IRC log for #maemo-ssu on 20151018

02:59.12*** join/#maemo-ssu FlameReaper (~assassin@175.137.196.31)
02:59.41*** join/#maemo-ssu FlameReaper (~assassin@175.137.196.31)
04:57.51*** join/#maemo-ssu ShadowJK_ (~jk@212.7.198.194)
04:57.58*** join/#maemo-ssu gregoa_ (~gregoa@colleen.colgarra.priv.at)
05:11.26*** join/#maemo-ssu ZetaR (~user@2602:47:3fb:fc00:d5a0:81ac:c12d:3965)
05:14.52*** join/#maemo-ssu XDS2010 (sid1218@gateway/web/irccloud.com/x-hhdgdjokonotjpof)
08:16.57*** join/#maemo-ssu Pali (~pali@Maemo/community/contributor/Pali)
08:22.38*** join/#maemo-ssu LauRoman (~LauRoman@5-14-22-54.residential.rdsnet.ro)
08:44.48freemangordonPali: do you have any idea how to compare binaries?
08:45.02freemangordonI mean - in a sane way
08:45.04Palifreemangordon: objdump -D
08:45.15Paliand check assembler parts
08:45.27Paliideally try to use same compiler, version, compile flags, etc
08:45.39Palisomebody told me that gcc is deterministic in assembler output
08:45.40freemangordonyeah, I know that much :)
08:46.14freemangordonhmm, maybe I should tryt it
08:46.26freemangordonand see what the result is
08:46.44freemangordonanyway, I was asking if you know more easy way
09:31.09*** join/#maemo-ssu futpib (~futpib@176.104.194.78)
11:11.15*** join/#maemo-ssu jonwil (~jonwil@27-33-80-219.tpgi.com.au)
11:18.16jonwilhi
11:23.39freemangordonjonwil: hi!
11:24.07freemangordonjonwil: what did you use to compare mce binaries (stock vs FOSS)?
11:24.27jonwilIDA with some plugins that I forget which ones
11:24.33jonwildo you have a binary you want to compare?
11:24.38freemangordonjonwil: I am trying to put libspeex in cssu
11:24.44freemangordonyes, I have :)
11:25.36jonwilok so whats special about libspeex that you need to compare it to the stock binary?
11:26.24freemangordonthe one in maemo repos doesn't have sources
11:26.39freemangordonand there are some NEON optimizations compared to upstream
11:39.07jonwildid you reverse engineer that stuff or did you use patches from http://lists.xiph.org/pipermail/speex-dev/2011-September/008242.html ?
11:41.49freemangordonI applied the above patches on top of debian lenny speex
11:42.06freemangordonat least part of them as some stuff does not apply
11:42.34jonwilok
11:43.56freemangordonand it seems gcc in SB is different than gcc used to build speex
11:44.03jonwilI dont think trying to match what you have to the Nokia binaries is really possible since they probably differ too much.
11:44.10jonwilyeah its possibly also that
11:44.21freemangordondiffer too much? why should they?
11:44.46jonwilDo you have the same version of speex as Nokia started with?
11:44.50freemangordonyes
11:45.04freemangordonspeex (1.2~rc1-1) unstable; urgency=low
11:46.03freemangordonjonwil: http://pastebin.com/m1n7ZLUU
11:46.40freemangordonhmm, wasn't that guy (Jyri Sarha) on #neo900?
11:46.49jonwilpossibly, I forget
11:47.07freemangordonjusa?
11:47.40jonwilin any case since the patches from that ml post didn't apply cleanly, its clear that the patches in question aren't a direct match for what Nokia shipped (otherwise they would apply cleanly to the 1.2~rc1-1 version of speex)
11:48.09freemangordonthose patches are rebased on top of what was cuurent by the time of submission
11:48.45freemangordonNEON parts apply clean, it is just that http://lists.xiph.org/pipermail/speex-dev/2011-September/008240.html that does not apply
11:49.02freemangordon+#define SATURATE32PSHR(x,shift,a) (x) and co
11:50.35freemangordonthose are aimed to be applied on top several commits above whet we have in maemo
11:50.41freemangordon*what
11:51.09jonwilwell I cant find where I put the tools I used to do the diffing anyway
11:51.40freemangordon:(
11:51.43jonwiland they probably wouldn't work with the version of IDA I now have
11:51.52freemangordon6.7?
11:52.05jonwil6.6 is the latest leaked one unless I missed something
11:52.11freemangordonoh, yes, 6.6
11:53.26jonwiloh and I suspect the tools would choke on ARM NEON in any case since it didn't like ARM FPU all that much (from memory)
11:55.08freemangordonhmm, that one seems closer to what we have :) https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/master/media-libs/speex/files/speex-1.2_rc1-add-neon-optimization.patch
11:57.13jonwilbtw it is sad to see that neo900 software stack work seems to have stopped (or at least is going very slowly). Unless of course there is a repo somewhere I totally missed...
11:58.05freemangordonwork needs workforce :)
11:58.09jonwiltrue
12:08.00jonwilreally wish there was someone out there with the time, interest and skills to finish reverse engineering the pulseaudio-nokia stuff... I have the interest but not the skills needed to reverse engineer ARM FPU/NEON stuff
12:12.07freemangordonjonwil: you have the skills to RE tha PA part
12:12.16freemangordonalso, Wizzup is wishing to help
12:12.35freemangordonand I'll jump on the NEON/VFP stuff
12:14.59jonwilI dont have the time to get deep into reverse engineering Maemo stuff right now
12:15.16freemangordonok
12:15.45freemangordondammit, it is impossible to diff speex, I am almost sure gcc is different :(
12:27.13jonwilSeems like pulseaudio-nokia and the cellular modem stuff are still the 2 main pieces missing
12:46.46jonwilgood to know that there are still people interested in a complete clone of pulseaudio-nokia and all the other things neo900 needs, even if the resources aren't there to make it happen right now
12:49.46jonwilbtw the reason I stopped doing anything with pulseaudio-nokia before was the difficulty in figuring out the layout of the giant private data structure inside pulseuaudio-module-nokia-voice (figuring out unknown data structures is one area my skills are very lacking)
12:56.23jonwilmy work on the cellular services daemon got stuck because the dbus calls I wanted to figure out either 1.ended up making undocumented calls to the cellular modem (meaning the only real way to figure out what those dbus calls do would be to reverse engineer the cellular modem, not an easy task) or 2.Ended up getting lost in the cellular services daemon code in ways that made figuring out what...
12:56.24jonwil...they do impossible
12:56.55freemangordonyeah, that PA userdata is a nightmare
12:58.32jonwilin fact thinking about it, it seems like every time I try to work on any of the interesting stuff (audio, cellular, status bar widgets, microb, maps, whatever else) I discover my skills aren't up to the task of reverse engineering it and stop
12:58.48jonwilpretty much why I haven't done any n900 stuff in ages
13:00.30jonwilI suspect we will never know what com.nokia.phone.sim.get_service_provider_info actually does
13:01.07freemangordonhmm. speex getting closert
13:01.31freemangordonthere are 2 more neon patches which can;t be found on the inet
13:01.51jonwilWhat is the advantage of having FOSS libspeex?
13:02.30freemangordonwe can move to a newer release
13:02.40freemangordonwe're on rc1, there is rc3
13:02.52jonwilok
13:03.02bencohand not having to redistribute a blob via cssu
13:03.03jonwilbut does newer release contain bug fixes or changes we actually need?
13:03.36jonwiloh yeah I forgot libspeex isn't installed on the phone by default
13:05.29jon_yhow does libspeex help on the n900?
13:06.05jonwilit helps if you want to use ogg speex codec for some reason
13:06.52jon_yanything to do with the telephony part?
13:07.04freemangordonno idea
13:07.09jon_yok
13:07.42jonwilspeex definatly has nothing to do with the telephony stuff on N900
13:07.52jonwilits possible there are voip systems using speex
13:08.00jonwilbut its definatly not used for any of the stock stuff
13:08.25jon_yunderstood
13:09.07freemangordonjonwil: but then, I wonder why Nokia wasted tim on optimizing it
13:09.20jonwilyou would have to ask Nokia
13:10.35jon_ymaybe they planned to use it but abandoned it
13:10.59bencohwhich cssu package depends on it?
13:11.33jonwilbut yeah if I ever get the enthusiasm to work on N900/Neo900 stuff again (and something happens that pushes me past the bits I was stuck at before such as some old Nokia people showing up with info or code) then I would contribute more
13:15.28freemangordonbencoh: noone, but there are packages in maemo-extras that depend on it
13:16.04freemangordonbut yeah, I guess it will be easier to just pu the binary blob in the repo, until we really need a newer version
13:16.13freemangordonmerlin1991: ^^^
13:35.21*** join/#maemo-ssu sparetire_ (~sparetire@unaffiliated/sparetire)
15:12.10Palispeex is/was used for google talk voice calls
15:12.14Paliand also in SIP
15:34.32freemangordonhmm, ok
15:34.49freemangordonanyway, I am on modest patches already, may return on it when done
15:55.49drathirPali: they not changed after migrate to hangouts>
15:56.01drathir>/?*
16:17.41Palino idea, if it still works on gmail side
16:58.46Sicelovoice doesn't work on n900 gtalk anymore due to hangouts. even chat has stopped working. you can't see peoplems online state
17:00.08PaliSicelo: voice & video in jabber protocol is dependent on client side
17:00.23Palidoes not depend on any server support
17:00.40Paliso if gmail still has jabber server, then you can use it
17:00.59Palialso you can do voice & video calls if you have account on any jabber server
17:01.15Siceloyes, jabber still works fine
17:01.36Sicelogtalk .. they removed jabber server afaik
17:02.06bencohnot sure whether they shut it down or just moved everyone to hangout
17:02.33Paligoogle's jabber server is still active
17:03.09Palifrom @gmail account you can still connect
17:03.15Paliand you can talk to hangout users
17:04.48Siceloi can't.
17:05.21Sicelohow do you do it? i don't see their online status for example. nor do i receive anything from them
17:06.44Siceloreenables gtalk on n900 to retest
17:16.12Siceloweird. today i see their online status .. i could swear i couldn't
17:16.33Palimaybe bug in telepathy-gabble
17:16.47Palirekilling it on N900 (or reboot) could help
17:17.02Palisometimes that telepathy process goes into bad state
17:17.29Sicelodon't know how i could test voice ... only IM is advertised on their names
17:21.50bencohdoes IM work?
17:22.22Sicelolooks as if it should. no one responding yet (but maybe they are just busy)
17:23.59freemangordonSicelo: voice and video n900<->n900 works
17:24.08freemangordontried it a couple of days ago
17:24.14Sicelovia gtalk?
17:25.12freemangordonyep
17:25.15Sicelowait .. n900 <-> n900 .. meaning the other party must use N900?
17:25.33freemangordonmeans that I tried n900<->n900
17:31.45SiceloIM works.
17:33.21Sicelothey don't see my online status though
17:36.18Sicelothey can't call... n900 doesn't ring
17:38.00Siceloalthough on their side it acts to be ringing
17:43.24PaliSicelo: other side must have jabber/jingle compatible client
17:44.32Palijingle protocol (for voice & video) depends just on client, not server
17:47.36Sicelook. sucks then, because your 'average' gtalk person is using android, and that default client doesn't help us
18:07.32freemangordonSicelo: android does not use jabber but hangouts
18:09.40Sicelothat's what i mean :)
18:11.28Siceloby the way, for some reason i can't add my 2nd gmail accoount on my 2nd N900 ... says auth error. even on PC empathy. but password is correct and account not locked
18:12.23freemangordonSicelo: open gmail on desktop and enable that device.... or stop using google services ;)
18:13.03Siceloi think i need this https://accounts.google.com/DisplayUnlockCaptcha
19:18.07freemangordonPali: modest PR1.3 patches are ported to master :), going to put it in cssu-devel
19:18.18Paliooo nice!
19:19.30freemangordonPali: som (3 iirc) patches I didn't forward-port, please check and comment http://pastebin.com/byxwNp3P
19:19.34freemangordon*some
19:42.15Palifreemangordon: can you push them to git?
19:48.43freemangordonsure, just need to fix a bug I introduced with forward-porting
19:50.47kerioforward-poettering? :P
19:51.08freemangordonkerio: hi! long time no see
19:51.17keriohiho
19:52.47*** join/#maemo-ssu NIN101 (~core@mail2.quitesimple.org)
19:55.21Palibug introduced by --forward-poettering?
20:01.26*** join/#maemo-ssu trx (ns-team@devbin/founder/trx)
20:03.22freemangordonPali: https://github.com/community-ssu/modest/commits/master
20:03.50freemangordonhmm, when is the "show more" button supposed to appear?
20:04.26Palithanks!
20:04.34Paliwhen messages are not downloaded in modest yet
20:04.43Paliadd new account into modest
20:04.59Pali(so without cached emails)
20:05.44freemangordonisn't it possible to just delete the cache for an existing account?
20:08.25Palido not know
20:08.39Paliclean setup is always better to do tests
20:10.19freemangordonPali: ok, so, what after setting up a new account?
20:10.27freemangordonI need 250 messages there iiuc
20:10.48Paliyou can add same existing account, or not?
20:11.06freemangordonyou can't, seems like username is a primary key
20:11.11Palicache should be in ~/.modest
20:11.29Palior somewhere
20:11.43Palimaybe renaming folder could help?
20:12.14freemangordonPali: hmm, are you sure this button is for not downloaded mails?
20:12.28Palido not remember
20:13.26freemangordoniiuc this is a limit to the number of headers shown in tree view
20:17.08freemangordonoh, it is me that screwed it up
20:19.43freemangordoncopy/paste error
20:21.41kerioopenssl_0.9.8zf-1+maemo1+0m5+0cssu0+thumb0 huh
20:22.57freemangordonmhm
20:24.00kerioi just yolo upgraded to what's likely in -devel, btw
20:25.39keriophew it boots
20:27.12merlin1991shakes fist at alx kernel developers
20:41.55keriowhat's the state of *bsd on n900?
22:10.06*** join/#maemo-ssu jonwil (~jonwil@27-33-80-219.tpgi.com.au)
22:12.42*** join/#maemo-ssu M4rtinK (~M4rtinK@ip-89-177-124-111.net.upcbroadband.cz)
22:14.16*** join/#maemo-ssu amiconn (~amiconn@rockbox/developer/amiconn)
22:17.07jonwilhi

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.