**** BEGIN LOGGING AT Fri Oct 27 03:00:02 2017
Oct 27 05:55:13 <Maxdamantus> Weird. Seem to have missed some SMSes that were sent to me.
Oct 27 05:55:55 <Maxdamantus> Can see them being received and acknowledged at the right time in syslog, but never saw them through the UI.
Oct 27 07:00:43 <Vajb> Maxdamantus: they probably contained smileys
Oct 27 07:00:53 <Maxdamantus> They did.
Oct 27 07:01:05 <Vajb> there was a fix in tmo
Oct 27 07:01:50 <Vajb> also i've noticed that some new phones send sms in wrong format
Oct 27 07:02:12 <KotCzarny> isnt it priority issue for cssu?
Oct 27 07:02:14 <Vajb> so sometimes they r treated as mms
Oct 27 07:02:36 <KotCzarny> vajb, that's because mms is often sent for free
Oct 27 07:02:36 <Maxdamantus> It was an iPhone that sent them.
Oct 27 07:02:52 <KotCzarny> while sms is most often overpriced
Oct 27 07:02:57 <Maxdamantus> So what's special about the smileys in iOS?
Oct 27 07:03:01 <KotCzarny> maxd: utf
Oct 27 07:03:04 <KotCzarny> most likely
Oct 27 07:03:09 <Maxdamantus> I'd've imagined they'd've just been Unicode, yeah.
Oct 27 07:03:19 <KotCzarny> utf encoded in unicode?
Oct 27 07:03:20 <KotCzarny> ;)
Oct 27 07:03:44 <Maxdamantus> Unicode encoded in whatever SMSes typically use to encode Unicode, which I suspect is UTF-8.
Oct 27 07:04:32 <KotCzarny> sms uses plain ascii when doesnt contain any special chars (and then you are charged per ~160 chars), when it switches to utf, you pay twice as much
Oct 27 07:04:46 <Maxdamantus> .. so it uses UTF-16?
Oct 27 07:04:52 <KotCzarny> yup
Oct 27 07:05:06 <Maxdamantus> Who came up with that idea?
Oct 27 07:05:10 <KotCzarny> dont ask me
Oct 27 07:05:29 <KotCzarny> but it was big money cow for telcos
Oct 27 07:05:32 <Maxdamantus> Surely they must've only started doing that long after UTF-8 was a thing.
Oct 27 07:05:58 <KotCzarny> as people were trying to add language specific characters and weren't expecting
Oct 27 07:06:06 <KotCzarny> nope
Oct 27 07:06:24 <KotCzarny> it was when people started demanding localized phones
Oct 27 07:06:25 <KotCzarny> ;)
Oct 27 07:06:38 <KotCzarny> even old nokia 3310 has it
Oct 27 07:06:48 <KotCzarny> and i suspect even older
Oct 27 07:07:00 <Maxdamantus> > The first commercially sold SMS service was offered to consumers, as a person-to-person text messaging service by Radiolinja (now part of Elisa) in Finland in 1993.
Oct 27 07:07:26 <Maxdamantus> which is about the same time as UTF-8 was invented.
Oct 27 07:08:21 <Maxdamantus> 3310 was apparently made in 2000, again, long after UTF-8.
Oct 27 07:11:38 <Maxdamantus> I can imagine devices just having options to set some 8859-like encoding before there being an effort to use Unicode instead .. find it hard to imagine how everyone would switch to Unicode but end up using UTF-16.
Oct 27 07:11:47 <Maxdamantus> unless it really was at the very beginning of commercial SMS.
Oct 27 07:17:06 <Maxdamantus> https://en.wikipedia.org/wiki/SMS#cite_ref-3GPP_23.038_44-0 suggests that it's actually UCS-2, which would explain it.
Oct 27 07:17:09 <Maxdamantus> Fucking idiots.
Oct 27 07:17:20 <KotCzarny> money money money
Oct 27 07:18:21 <CatButts> it's so funny
Oct 27 07:19:07 <Maxdamantus> Well, UCS-2 doesn't give a monetary advantage over UTF-16.
Oct 27 07:19:54 <Maxdamantus> (I'm guessing you're referring to the fact that it more often has to send things as multiple messages because almost half of the messages are just zeroes)
Oct 27 07:20:30 <Maxdamantus> UCS-2 is just completely the wrong thing to use.
Oct 27 07:20:35 <KotCzarny> yup
Oct 27 07:21:03 <KotCzarny> even single localized character switches whole message into ucs-2
Oct 27 07:21:45 <KotCzarny> and since most people have some internet or mms package on their phones already, it's cheaper than sending sms
Oct 27 07:21:55 <KotCzarny> apparently.
Oct 27 07:22:01 <KotCzarny> which is just wrong. eh
Oct 27 07:25:41 <Maxdamantus> Mm .. if I try to send a message to myself that includes U+10000 in it, it says it fails to deliver, but syslog just gives stupid dbus errors.
Oct 27 07:26:21 <Maxdamantus> Oct 27 20:24:35 orcus user.crit rtcom-messaging-ui[26921]: GLIB CRITICAL ** default - dbus_g_error_get_name: assertion `error->code == DBUS_GERROR_REMOTE_EXCEPTION' failed
Oct 27 07:27:04 <Maxdamantus> Don't tell me dbus is restricted to UCS-2 too.
Oct 27 07:27:28 <KotCzarny> yaaaay
Oct 27 07:38:07 <Maxdamantus> Can't access the tars linked to from http://talk.maemo.org/showthread.php?t=93427&page=2 , but it sounds like it's just a matter of replacing a "UCS-2" string passed to iconv with "UTF-16"
Oct 27 07:43:09 <Maxdamantus> Yep, that seemed to do it.
Oct 27 07:43:45 <Maxdamantus> Hopefully that zero byte that was after "UCS-2\0" wasn't important.
Oct 27 07:45:44 <Maxdamantus> Will try it properly with an iPhone later, but can at least send U+10000 to myself now.
Oct 27 07:47:47 <Maxdamantus> (for the record, I just used a hex editor to replace "UCS-2\0\0" with "UTF-16\0" in /usr/lib/libsms-utils.so.0.0.0)
Oct 27 07:48:37 <Maxdamantus> I'm guessing "UTF16\0" would probably also work if it's just passed to iconv, because I think iconv is fairly lenient about those names.
Oct 27 07:48:51 <Maxdamantus> (since "UTF16\0" and "UCS-2\0" have the same length)
Oct 27 07:56:09 <Maxdamantus> Ah, I guess this is the thing linked to by that thread: https://github.com/harmattan/emojifix
Oct 27 07:57:13 <Maxdamantus> Seems to effectively do the same thing, but it does it by linking to a different library instead of just changing the string directly.
Oct 27 08:44:57 <sixwheeledbeast> ~emojifix
Oct 27 08:45:07 <sixwheeledbeast> hmm.
Oct 27 08:45:30 <sixwheeledbeast> there was a thread on it somewhere Thomas Perl
Oct 27 08:45:59 <sixwheeledbeast> Ah yesy that github link will do
Oct 27 08:47:09 <ShadowJK> More importantly, unicode wasn't anything invented by phone companies. ;-)
Oct 27 08:47:33 <DocScrutinizer05> unicode in SMS was always in the spec afaik
Oct 27 08:48:02 <DocScrutinizer05> hmm, not sure about *unicode*
Oct 27 08:48:29 <DocScrutinizer05> might be other specific codepages
Oct 27 08:49:16 <DocScrutinizer05> but the switch of whole SMS from one-byte to two-byte chars been there from beginning I'd say
Oct 27 08:49:26 <ShadowJK> ah okay
Oct 27 08:50:27 <ShadowJK> Didn't windows use UCS-2 before it became clear that unicode wouldn't stay at 16 bit?
Oct 27 08:50:48 <Maxdamantus> No.
Oct 27 08:50:53 <Maxdamantus> Because Windows didn't exist back then.
Oct 27 08:50:58 <DocScrutinizer05> I can't recall details. UCS rings a bell
Oct 27 08:51:07 <DocScrutinizer05> ~phonecontrol
Oct 27 08:51:07 <infobot> it has been said that phonecontrol is http://wiki.maemo.org/Phone_control
Oct 27 08:52:33 <Maxdamantus> actually, nvm, it did. I thought UTF-16 was older than it was.
Oct 27 08:52:57 <Maxdamantus> Hm, so UTF-16 was made in 1996 .. that's .. after UTF-8.
Oct 27 08:53:09 <ShadowJK> I remember someone saying that windows is UCS-2 .. but then switched to UTF-16 at some point
Oct 27 08:53:17 <ShadowJK> Maxdamantus, yep
Oct 27 08:53:43 <DocScrutinizer05> sth along >> ADDR_TYPE = 129 #unknown format<<
Oct 27 08:55:12 <DocScrutinizer05> https://www.startpage.com/do/search?q=SMS+encoding
Oct 27 08:55:32 <Maxdamantus> I didn't get something that suggested an unknown format.
Oct 27 08:55:56 <DocScrutinizer05> https://en.wikipedia.org/wiki/GSM_03.38
Oct 27 08:56:01 <Maxdamantus> Oct 27 16:51:05 orcus user.notice telepathy-ring[1204]: GLIB MESSAGE Modem-SMS - deserializing SMS-DELIVER "..." failed: conversion failed (-10@SmsUtilssError)
Oct 27 08:56:41 <Maxdamantus> Oct 27 16:51:05 orcus user.warn sms-manager[857]: GLIB WARNING ** default - Error decoding TPDUs. Discarding message
Oct 27 08:57:50 <DocScrutinizer05> >>However, since modern programming environments do not provide encoders or decoders for UCS-2, some cell phones (e.g. iPhones) use UTF-16 instead of UCS-2<<
Oct 27 08:59:01 <DocScrutinizer05> >>his works, because for characters in the Basic Multilingual Plane (including full alphabets of most human languages) UCS-2 and UTF-16 encodings are identical. To encode characters outside of the BMP, such as emoticons, UTF-16 uses surrogate pairs, which when decoded with UCS-2 would appear as two valid but unmapped code points.<<
Oct 27 08:59:02 <Maxdamantus> That sounds a bit misleading.
Oct 27 08:59:26 <ShadowJK> "n a standard GSM text message, all characters are encoded using 7-bit code units, packed together to fill all bits of octets" <- Ah this is the headache I remember
Oct 27 09:00:03 <Maxdamantus> UCS-2 isn't really designed to be encoded/decoded.
Oct 27 09:00:20 <DocScrutinizer05> ShadowJK: see http://wiki.maemo.org/Phone_control#Send_SMS for a fine (de)octettifier
Oct 27 09:00:21 <Maxdamantus> Because what would you encode/decode it to?
Oct 27 09:01:00 <Maxdamantus> You can convert between UCS-2 and UTF-8, or between UCS-2 and UTF-16, but some things in either direction simply can't be converted.
Oct 27 09:01:17 <DocScrutinizer05> def semi_octify(str):
Oct 27 09:01:24 <DocScrutinizer05> and 
Oct 27 09:01:25 <DocScrutinizer05> def deoctify(arr):
Oct 27 09:01:54 <Maxdamantus> whereas UTF-8 and UTF-16 are real encodings, since they are both able to losslessly convert between each other and conceptually into "Unicode" code points.
Oct 27 09:02:03 <Maxdamantus> (as long as the UTF-8/UTF-16 strings are well-formed)
Oct 27 09:02:36 <DocScrutinizer05> oops  def octify(str):
Oct 27 09:06:07 <DocScrutinizer05> also see my slightly patched version: http://maemo.cloud-7.de/maemo5/usr/local/bin/smscb.py
Oct 27 09:13:05 <Maxdamantus> So you can't octify one-character strings?
Oct 27 09:13:12 <Maxdamantus> UnboundLocalError: local variable 'bitstocopy' referenced before assignment
Oct 27 09:17:48 <Maxdamantus> This octify function in that Phone_control link looks like nonsense to me.
Oct 27 09:19:34 <Maxdamantus> Seems to work as the inverse of deoctify at least, as long as the string is not of length 1.
Oct 27 09:21:18 <Maxdamantus> or not.
Oct 27 09:22:46 <Maxdamantus> >>> oct.deoctify(oct.octify("a"*7))
Oct 27 09:22:47 <Maxdamantus> 'aaaaaaaB'
Oct 27 09:28:07 <Maxdamantus> and yeah, the SMS from the iPhone works now.
Oct 27 09:41:42 <Langoor> I have my n900 on the newest cssu now (21.2011.38-1Tmaemo11) (testing branch)
Oct 27 09:42:02 <Langoor> kernel-power
Oct 27 09:42:16 <Langoor> now to uboot, I couldn find a lot about that, how would I install this?
Oct 27 11:19:01 <DocScrutinizer05> decompression 7->8: http://paste.opensuse.org/25469617
Oct 27 11:20:09 <DocScrutinizer05> oops sorry
Oct 27 17:43:48 * CatButts wonders if running Discord on the Maemo is even feasible
Oct 27 17:54:34 <pkill9> i'd like that
Oct 27 17:55:45 <pkill9> i would think the hardware could handle it
Oct 27 19:10:48 <CatButts> Discord is a piece of doo-doo, anyway
Oct 27 19:11:16 <CatButts> but hey, if it's what your friends use, that's life
Oct 27 19:11:38 <pkill9> yeah 
Oct 27 19:11:53 <CatButts> oh, this reminds me
Oct 27 19:12:03 <CatButts> does Skype for N900 still work?
Oct 27 19:24:25 <sixwheeledbeast> Discord refer you to the web app for linux
Oct 27 19:24:51 <sixwheeledbeast> looks like there is a beta for 64 bit only
Oct 27 19:26:06 <CatButts> 64butts
Oct 27 19:26:49 <CatButts> https://github.com/EionRobb/purple-discord
Oct 27 19:27:05 <CatButts> this might be a starting point for such an endeavour
**** ENDING LOGGING AT Sat Oct 28 03:00:02 2017