Saturday, 2020-05-30

riniguspiggz: just debugging a bit further. looks like I can list the stored fingers using enumerate (made some small additions). cannot remove though (yet)09:11
riniguspiggz: I can also remove single fingerprint from store.09:26
rinigus... and remove them all one-by-one. I'll ask around, maybe its some bug with remove(0,0)09:35
T42sky_sagit was added by: sky_sagit11:14
riniguspiggz: from help of Sony AOSP devs - Android is removing by iterating one by one, last FP is removed using remove(0), the one that fails for me even with one FP. Marijn from Sony AOSP wanted to check out if there is some issue on their side with it.11:16
riniguspiggz: I will add the support for listing stored FPs and should add also to your test GUI support for listing+removal one by one.11:17
rinigusI presume you work on auth bit11:17
riniguspiggz or someone else: do you know whether SFOS Settings list fingerprints and whether they allow to remove them one by one?11:18
riniguspiggz: how can I hook SFOS settings to FP? as you did for enrolling them11:18
piggzrinigus: thx... send a PR ...12:15
piggzi willw ork on auth bit later, been mtb :)12:15
piggzwe need to enumerate, then convert them to finger1, finger2, finger3 etc for sfos i thin12:16
riniguspiggz: I wonder where should we keep a map between FPids and names...12:20
rinigusbut OK, I continue with enumerate12:20
riniguspiggz: I am cleaning up the code a bit. question: in the logs, do you also have /data/system/users/0/fpdata//user.db <- double //? I'd drop trailing slash in the path that is given to android12:29
piggzrinigus: i was thinking about where to keep a map between names and ids, bt then thought we dont need to....we just enumerate them from the device, and name them in order ?12:57
piggzrinigus: you will probbaly beat me to impementing everything, i wont get a chance until later!!12:58
riniguspiggz: looks to me the API expects finger_string -> uint32_t id on daemon side though. we can't make up the map as user may want to remove some specific finger... if only we could let the client to keep the map12:59
riniguspiggz: but let me submit PR with cleanup and enumerate and then we can discuss13:00
riniguspiggz: just pushed into existing PRs (daemon + test code). will make a break now. feel free to review ...13:19
piggzrinigus: looking.....the Enumerate call....that isnt part of the closed-source api is it?13:20
piggzthey use GetAll i think13:21
piggzgoing out for a few hours now....will do more later13:21
riniguspiggz: that's correct. thinking of it, I will hide it this part and make it internal14:12
T42<ankaos> device is bootloop. device not open.15:33
T42<ankaos> this is kernel config15:35
T42<ankaos> @elros34 where is the problem?15:48
riniguspiggz: close with the list, have to make a break16:00
T42<ankaos> rinigus: please look links16:00
T42<elros34> @ankaos seems like problem is that you removed well working droid and kernel repos and created everything from scratch without any reason16:23
T42<ankaos> @elros34 [@ankaos seems like problem is that you removed …], `'Enes'` `'hodja'`, [30.05.20 18:33]16:29
T42<ankaos> this is bad?16:29
piggzrinigus: ready to start....16:47
rinigus5 mins16:49
rinigusand I'll submit current state16:49
riniguspiggz ^16:50
piggzrinigus: gr816:50
riniguspiggz: daemon submitted, not tested much yet16:51
T42<elros34> @ankaos all these wlan related changes in your defconfig looks incorrect for me: like enabling pronto/prima or using builtin wil6210 and so on16:56
riniguspiggz: test submitted. had to cut short as I have a small meeting in few minutes.16:59
riniguswill be able to join in 20-45 mins, probably16:59
piggzrinigus: np, thx for helping17:01
piggzim sure we can nail this shortly17:01
piggzrinigus: identifying works17:07
piggzits a hack atm, need some other groundwork17:07
riniguspiggz: sounds great!17:09
piggzrinigus: that is pushed .... you currently have to re-add the finger every boot becuase the GetAll isnt implemented17:10
riniguspiggz: check prs17:11
piggzon it17:12
piggzrinigus: i think GetAll needs to return "finger1,finger2,finger3..." .... but i might merge your PR than work on the mapping between strings and id's17:15
piggzwhat do you think?17:15
riniguspiggz: please do17:15
riniguspiggz: can join when you are ready. to get the mapping, we will need to store strings with which keys were created by clients by calling `Enroll(const QString &finger);`.17:52
riniguspiggz: should I look into the test client?18:13
rinigusor anything else I could help with?18:14
piggzrinigus: im just thinking ....18:15
piggzrinigus: i guess we need a file store18:17
riniguspiggz: for string <--> uint32_t?18:17
piggzyes .... until you mentioned using the name the client sends , i was going to fudge it internally and assume fingerX18:18
rinigusI agree. we do need it, assuming that jolla's settings does send some proper unique string18:19
riniguspiggz: looking at API, does Remove have an argument? shouldn't there be one string?18:20
rinigus(all relates to that filestore with the map)18:20
piggzrinigus: I assume so .... reading mce will say for sure18:21
piggzprobably some of those functions also need to return int not void18:21
riniguspiggz: they haven't sent you privatly API?18:21
piggzrinigus: no, im reverse engineering it!18:22
riniguspiggz: OK, but then we can make at least the test app working as we expect and later see how well it maps to SFOS Settings...18:23
riniguspiggz: do you know how identification works? which api is for that?18:23
riniguswith the map, where are expected to store it? somewhere in /var/cache?18:24
rinigusmaybe /var/db18:25
rinigusjust looked in /var and found .... /var/yp!!!18:26
rinigusnow that's a blast from the past18:26
piggzrinigus: just testing load/save code19:45
piggzrinigus: first test looks good!19:50
piggzi will push and you can test too ... i rebooted and was able to unlock19:51
piggzmal: do you want to jump on the community fingerprint band wagon for your port?20:01
malpiggz: I have no community device with fingerprint support20:43
riniguspiggz: we should make a list of outstanding issues and I can help with fixing them20:47
rinigusas packaging android bits (pull into sailfish-fpd-community?) and the rest.20:47
piggzrinigus: sure, create an issue for each one20:48
piggzrinigus: yes, was thinking we could pull in that android file20:48
riniguswill make issues as I go along20:48
rinigusgood, will you do that?20:49
riniguspiggz: what's a plan with test code? leave it as it is?20:50
piggzrinigus: i will keep it for now, it may proove useful, but hopefully not needed much now ....20:51
piggzrinigus: i wonder how well it runs against the official fpd!20:51
riniguspiggz: how about removing FPs? is it possible one-by-one?20:51
riniguspiggz: re official fpd - good question20:52
piggzrinigus: that still needs implemented i think ... fell free!20:52
rinigusall thing looks bit unsecure. as soon as your phone is lying around unlocked, anyone can add FPs20:53
piggzrinigus: you have to enter the pin to enter a finger20:53
riniguspiggz: in official settings. not in DBus API20:53
piggzheh, i wonder how offical device copes with that20:54
rinigus... as your app can be used without pin code20:55
piggzyes .... but an unlocked phone could also issue dbus-send commands20:56
piggzrinigus: android code imported and pushed21:04
riniguspiggz: let's just warn / document. or at least remember that concern21:05
riniguslooks like I cannot remove FPs from official settings. otherwise it works. will start filing issues21:05
piggzrinigus: remove looks easy to fix....21:10
piggzsignature needs changed for a start21:10
riniguspiggz: great. I am off to sleep, will be able to continue tomorrow on where you stopped21:11
piggzrinigus: Remove() implemented and pished21:42

Generated by 2.17.1 by Marius Gedminas - find it at!