IRC log of #maemo-ssu for Wednesday, 2013-01-02

* DocScrutinizer05 idly wonders if awk...00:00
FIQ|n900awk binary patching?00:00
DocScrutinizer05sed is clearly too line oriented00:00
FIQ|n900super ed00:00
DocScrutinizer05FIQ|n900: *nod*00:00
FIQ|n900(I know what sed is, just not what the s part of it means)00:01
DocScrutinizer05dunno if awk could get configured to use one-byte "lines"00:01
DocScrutinizer05serial afaik00:01
DocScrutinizer05or stream00:01
DocScrutinizer05since it basically doesn't do any skip-back on input00:02
DocScrutinizer05so movin last line of sourcefile to the head of destination file is a tad cimbersome with sed00:03
DocScrutinizer05cumber*00:03
DocScrutinizer05unlike ed where you could actually do that00:03
DocScrutinizer05hmmm, ed00:04
*** macmaN has joined #maemo-ssu00:10
DocScrutinizer05I wish there was a hex_ed which doesn't work on lines but byte adresses, and not displays raw text but hex reprentation00:11
FIQ|n900make it!00:12
DocScrutinizer05already pondering it00:12
DocScrutinizer05can't be hard00:12
DocScrutinizer05read in from <file> to buffer assumes/fakes a \n after very byte, and the most basic print command converts each "char" to a 2char hex representation00:13
jon_yDocScrutinizer05: speaking of \n, here's the perl bq2k query script00:16
jon_yhttp://paste.debian.net/220910/00:16
DocScrutinizer05:-)00:16
DocScrutinizer05I'll add it to my host, if you allow00:16
jon_yhost?00:17
jon_ywell, just take it, it is your script00:18
DocScrutinizer05I'd like to add your (C) to the header however00:18
jon_yoh, sure, thanks00:19
kerioit should be 27x00 then00:24
DocScrutinizer05jon_y: what author name shall I use for your (c)?00:26
*** Sc0rpius has quit IRC00:27
*** FIQ|n900 has quit IRC00:30
*** FIQ|n900 has joined #maemo-ssu00:31
DocScrutinizer05hmm, hope that's ok: http://maemo.cloud-7.de/maemo5/usr/local/sbin/bq27k-detail-perl00:35
*** kolp has quit IRC01:10
*** Sc0rpius has joined #maemo-ssu01:15
keriofreemangordon: have you noticed a regression with whatever is responsible for phone notifications since a bit ago?01:18
DocScrutinizer05phone notifications?01:31
DocScrutinizer05could it be that rotation stuff got way more laggy for dialer in forced portrait mode, since last (2) update(s) of CSSU-T?01:32
DocScrutinizer05(dialer is the only thing in portrait mode on my IroN900, always been)01:33
kerionah, i'm talking about hildon-home taking up 100% of cpu01:33
DocScrutinizer05ouch01:33
kerioand hildon-desktop locking up when you try to access the tasknav01:33
keriopossibly because of ^01:34
DocScrutinizer05sure01:34
keriobut i couldn't even pull up the power key menu01:34
DocScrutinizer05o.O01:34
DocScrutinizer05and that's all the time?01:34
kerioi've had a slowdown and a lockup in two separate occasions01:35
kerioboth after i didn't answer a call01:35
kerio(left it ring, didn't hang up)01:35
DocScrutinizer05:nod:01:35
DocScrutinizer05I always thought that's due to recaller writing a new file and tracker going mad in indexing, when call gets finished01:36
DocScrutinizer05but since last update which for me been basically 2 in 1, it feels a lot more laggy on start of ringing, with rotating dialer to portrait mode. It takes like 5..10s now, while former lag was only ~1s01:37
DocScrutinizer05it might be a singular event, maybe caused by HAM or whatever still active in background01:39
DocScrutinizer05maybe rsync backup was active01:39
DocScrutinizer05anyway it sucked, and I hope it's not a regression in latest (2) update(s)01:40
kerioas a sidenote: is it normal that i don't get a name for a number in multiple contacts?01:40
*** Sc0rpius has quit IRC01:45
*** Sc0rpius has joined #maemo-ssu01:50
DocScrutinizer05alas yes01:52
arceanDocScrutinizer05: rotation should be slightly faster with the latest h-d01:52
DocScrutinizer05arcean: then I give it the benefit of doubt until it happens again01:53
DocScrutinizer05who knows what's going on in this little phone all the time ;-D01:53
DocScrutinizer05might actually been HAM still active01:54
DocScrutinizer05and we all know HAM brings *everything* to a grinding halt01:54
arceanyeah, sometimes strange things happens01:54
DocScrutinizer05kerio: the resolving of number->name is not very smart, actually it sucks donkeyballs01:55
DocScrutinizer05since it doesn't tolerate any small differences in number to match01:56
DocScrutinizer05and also fails completely when same number is in multiple contacts01:56
kerioi kinda like the fact that if a number is associated to multiple contacts, it doesn't just choose one at random01:56
DocScrutinizer05I'd prefer some unambiguous priority thing, like maybe wildcards in numbers or sth01:57
DocScrutinizer05or a dedicated numerical priority field01:58
DocScrutinizer05or a simple selector popping up when same number gets added to a second (third...) contact, to select which one is primary contact now01:59
kerioforeign keys without enforcing will end up in a giant mess01:59
kerio*enforcement01:59
DocScrutinizer05after all if number->name works during call inbound, it can also get used during adding a number to arbitrary contacts, to detect twin entries02:00
DocScrutinizer05obviously that again is doomed to fail due to missing sourcecode of contacts app02:00
DocScrutinizer05much the same basic issue like with fixed hardcoded set of default fields shown with new contacts. Each time I have to add the same friggin "address" fields etc again when creating a new contact, I feel like smashing the N900 on elop's head02:02
DocScrutinizer05(though elop is innocent this single time)02:03
DocScrutinizer05also did you ever notice that the only way to read a freetext note added to a contact is to open the contact for editing?02:05
*** freemangordon has quit IRC02:06
keriothe os x addressbook lets you choose which fields to display in editing mode, even without any data02:07
*** FIQ|n900 has quit IRC02:25
DocScrutinizer05alas this isn't os x02:34
DocScrutinizer05err02:34
DocScrutinizer05well, I should actually look into this *%$"G" contacts database, maybe the default empty record is in there and could get edited02:36
DocScrutinizer05GOD! would I like to change the default set if fields shown in a new contact02:37
DocScrutinizer05s/ if / of /02:37
infobotDocScrutinizer05 meant: GOD! would I like to change the default set of fields shown in a new contact02:37
DocScrutinizer05actually I couldn't think of any special complications in REing the contacts frontend02:38
DocScrutinizer05should probably be easier than even calendar02:39
DocScrutinizer05and a completely different class than dialer02:40
DocScrutinizer05probably I should check evolution02:43
*** NIN101 has quit IRC02:55
RaimuIPS is a horrible ROM patching dinosaur03:30
RaimuNo way to discern if your ROM dump is the one the patch is created for03:30
RaimuIt just tells what bytes of the file to change into which and then good luck running the file03:31
RaimuIf there's like a different header than expected taking up space then IPS here patches the wrong bytes03:31
RaimuOr maybe the guys making the patch used a slightly older or newer dump of the ROM and hunting for that one you go03:32
*** arcean has quit IRC03:42
*** jonwil has joined #maemo-ssu03:53
jonwil~seen pali03:53
infobotpali <~pali@Maemo/community/contributor/Pali> was last seen on IRC in channel #maemo-ssu, 4d 14h 33m 43s ago, saying: 'freemangordon, do you know state of portrait keyboard?'.03:53
*** ekze has quit IRC04:01
*** ekze has joined #maemo-ssu04:18
*** jonwil has quit IRC04:38
*** jonwil_ has joined #maemo-ssu04:38
*** jonwil_ is now known as jonwil04:38
gregoaDocScrutinizer05, keri: sketch for getting the first parapgraph of a wikipedia page as plain text for infobot: http://paste.debian.net/220949/ ; requires the not very standard Web::Scraper module04:48
gregoas/keri/kerio/04:48
infobotgregoa meant: DocScrutinizer05, kerio: sketch for getting the first parapgraph of a wikipedia page as plain text for infobot: http://paste.debian.net/220949/ ; requires the not very standard Web::Scraper module04:48
*** nox- has quit IRC05:18
*** M4rtinK has quit IRC05:23
*** amiconn_ has joined #maemo-ssu05:50
*** amiconn has quit IRC05:50
*** amiconn_ is now known as amiconn05:50
*** Martix has quit IRC05:53
*** DocScrutinizer05 has quit IRC06:04
*** DocScrutinizer05 has joined #maemo-ssu06:04
*** tg has quit IRC06:14
*** tadzik has quit IRC06:41
*** tg has joined #maemo-ssu06:52
*** tadzik has joined #maemo-ssu06:54
*** sourav has joined #maemo-ssu07:03
*** sourav has left #maemo-ssu07:05
*** luf has joined #maemo-ssu07:39
*** kolp has joined #maemo-ssu07:49
*** FlameReaper has joined #maemo-ssu09:19
*** FlameReaper has quit IRC09:33
*** FlameReaper has joined #maemo-ssu09:38
kerio:D09:59
kerionow i feel kinda bad for dissing perl09:59
keriobut not too much09:59
merlin1991luf: libxml2 is merged10:05
lufmerlin1991: I see in mailbox. Thanks. I'll put it into cssu-devel in near future also new libpng and hopefully zlib.10:06
*** dhbiker has quit IRC10:09
*** dhbiker has joined #maemo-ssu10:10
merlin1991luf:  you also might like this commit https://gitorious.org/community-ssu/libcpaboutcssu/commit/cbd845fbc731bff84e8d5423e4fa5bca4121327f ;)10:13
lufOh no ... I'm not the grey man behind ... Now they know me :D10:14
*** dhbiker has quit IRC10:27
*** Milhouse has quit IRC10:29
*** Milhouse has joined #maemo-ssu10:30
*** kolp has quit IRC10:36
*** drath_n900 has joined #maemo-ssu10:43
*** drath_n900 has quit IRC10:48
*** drath_n900 has joined #maemo-ssu10:55
*** dhbiker has joined #maemo-ssu11:01
*** drath_n900 has quit IRC11:04
jon_yDocScrutinizer05: jon_y is good11:22
keriomy, how modest11:31
jon_ykerio: DocScrutinizer05 put DocScrutinizer, so my irc nick is fine too :)11:34
kerioi meant the absolute statement11:35
kerioinfobot: jon_y is <jon_y> jon_y is good11:35
infobotokay, kerio11:35
kerio~jon_y11:35
infobotfrom memory, jon_y is <jon_y> jon_y is good11:35
kerio:D11:35
jon_y:)11:35
jon_yI was referring to the bq2k query perl conversion11:36
*** dhbiker has quit IRC11:46
*** _rd has joined #maemo-ssu12:02
keriojon_y: i know :)12:05
kerio'twas a joke12:05
*** FlameReaper has quit IRC12:06
*** arcean has joined #maemo-ssu12:28
*** NIN101 has joined #maemo-ssu12:58
*** andre__ has joined #maemo-ssu13:03
luffreemangordon: to the /dev/mtd2 - the first line is interesting: [ 1118.429321] Unable to handle kernel NULL pointer dereference at virtual address 0000000013:22
keriodereferencing NULL, yay13:24
lufkerio: exactly ...13:31
*** jonwil has quit IRC13:46
*** _rd has quit IRC13:52
*** arcean has quit IRC13:54
*** arcean has joined #maemo-ssu13:55
*** dhbiker has joined #maemo-ssu13:55
*** lizardo has joined #maemo-ssu13:56
*** arcean_ has joined #maemo-ssu13:58
luf~seen pali13:59
infobotpali <~pali@Maemo/community/contributor/Pali> was last seen on IRC in channel #maemo-ssu, 5d 39m 52s ago, saying: 'freemangordon, do you know state of portrait keyboard?'.14:00
kerioD:14:00
*** arcean has quit IRC14:01
*** arcean has joined #maemo-ssu14:01
*** arcean_ has quit IRC14:03
lufkerio: it's KP52 ;) but maybe not the latest ...14:04
*** drath_n900 has joined #maemo-ssu14:07
*** arcean_ has joined #maemo-ssu14:11
*** arcean has quit IRC14:14
*** luf has quit IRC14:17
*** luf has joined #maemo-ssu14:18
*** M4rtinK has joined #maemo-ssu14:19
*** arcean_ is now known as arcean14:19
*** luf has quit IRC14:22
*** luf has joined #maemo-ssu14:23
*** Martix has joined #maemo-ssu14:29
*** _rd has joined #maemo-ssu14:36
kerioluf: yeah, prereleases don't have an actual version number14:40
lufkerio: are you sure? [  464.625366] KERNEL_VERSION=Linux Nokia-N900 2.6.28.10-power52 #1 PREEMPT Fri Oct 12 16:06:34 CEST 2012 armv7l GNU/Linux14:41
kerioluf: there are at least three KP5214:41
*** M4rtinK has quit IRC14:48
*** freemangordon has joined #maemo-ssu14:51
freemangordonluf: could you pastebin /dev/mtd2 so I can take a look at it?14:52
luffreemangordon: sure14:53
freemangordonluf: BTW there is zlib repo on gitorious ;)14:55
luffreemangordon: I know.14:59
*** drath_n900 has quit IRC15:00
luffreemangordon: you don't need to wait for sources in git as I sent you the link to test the patch ...15:00
*** _rd has quit IRC15:05
*** drath_n900 has joined #maemo-ssu15:06
freemangordonluf: I'll wait for the source in git, I want to do thumb branch too15:07
*** jonwil has joined #maemo-ssu15:08
lufIt will be better if you can test it without git. Maybe we don't want the patch in git ...15:08
luffreemangordon: ^^^15:09
freemangordonhmm, yeah, makes sense15:11
luffreemangordon: http://pastebin.com/F5cEt22815:11
luf3 times the same oops ;)15:12
luffreemangordon: however for sure I installed this kp52 some time prior the official release.15:13
freemangordonwhat you were doing?15:13
lufNothing special.15:13
freemangordonaah, make sure to upgrade15:13
freemangordonThere were some fixes15:13
lufOnce some dpkg -i (through sshd)15:13
freemangordoncharger connected?15:14
lufnetwork over usb.15:14
lufSo USB cable connected (N900 and PC)15:14
*** _rd has joined #maemo-ssu15:14
freemangordonhmm, QtMobileHotspot?15:14
freemangordonhowever, install the latest KP52 and if the problem continues, report it again (to me and Pali)15:15
lufI don't think so. I'm going to check. Is it package name?15:15
freemangordonqtmobilehotspot iirc15:15
lufIt hasn't appeared again.15:15
lufJust 3 times in one day.15:16
lufBefore that the phone was fully discharged.15:16
freemangordondespite that, install the latest, Pali has fixed sume related stuff15:16
freemangordon*some15:16
lufThe first oops during charging. The last one was with nearly full battery.15:17
freemangordon^^^15:17
lufI see.15:17
luffor record: no QtMobileHotspot (checked with dpkg -l | grep hot)15:18
freemangordonluf: how do you setup USB netowrking?15:19
kerioluf: oh really? you installed kp52 before the official release? i would've never guessed15:20
kerioexcept for the fact that it still hasn't been released :)15:21
lufkerio: because of BT + NAP and bridge module ;)15:24
*** M4rtinK has joined #maemo-ssu15:28
jonwilAnyone here seen Pali lately or know if/when he is most likely to show up next?15:30
*** drath_n900 has quit IRC15:33
lufjonwil: It's just 5 days. Maybe some vacation as no regular school hours now ...15:33
jonwilquite possible :)15:34
*** arcean has quit IRC15:42
*** arcean has joined #maemo-ssu15:44
freemangordonluf: I am not sure NEON patch in zlib is used at all15:47
freemangordonat least there is no adler32.o in arm/ directory15:47
freemangordonaah, wai, it is #included15:48
freemangordon*wait15:48
amiconnHmm, somehow I can't install cssu-tjumb :\15:54
amiconnI tried the 1st step, but now HAM has no catalogs at all?15:54
kerioamiconn: HAM is being silly as usual15:54
kerioclose it, wait for the cpu usage to go down, start it again15:54
luffreemangordon: I hade the same feeling prior to include :D15:55
luffreemangordon: but there is no problem with test and qemu/sb should have problem with neon ... So I'm not sure if there is any neon opts.15:56
*** luf has quit IRC16:02
*** arcean_ has joined #maemo-ssu16:17
*** FIQ|n900 has joined #maemo-ssu16:18
*** arcean_ has quit IRC16:20
*** arcean has quit IRC16:20
*** arcean has joined #maemo-ssu16:21
amiconnkerio: Seems to have worked now, thanks!16:22
*** arcean has quit IRC16:22
*** arcean has joined #maemo-ssu16:24
amiconnThumbified fennec doesn't seem to be any faster than plain fennec though (iow, still unusable)16:27
kerio:(16:28
amiconnIt's not a real problem, since I'm using opera mobile 95% of the time anyway (and maybe 5% microb - due to needing flash or ovi compatibility)16:32
kerioweird, i can't stand opera16:33
amiconnIt's the fastest browser on N900 by far16:34
keriomicrob is fast enough16:34
kerioand i doubt that opera mobile can actually deal with weird websited16:34
kerio*sites16:35
amiconnMicrob is considerably slower, but still okay16:35
amiconnOpera mobile has page reflow, which is really useful16:35
amiconnBtw, does cssu-thumb have thumbified microb?16:37
keriothe rendering engine, not the UI16:37
keriothe UI is closed16:37
*** FIQ|n900 has quit IRC16:51
*** _rd has quit IRC16:54
*** _rd has joined #maemo-ssu16:59
*** _rd has quit IRC17:09
*** _rd has joined #maemo-ssu17:15
*** peetah has quit IRC17:17
*** peetah has joined #maemo-ssu17:18
*** luf has joined #maemo-ssu17:32
*** jonwil has quit IRC17:33
freemangordonkerio: and it is the engine that matters in terms of speed usually17:38
luffreemangordon: have you noticed any speedup with the zlib and the patch from meego? :)17:38
freemangordonno :(. well, some 2-3 %17:39
freemangordongcc does it better with -O317:39
freemangordonluf: BTW -Os is size17:39
freemangordon-O2 is better17:39
freemangordonluf: that NEON patch does not worth the bytes used to be stored on the device :D:D:D17:40
freemangordonluf: BTW did you see the comments from the author, something like "I don;t have ARM device nearby, so I didn;t test my code" :D17:41
freemangordonluf: we'd rather use gcc 4.7.2 with -ftree-vectorize -ffast-math to produce assmby output and use that17:42
freemangordon*assembly17:42
freemangordonluf: did you try code aurora patch?17:43
luffreemangordon: so change -O2 flag and don't use the patch. To be honest I don't study this patch very deeply as I get it from meego repository.17:44
freemangordonluf: I still think NEON could be useful, if we only can find a good code17:45
luffreemangordon: Yes I tested code aurora patch. It's in history (it's slower), I hope we both are talking about the fast patch.17:45
freemangordonI am using the source in your home dir17:45
luffreemangordon: np about NEON.17:45
luffreemangordon: You can read it in changelog what I tested and doesn't work :)17:46
lufhttps://gitorious.org/0xdroid/external_zlib/commit/5d42d250693f443ba3ad0f663d269177b2d02300/diffs - this codea aurora patch just slow down the deflate.17:46
freemangordonluf: BTW i'll google around if I can find something better17:46
freemangordonwell, adler is used only in decompress aiui17:47
luffreemangordon: ok.17:47
luffreemangordon: so the question is if the fixes in zlib 1.2.7 are enough to upgrade zlib on N900.17:48
freemangordonluf: wait, that patch is only partial17:48
freemangordonthere is code aurra adler neon implementation17:48
freemangordonluf: would you rephrase please17:49
freemangordonthe question that is17:49
luffreemangordon: I see no adler neon patch from code aurora.17:59
luffreemangordon: rephase: do we want upgrade zlib on N900 to 1.2.7 (without neon patch = without speed up)? There are just some fixes (mem_leaks - no proove if againist N900 1.2.3-9 code) and some new features.18:01
luffreemangordon: please post me the code aurora adler neon implementation. I can try it.18:01
freemangordonluf: ok18:16
*** peetah has quit IRC18:16
*** Martix has quit IRC18:16
*** peetah has joined #maemo-ssu18:17
freemangordonluf: https://codereview.chromium.org/5176006/patch/1/218:17
freemangordonhttps://codereview.chromium.org/517600618:17
freemangordonluf: BTW this is against 1.2.3 :)18:20
*** BCMM has joined #maemo-ssu18:34
*** BCMM has quit IRC18:47
luffreemangordon: is it really different patch? When I took a look last time I think it's just subset of patches I tried.18:59
lufI'll try.19:00
freemangordonluf: don;t19:00
freemangordonI am porting it, will be ready in a couple of minutes19:00
luffreemangordon: I bet it'll help in that case :D19:01
freemangordonthe link you posted contains only fast_copy part19:01
freemangordondon't :P19:01
freemangordonyou'll lose your money19:01
lufI know that. I spent several hours or days on libpng with slowdown all the time and you tried one patch in few minutes and it saves 30%. It's the fact ;)19:02
freemangordonwell, I guess I am just lucky :D19:03
lufBTW to be honest I vote for zlib 1.2.7 as it's backword compatible (in public API/ABI). Unfortunatlly libxml2 used internal stuctures (but we have it patched now).19:03
freemangordonBTW I vote for 1.2.7 too19:04
luffreemangordon: don't belieave you're lucky. You're so good.19:04
*** toxaris has joined #maemo-ssu19:21
*** kolp has joined #maemo-ssu19:27
*** qwazix has left #maemo-ssu19:43
*** kolp has quit IRC19:45
*** luf has quit IRC19:55
*** M13 has joined #maemo-ssu20:06
*** Martix has joined #maemo-ssu20:15
*** _rd has quit IRC20:36
*** _rd has joined #maemo-ssu20:37
*** _rd has quit IRC21:10
*** _rd has joined #maemo-ssu21:11
*** _rd has quit IRC21:30
*** nox- has joined #maemo-ssu21:50
*** _rd has joined #maemo-ssu22:05
*** AndrewX192_ has joined #maemo-ssu22:25
*** AndrewX192_ has joined #maemo-ssu22:25
*** AndrewX192 has quit IRC22:28
*** int_ua has joined #maemo-ssu22:29
*** M13 has quit IRC22:34
*** jester` has joined #maemo-ssu22:34
*** wumpwoast has joined #maemo-ssu22:36
*** toxarisswe has joined #maemo-ssu22:54
*** AndrewX192_ has quit IRC22:54
*** AndrewX192 has joined #maemo-ssu22:55
*** AndrewX192 has joined #maemo-ssu22:55
*** toxaris has quit IRC22:57
*** AndrewX192 has quit IRC22:59
*** lizardo has quit IRC22:59
*** AndrewX192 has joined #maemo-ssu23:00
*** AndrewX192 has joined #maemo-ssu23:00
*** _rd has quit IRC23:26
*** _rd has joined #maemo-ssu23:35
*** toxarisswe has quit IRC23:35
*** _rd has quit IRC23:44
*** arcean has quit IRC23:53
*** arcean has joined #maemo-ssu23:56

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!