | ....I know this sounds weird | 13:19 |
| Tekk_ | But does it just not like going into fastboot mode with an unlocked bootloader when sailfish is installed? | 13:19 |
| mal | so you keep the volume up pressed and then only plug the cable and keep the volume up pressed for a moment until you see the blue led, not pressing of power button | 13:19 |
| r0kk3rz | mal: that information isnt complete, i know its wrong for scorpion at least | 13:19 |
| mal | r0kk3rz: how is it wrong? | 13:20 |
| Tekk_ | mal: Yeah, doesn't work for me. It goes straight to red LED and boots as soon as I plug the cable in. | 13:20 |
| r0kk3rz | mal: as in, that isnt the combination | 13:20 |
| r0kk3rz | Tekk_: vol key works ok? | 13:20 |
| Tekk_ | Yeah. | 13:20 |
| Tekk_ | Maybe if I can get it into recovery it has an option to reboot into fastboot. | 13:21 |
| mal | Tekk_: I also get red led after volume + usb plug but it changes to blue after a second or two | 13:21 |
| mal | Tekk_: this is on xperia x | 13:22 |
| Tekk_ | mhm | 13:22 |
| mal | maybe you release volume button too early? | 13:22 |
| r0kk3rz | hold it until it goes blue | 13:23 |
| Tekk_ | I'm holding the volume button until it gets to the sailfishOS charging screen. | 13:23 |
| mal | so either your volume key is broken or you are using it wrong | 13:23 |
| Tekk_ | I'm leaning towards the latter, considering how I can adjust volume fine. | 13:24 |
| mal | boot to UI and check that the button works | 13:24 |
| mal | ok | 13:24 |
| Tekk_ | Flipping it right now :) | 13:24 |
| Tekk_ | Volume up is the one *towards* the power button? | 13:24 |
| yofuh | had a bad time finding out how it works my self, but it's actually quite simple once you know | 13:24 |
| Mister_Magister | mal: cant he just run reboot bootloader from sfos? | 13:24 |
| mal | the instruction on sony website is very clear | 13:25 |
| Tekk_ | Hm. Do I need to have the middle click down too? | 13:25 |
| Mister_Magister | yofuh: not as simple as on motorola phones tho :P | 13:25 |
| Tekk_ | If you run your finger along the volume button, there's a click for volup, then another click, then another for voldown. | 13:25 |
| Tekk_ | I wonder if this cable is bad.. | 13:26 |
| yofuh | just volup on the disabled phone and then plug in usb while buton is pressed, dont use the powerbutton | 13:26 |
| Tekk_ | Sec, gonna grab another just in case. | 13:26 |
| mal | Tekk_: so you just press volume up button (the upper side of the button) and keep it pressed and plug cable, wait for a moment until led changes and the release the volume up button? | 13:27 |
| Tekk_ | mal: Hold on, let me run through it one more time to make sure the order is right. | 13:34 |
| Tekk_ | Plus new cable | 13:34 |
| Tekk_ | So cable's plugged into the laptop. | 13:34 |
| Tekk_ | Hold down volup | 13:35 |
| Tekk_ | Plug in cablne. | 13:35 |
| Tekk_ | Red LED turns on. | 13:35 |
| Tekk_ | "Your device is unlocked, device will boot in 5 seconds" screen comes up | 13:35 |
| Tekk_ | After 5 seconds sony screen comes up, red LED stops. | 13:35 |
| Tekk_ | White (blue?) LED starts | 13:35 |
| Tekk_ | After a few seconds, LED dims slightly. | 13:36 |
| Tekk_ | Then Sailfish charge screen. | 13:36 |
| Tekk_ | I suppose the low level firmware could be broken. I remember having a hell of a time getting this to work back when I first got it too. | 13:37 |
| Tekk_ | So if there's a way (like there is in android) to enable a "reboot to fastboot/reboot to recovery/reboot to bootloader" that'd be great :) | 13:38 |
| abranson | that's not the usual behaviour. it's usually device off > press vol+ > plug cable > blue led. no intermediate stages | 13:38 |
| Tekk_ | Guess I'll just keep trying until it works, which I think is how it went last time.. | 13:41 |
| elros1 | @Tekk_: can't you just type in sfos "reboot bootloader"or it not works for xperia phones? | 13:41 |
| abranson | it sounds like it's not registering the vol+ | 13:42 |
| Mister_Magister | elros1: same what i said | 13:43 |
| elros1 | right | 13:44 |
| Tekk_ | Sorry, work got in the way. reboot {fastboot,bootloader} did nothing | 13:58 |
| Tekk_ | By which I mean they rebooted normally. | 13:59 |
| Tekk_ | Just ran the jolla-camera app from a terminal for fun. No obvious culprit in terminal output, but it does have an error about assigning undefined into a pulley. | 14:27 |
| Tekk_ | 2.2.1.18 is still downloading though. | 14:27 |
| Tekk_ | So maybe it'll be different on Nurmonjoki | 14:27 |
| Tekk_ | Ooh, Miracucam is a bit more useful. Seems like it's a gstreamer issue somewhere. "GStreamer encountered a general supporting library error" and "Camera is not ready" on there. | 14:32 |
| abranson | Tekk_: /system/bin/logcat might give you more of an idea of what's going on | 16:05 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!