Wednesday, 2022-02-02

T42_<Delbert Underwood> Forwarded from unknown: 𝚏𝚎𝚎𝚕 𝚒𝚖𝚙𝚛𝚎𝚜𝚜𝚎𝚍 𝚠𝚒𝚝𝚑 𝚃𝚑𝚎 𝚠𝚘𝚗𝚍𝚎𝚛𝚏𝚞𝚕 𝚠𝚘𝚛𝚔 𝚠𝚎𝚕𝚕𝚍𝚘𝚗𝚎 . 𝙸 𝚝𝚘𝚘𝚔 𝚊 𝚛𝚒𝚜𝚔 𝚝𝚘 𝚐𝚒𝚟𝚎 𝚝𝚛𝚢 𝚘Ϝ $➐⓪⓪, 𝙸 𝚛𝚎𝚌𝚎𝚒𝚟𝚎𝚍 𝚊 𝚙𝚊𝚢𝚖𝚎𝚗𝚝 𝚘𝚏 $➊➍,⓪⓪⓪ 𝚃𝚑𝚊𝚗𝚔 𝚢𝚘𝚞 �07:51
T42_<Delbert Underwood>
jules[m]<deathmist> "@jules[m] I'm interested in..." <- Thanks :)09:09
maldeathmist: then we probably need to do some magic to strace for example that gst-inspect-1.0 so we see if loading everything works (there was some need for special parameters to make it work correctly with strace)09:13
Verevkadeathmist: Need help11:00
Verevka* No git url specified, assuming
Verevka* pulling updates...11:02
Verevka!! Could not get version for miniaudiopolicy: fatal: --unshallow on a complete repository does not make sense11:02
VerevkaERROR: Failed to run post-build-mw hook for xiaomi-sagit!11:02
T42_<XAP2P> :D11:05
deathmistVerevka: replacing "--mw=miniaudiopolicy" with "--build=hybris/mw/miniaudiopolicy" should fix build, I'm not sure if using --mw this way is even correct11:37
maldeathmist: -b hybris/mw/miniaudiopolicy is correct afaik11:45
maldeathmist: sorry, misread your comment11:46
deathmistVerevka: something is wrong with sfbootstrap and passing arguments to build_packages: it appears to only read the first argument. I'll investigate11:49
deathmistVerevka: fixed passing of >1 args to build_packages via, now you should also be able to do --mw=miniaudiopolicy still but simply additionally also pass --no-auto-version (or -N) for short so it won't try to correct version; this was a problem for you because the repo has no tags so it can't correct the version12:24
VerevkaThanks, I'll try now12:25
deathmist(btw you can drop the --spec arg for miniaudiopolicy since that is the only rpm spec file in the repo, build_packages can find it)12:29
deathmistI specified --spec files for fpd on my port to avoid having to build extra HAL side & RPM packaging for fake_crypt which may be needed by newer devices12:32
*** Ischwitch is now known as Ingvix14:38
deathmistrinigus: have you seen "hybris/mw/hwcrypt/main.cpp:81:68: error: no member named 'getErrorCode' in 'keystore::KeyStoreNativeReturnCode'" before? I get this on 16.0 (A9), so I assume it's only been tested with A10+?16:22
rinigusdeathmist: no, I haven't. I am using it with aosp10, correct16:43
rinigusdeathmist: maybe you could check whether there is something similar in aosp9? I wonder if you have system/security/keystore/keystore_cli_v2.cpp for it16:45
deathmistrinigus: yeah:
deathmistI suppose result.isOk() is a thing16:57
deathmistoh wait that's just for checking it, eh gonna have to int32_t(result)16:58
deathmistwe'll need some ErrorCodeFromResult() method everything uses so it can be wrapped with a #if ANDROID_VERSION_MAJOR > 9 or similar17:00
maldeathmist: did you try strace or something with gst-inspect?17:05
malyou probably need the remove the cache file before trying it17:05
deathmistmal: not yet, I'll check it a bit later tonight17:14
T42_<BenKerry> Ist there a SailfishOS Port Form Galaxy S5?17:35
rinigusdeathmist: I can see that at least your link to keystore_cli_v2 doesn't have getErrorCode call.18:23
*** Daanct12 is now known as Danct1218:24
rinigusdeathmist: looks like it was added here:
rinigusdeathmist: reading your reply again - yes, ErrorCodeFromResult wrapper should work18:29
deathmistrinigus: would you mind writing one? I suppose I can take a crack at it today still but tomorrow my military service still continues for another month or so :)18:39
rinigusdeathmist: ha! will see. I suspect I'll forget about it and we will be back with the discussion regarding it in one month as I just started working on other issues. just to save us in future, I'll write an issue describing a problem19:16
deathmistrinigus: looks like keystore droid service doesn't want to stay as running for me, so I can't test hwcrypt (or use it for that matter) :/20:39
deathmist(keeps dying and restarted by d-h-i every 5 seconds)20:40
rinigusdeathmist: weird. I thought that keystore is a vital for keeping android bits working happily in the port. or am I missing something?20:41
deathmistI think android 9 ports used to straight out just disable it20:41
deathmistthat's what I had in my disabled_services.rc anyway20:41
rinigusdeathmist: let me check for tama20:42
rinigusdeathmist: doesn't look like keystore is disabled in tama nor seine:
deathmistrinigus: but that's sparse-10, aka android 10, I'm on android 920:47
rinigusdeathmist: not for
deathmisthmm, guess it just kept dying for me and I couln't figure out why and just disabled it to stop battery drain etc since other parts seemed to work at the time, but now for encryption support it's a blocker :/20:49
deathmist'll try using the linked disabled_services.rc file and see what happens20:50
rinigusyeah, that one you would want to get running20:51
deathmistwith that file I've got the following crash looping: time_daemon, keystore & vendor.dataqti20:54
deathmistdoesn't look like keystore got any happier with less disabled services20:54
deathmisteh I don't really know how I'd debug this, maybe hybris-17.1 port would fix this issue as well20:55
deathmist(considering this device's vendor HAL still got updated as well for Android 10)20:56

Generated by 2.17.1 by Marius Gedminas - find it at!