DocScrutinizer05 changed the topic of #neo900 to: http://neo900.org | conversations are logged to http://infobot.rikers.org/%23neo900/ and http://irclog.whitequark.org/neo900 | 2013-11-04 - the day our fundraiser reached its goal | 2014-05-01 360 devices 75k€| 0712 183 ~30k | 0810 300 ~49k | 0914 346 ~56k
* DocScrutinizer05 just updates his Jollaphone and notices how sailfish is a closed-surface wholesale product not allowing any access to the system
<Wizzup> indeed
b1101 has quit [Quit: b1101]
jonwil has quit [Remote host closed the connection]
Pali has quit [Remote host closed the connection]
<DocScrutinizer05> ohwow, fingerterminal
<DocScrutinizer05> no redial of last number, incredible. As if the last 30 years didn't exist
che1 has quit [Remote host closed the connection]
che1 has joined #neo900
che1 has quit [Remote host closed the connection]
<DocScrutinizer05> how hard could it be to have the "dial2 button do a redial when not a single digit been enetered in dialer, instead of disabling it when no digit
<DocScrutinizer05> also size of digits entered + displayed in dialer is pathetic
<wpwrak> couldn't that result in a lot of unintentional calls ?
<DocScrutinizer05> with like 5 actions/gestures to even *open* the dialer? hardly
<DocScrutinizer05> you think pressing a few random digits in your pocket is a sufficient protection against unintended dialing by randomly pressing the "green button"?
<DocScrutinizer05> making that "call" button a slider instead would make much sense, yes
<DocScrutinizer05> requiring entering of a number to fial instead of simply redialing last number when no number entered before engaging "dial" button, not so much
<DocScrutinizer05> dial*
<wpwrak> dialling random digits usually isn't such a big issue. those few places that have three-digit numbers probably aren't bothered much by such calls :)
<DocScrutinizer05> eventually it seems i have to knock together a decent dialer UI
<wpwrak> but dialling a perfectly valid regular number can be undesirable
nicksydney has quit [Quit: No Ping reply in 180 seconds.]
nicksydney has joined #neo900
<DocScrutinizer05> note that in sailfish (just like in fremantle) you can reach dialer UI only by hitting the "enter a number" button in inbound call history
<wpwrak> e.g., call your wife to tell her that you have to work late, then mess with your girlfriend and in the throes of passion accidently redial ...
<DocScrutinizer05> or generally call history. Which conveniently may lack the calls that didn't even get established
<DocScrutinizer05> wpwrak: that's all no argument for having a nonsensical UI paradigm
<wpwrak> well, if the path to the accident is always large, then it's probably okay to have a "quick redial"
<DocScrutinizer05> relying on "how likely is it to dial a valid number" is a pretty poor approach to avoid unintentional calls
<wpwrak> or maybe have two steps: dial -> last number appears, dial again -> call is made
<DocScrutinizer05> that's actually the correct way, juust that last 5 or 10 calls appear and the most recent one gets dialed when you hit "dial", unless you select another one
<DocScrutinizer05> for good reference, compare to ancient Nikia phones (pre smartphone age) that indeed had a decent UI
<DocScrutinizer05> Nokia*
<DocScrutinizer05> you had basically red, green, up, down, NK-pad, and a power button. Maybe two softkeys
<DocScrutinizer05> the "modern" touch" UIs fail due to too many alternative or differing input methods
<DocScrutinizer05> on a phone I'd expect dialer to be one of my primary homescreens
<DocScrutinizer05> with a hardware button to open it up, at any time
<DocScrutinizer05> dialing my favorite 10 numbers must be a matter of no more than 2 actions
<DocScrutinizer05> dialing an arbitrary "spelled out" number as well
<DocScrutinizer05> (until strart of keying in the 10some digits of number)
<DocScrutinizer05> and why the heck are the digit chars a 5mm high on a screen with 5" diagonale?
<DocScrutinizer05> even 10$ feature phones know to zoom the digit display to always use max area of the 180x55 display
<DocScrutinizer05> I've seen only one decent "smartphone dialer" so far, and that's been a windows phone. where entering digits into dialer also fulltext-searched and displayed the contacts for anything matching the T9 equivalent of the number you entered
<DocScrutinizer05> contacts and call history actually
<DocScrutinizer05> protection against unintended operation of device is a duty of keylock
<DocScrutinizer05> long press of any digit key shall dial the speed-dial number on storage #1..#10
<DocScrutinizer05> entering #00 to #99 shal invoke (but not yet dial) the storage (aka contacts) of that index number
<DocScrutinizer05> evidently only applicable on a list sorted by index, not on a list sorted by alphabet or whatever
<DocScrutinizer05> ooops, that's 00# to 99# iirc
<DocScrutinizer05> even in line with 3GPP
<DocScrutinizer05> I think our modem actually supports that (except for already dialing that number instead of just displaying it)
<DocScrutinizer05> as a general rule: always design an UI so you could operate it by (hot-)keys and never need to touch mouse or whatever pointer device
<DocScrutinizer05> in case of a phone, your number of hardware (hot)keys is limited to 16..20
<DocScrutinizer05> and usually no qualifier keys allowed
<DocScrutinizer05> dualtouch pinch-to-zoom is a regression in this regard
<DocScrutinizer05> try to pinch-to-zoom when you got only one free hand to hold and touch the phone
<DocScrutinizer05> wobbly backside that doesn't allow device to sit flat and stable on top of table: nogo
<DocScrutinizer05> (jollaphone is OK here)
<Oksana> What about 3.5mm connector? It's currently "????" on the block diagram?
<DocScrutinizer05> (N900 alas not, unless you attach feet: http://maemo.cloud-7.de/share-service/20141124_002.jpg )
<DocScrutinizer05> Oksana: the AVconn is a Nokia proprietary component
<Oksana> How so? Nokia manufactured it, and nobody else did anything compatible?
<wpwrak> what does actually make it so special ?
<DocScrutinizer05> wpwrak: you disassembled the device. Look at it
<DocScrutinizer05> I'd say the exact build of that thing makes it special
<DocScrutinizer05> length, height, width, and contacts
<wpwrak> creepy feet ?
<DocScrutinizer05> ?
<wpwrak> let's see what the competition has to say about it ...
<DocScrutinizer05> huh?
<DocScrutinizer05> HUH?
<DocScrutinizer05> what are you after?
<DocScrutinizer05> something that might maybe fir onto the bay in case?
<DocScrutinizer05> fit*
<DocScrutinizer05> sure maybe such component esists and maybe it's SMT so we could solder it in place and *maybe* we could still insert the board with soldered-on AV-jack into the case
<DocScrutinizer05> please read sevice manual
<wpwrak> hmm, not much luck with that mounting method it seems
<DocScrutinizer05> for sure not, and particularly not with exactly those outer dimensions so it would fit tight into the case bay, and has same electrical/contact/switch properties as well
<wpwrak> let's see what the SMT solutions would look like ...
<Oksana> I am after having a specification of "3.5 mm headset connector" on block diagram, about ability to survive ten years in hot-dry-humid environment without getting rusty. Because on my current N900 there is green-rust visible inside it.
<DocScrutinizer05> what they would look like? like something we cannot insert into case shell anymore
<wpwrak> that's what i'm trying to determine ...
<DocScrutinizer05> that's pretty obviouus
<wpwrak> yes, but great scientific progress is made by questioning the obvious ;-)
<DocScrutinizer05> the outermost contact (sleeve) of any such connector would lock the board inside case shell
<DocScrutinizer05> reassemble you N900 and watch it. Err nope you still have the AV jack inside case shell, so it is sufficient to closely inspect it
<DocScrutinizer05> I think it gets pretty tight to get the sleeve contact that close to the outer surface of case shell with a SMT mounted connector
<DocScrutinizer05> as long as the original Nokia component is widely available, I consider if a waste of time to even investigate the alternatives
<Oksana> So is the original Nokia's component widely available? Link to it?
<DocScrutinizer05> google is your friend
<DocScrutinizer05> used on half a dozen Nokia devices
<DocScrutinizer05> aka "widely available"
<wpwrak> you'll like this one. has even three (!!!) switches :) http://www.digikey.com/product-detail/en/SJ-43617-SMT-TR/CP-43617SJCT-ND/2625183
<wpwrak> audio jack compartment is about 14.9 x 7.4 mm. so the SJ-4361X-SMT or also the SJ-43504-SMT would fit easily
<wpwrak> which would leave a) Z positioning and b) getting it close enough to the outside wall
<DocScrutinizer05> and you also considered how to actually insert the complete PCB with AV jack soldered on top into the case shell?
<wpwrak> seems that they all expect to mate "tightly"
<wpwrak> that's part of b)
<DocScrutinizer05> no it isn't
<wpwrak> wall will bend a little. but we'll need some ... 1.1 mm at least. hmm.
<Oksana> Was the original Nokia's jack also copper alloy with gold plating? Is it "normal" that the jack is unshielded? Alright, I will look for waterproof jack-caps, since I do not use the headset connector much, and I do not want it to get rusty-dusty ten years later. The Nokia's jack was about 14.7x9.4x6.5mm, according to http://www.electronicproducts.com/Nokia_N900_Mobile_Phone-whatsinside_text-90.aspx
<DocScrutinizer05> inserting PCB nto shell is like: rotate board 5° counterclockwise, lift powerbutton side a 10mm. InsertUSB into hole in shell. Rotate PCB clockwise until lower long side fits into the latches in shell. Lower powerbutton side and press it into the shell latch at this side
<DocScrutinizer05> any soldered-toPCB AV component will collide with the walls of the shell cave where the jack lives in
<DocScrutinizer05> during that procedure
<DocScrutinizer05> Oksana: yes, the AV jack is goldplated springs, unshielded. using rotten plugs or plugging a lot in sandy environment may make the gold plating come off. Then it starts rotting
<Oksana> So how would Nokia's AV component be attached? And how do other modern phones solve the problem, since they do not use Nokia's connector, right? I do not recall plugging a lot, but the environment was sandy, for a long time.
<DocScrutinizer05> using a rubber plug is an excellent idea
<DocScrutinizer05> Nokia's AV jack is spring loaded feet that connect to gold plated pads on PCB
<Oksana> How do these feet help? | Why do most common rubber plugs look like a screw? For water-tightness, or for some other reason? https://images.duckduckgo.com/iu/?u=http%3A%2F%2Fts4.mm.bing.net%2Fth%3Fid%3DHN.608046878477254915%26pid%3D15.1&f=1
<DocScrutinizer05> http://neo900.org/git/?p=scans;a=blob_plain;hb=HEAD;f=pics/large/n900-bottom-inside-top.jpg
<DocScrutinizer05> I think those with a thread are meant for device case that has a matching thread
<DocScrutinizer05> (pics of AV: http://neo900.org/stuff/werner/scans/ )
<DocScrutinizer05> so much wasted energy on a non-risk part
<DocScrutinizer05> well, ok. How's that relevant?
<wpwrak> most competing parts are much narrower
<wpwrak> so collision in that direction probably wouldn't be a bit problem
<wpwrak> however, the 3 o'clock direction is where things don't look so good
<DocScrutinizer05> I'm not eager to even discuss such non-issues
<wpwrak> do we at least have some sort of part number ? then i can already make it look a lot prettier :)
<DocScrutinizer05> how about simply keeping those nokia part #5469246 on BOM and frequently check for availability, and not get a headache for alternatives
<wpwrak> #5469246 is the nokia number ?
<DocScrutinizer05> looks like
<wpwrak> kewl. that's all i need. thanks ! :)
<DocScrutinizer05> yw
<astr> DocScrutinizer05, umm what would you suggest for a more realistic/helpful donation/pre-order size? £200?
<astr> £300+.....?
<astr> baring in mind in my case I can't afford to lose *00 and gain nothing
<astr> and yet i would be after 1gb ram
<astr> seeing about trying to bite the bullet thats all
Kabouik_ has quit [Ping timeout: 256 seconds]
<Oksana> DocScrutinizer mentioned the possibility of changing "donation limit" from 100 EUR to 200 EUR, earlier: http://irclog.whitequark.org/neo900/2014-11-19#10901855
<wpwrak> we should do that after the "risky component purchase" campaign
<Oksana> Good night
nox- has quit [Ping timeout: 260 seconds]
arcean has joined #neo900
jonwil has joined #neo900
sparetire_ has quit [Quit: sparetire_]
norly has joined #neo900
Pali has joined #neo900
mvaenskae has joined #neo900
mvaenskae has quit [Client Quit]
che1 has joined #neo900
norly has quit [Quit: Leaving.]
mvaenskae has joined #neo900
che1 has quit [Ping timeout: 240 seconds]
che1 has joined #neo900
obsed has joined #neo900
mvaenskae has quit [Ping timeout: 272 seconds]
jonwil has quit [Remote host closed the connection]
mvaenskae has joined #neo900
che1 has quit [Ping timeout: 245 seconds]
slyon_ has joined #neo900
che1 has joined #neo900
slyon_ is now known as slyon
che1 has quit [Ping timeout: 240 seconds]
che1 has joined #neo900
PeperPots_ has quit [Read error: Connection reset by peer]
unclouded has quit [Ping timeout: 265 seconds]
PeperPots_ has joined #neo900
rZr is now known as RzR
unclouded has joined #neo900
mvaenskae has quit [Ping timeout: 250 seconds]
ShadowJK has quit [Ping timeout: 244 seconds]
ShadowJK has joined #neo900
che1 has quit [Ping timeout: 265 seconds]
louisdk has joined #neo900
che1 has joined #neo900
b1101 has joined #neo900
<Pali> ouch... 120 patches for n900 kernel is currently in my git tree...
<DocScrutinizer05> why "ouch"?
<pigeons> upstream them!
mvaenskae has joined #neo900
slyon has quit [Quit: Leaving]
<Pali> and +10 from pavelm
<Pali> first step is to compile kernel 3.18-rc6 with all those patches and start it in qemu...
<mvaenskae> oh, this weekend is the workshop :o
<mvaenskae> man, this is interesting
<Pali> qemu booted :-)
<mvaenskae> Pali: maemo 5 with newest kernel?
<Pali> no qemu with busybox init
<Pali> and after 125s:
<Pali> [ 125.413787] random: nonblocking pool is initialized
<Pali> :D
<mvaenskae> wtf?! are you calculating the addresses yourself and messed up a bunch of times using the calculator? =p
<Pali> booting kernel to busybox sh in qemu take [ 1.414825]
<Pali> so only 124s for random nonblocking pool :D
<mvaenskae> well, it was a blocking pool for 124 seconds
<mvaenskae> Pali: what are you using the qemu system for then? :)
<Pali> static linked busybox (with easy /sbin/preinit which only spawn shell) in unbifs image
<Pali> this is good for testing if it boots
<mvaenskae> unbifs?! is that like a bifi?
<mvaenskae> btw, any new device numbers? :)
<Pali> ubifs
<mvaenskae> ahhh
<DocScrutinizer05> mvaenskae: I think you failed to check
mvaenskae has quit [Ping timeout: 245 seconds]
mvaenskae has joined #neo900
Pali has quit [Ping timeout: 265 seconds]
Pali has joined #neo900
<Pali> freemangordon: n900 patches for 3.18-rc6 are on gitorious
<mvaenskae> Pali: n900 patches? so you are porting the newest kernel to the neo900 platform? :o
<Pali> I'm just rebasing n900 patches on top of current kernel
<mvaenskae> cool and nice; may i recommend basing them on current mainline kernels as well if possible and maybe even get them vanilla?
<mvaenskae> that would be awesomesauce :)
<mvaenskae> megaawesomesauce i mean; this already is awesomesauce :)
trx has joined #neo900
sparetire_ has joined #neo900
arcean_ has joined #neo900
arcean has quit [Ping timeout: 244 seconds]
arcean_ is now known as arcean
nox- has joined #neo900
SylvieLorxu has joined #neo900
peterh_ has joined #neo900
che1 has quit [Remote host closed the connection]
che1 has joined #neo900
peterh_ has quit [Quit: Page closed]
che1 has quit [Ping timeout: 240 seconds]
jonwil has joined #neo900
b1101 has quit [Quit: b1101]
che1 has joined #neo900
arcean has quit [Quit: Konversation terminated!]
xes has joined #neo900
Kabouik has joined #neo900
jonwil has quit [Remote host closed the connection]
<freemangordon> Pali: great
<freemangordon> did you test the kernel?
<Pali> only in qemu
<Pali> it booted :-)
<freemangordon> BTW someone (most probably me) should fix libossoproductinfo
<DocScrutinizer05> after 3 hours
<Pali> s/booted/compiled and booted/
<freemangordon> esp while jonwil takes a rest from PA
<Pali> whats wrong with ossoproductinfo?
<freemangordon> DT boot ;)
<freemangordon> and we should change mce and whatnot to use it
<Pali> ah rigth... read device name and hw version
<Pali> but you cannot fix it
<Pali> DT does not send us hw revision
<Pali> and other data
<freemangordon> mce, ke-recv. anything else?
<Pali> it is not possible to fix it
<freemangordon> Pali: wait, why?
<Pali> because kernel do not use ATAGs
<freemangordon> the guys on the ML explained what to use
<freemangordon> aah, right
<Pali> if you want it you need to hardcode these data to DT file
<Pali> which cannot be done as those data are generated by NOLO at runtime
<Pali> (+ from CAL)
<freemangordon> but we can revert the commit that removed atags
<Pali> this reminds me, I should report that problem (again)!
<Pali> no we cannot
<freemangordon> why?
<Pali> kernel just does not implement atag parsing for DT
<Pali> there is no commit which delete this support
<freemangordon> we'll implement it, I doubt it is a rocket science
<Pali> yes, needs implement
<freemangordon> we can reuse the code, iirc it was some ~100 lines
<freemangordon> (form omap1 kernel)
<freemangordon> Pali: how's your free time lately?
<Pali> I have some free time
<freemangordon> we need more people to work on PA
<freemangordon> voice module is huge
<freemangordon> if we want it REed ofc :)
<Pali> I rather look at kernel hacking...
<freemangordon> remember what sebastian said
<freemangordon> re cmt-speesh
<freemangordon> but sure, it is your choice how to spend your free time :)
<freemangordon> anyway /me goes afk, GF's calling :)
<Pali> freemangordon: I sent mail, see *nothing* is working
<Oksana> nothing is working?
<DocScrutinizer05> everybody is working ;-)
<DocScrutinizer05> Oksana: the concept of *nothing* 8nirvana) is actually working since ages ;-D
<DocScrutinizer05> so "nothing is working" is a real zen saying
<DocScrutinizer05> s/real/true/
<DocScrutinizer05> "a true zen saying: 'noting is what I want' " (F.Z.)
Oksana has quit [Read error: Connection reset by peer]
mvaenskae has quit [Quit: Lost terminal]
Oksana has joined #neo900
<Pali> ok, all informations from NOLO are on DT booted n900 dropped
<Pali> not only from NOLO but any bootloader
<Pali> and main ARM maintainer does not care about it
<DocScrutinizer05> wpwrak: ^^^ any comments? IIRC you're bootloader expert
<wpwrak> i can't quite parse that :)
<wpwrak> in general, i like it if the kernel doesn't need a lot of information from the boot loader. if this is what it's about
<wpwrak> (keep interfaces narrow)
<DocScrutinizer05> basically that is what it's about, yes. ATAGs
<DocScrutinizer05> bootloader telling kernel about device name, memory config, dunno what
<DocScrutinizer05> start reading backscroll at >>[2014-11-24 Mon 23:06:43] <Pali> it is not possible to fix it<<
<DocScrutinizer05> or even 5 lines before
<Pali> memory config is only one atag which is read by kernel
<Pali> all other configs are read from DTS
che1 has quit [Remote host closed the connection]
<DocScrutinizer05> so you say "it's only the memory config that's announced to kernel by atag"?
<DocScrutinizer05> would make quite a bit of sense since bootloader needs to initialize memory to have some space where to load kernel to
<Pali> looks like, yes
<Pali> and cmdline
<Pali> (and I belive initfs too, but not tested)
<DocScrutinizer05> werner is the bootloader crack :-)
<DocScrutinizer05> just we don't have initfs usually ;-)
che1 has joined #neo900
<DocScrutinizer05> wpwrak: ^^^
<DocScrutinizer05> cmdline makes sense since you can provide it via flasher/dfu-util
<DocScrutinizer05> initfs= is a cmdline parameter afaik
louisdk has quit [Read error: Connection reset by peer]
norly has joined #neo900
<DocScrutinizer05> wpwrak: CAL is for NOLO (Nokia bootLOader) what env is for uBoot
<DocScrutinizer05> just one difference: CAL is also used by maemo linux quite a lot
<DocScrutinizer05> ~cal
<infobot> cal is, like, a calendar. try $(cal 1752)
<DocScrutinizer05> ~#maemo cal
<infobot> somebody said #maemo cal was http://talk.maemo.org/showthread.php?t=20465
<bencoh> does that mean we could easily end up with CAL corruption and nolo unable to boot ?
<bencoh> (if maemo plays with CAL)
<wpwrak> as far as i get what those parameters are, nothing here sounds unsolvable. worst case: board-specific code has to do some decoding. and yes, the command line can also be helpful.
<wpwrak> for testing, just hard-code :)
<DocScrutinizer05> bencoh: NOLO will boot anyway (on N900), but you might miss R&D mode, device hw rev, BT & WLAN MAC addr, etc
<DocScrutinizer05> CAL is pretty much unlikely to corrupt, but yes it's possible. We had our used we called "nandtester" since he started nandtest on complete /dev/mtd and then wondered why his device acted weird
<DocScrutinizer05> CAL is actually the only data structure on N900 you cannot regenerate/recover with a reflash
<DocScrutinizer05> you however can (probably, unconfirmed afaik) restore CAL under maemo using mtdtools to write a CAL image you created before damage of CAL
<bencoh> hmm
<bencoh> do we have a public CAL image ?
<bencoh> (available for anyone messing its cal part)
che1 has quit [Remote host closed the connection]