Wizzup | sunshavi: agreed | 00:21 |
---|---|---|
Wizzup | err sixwheeledbeast agreed* | 00:21 |
freemangordon | tmlind: I looked in the code and WDS is already created by qmimodem/gprs.c | 10:01 |
freemangordon | it turned out there is a nasty bug the qmimodem code | 10:01 |
freemangordon | qmi_service_create_shared() is not written very well... to say at least | 10:02 |
freemangordon | it checks if a service of type 'type' is already created and if not, it creates one. *but* service creation is async, so while creation is going on, if another qmi_service_create_shared() call happens, we stil don;t have created service and another create_service() call is made. etc. | 10:08 |
freemangordon | Wizzup: ^^^ is the reason for missing contexts end client ids exhaustion | 10:09 |
freemangordon | if a rainy day happens while I am on holiday I will fix it, if not - when I am back. | 10:17 |
Wizzup | good sleuthing | 10:32 |
tmlind | freemangordon: unless we kick the usb qmi modem on n_gsm wds notifications, the usb qmi modem won't notice any changes | 11:16 |
dsc__ | I updated conversations with support for sending multi-line messages and (optionally) forcing send when Enter/RETURN is pressed. | 13:32 |
dsc__ | and the text on the button 'Enter' has been changed to 'Send' as per the rest of the world | 13:32 |
dsc__ | https://plak.infrapuin.nl/selif/d401b9md.png (default is disabled) | 13:34 |
sunshavi | dsc__: quite nice. What I missed on old fremantle is an api for sending sms's from the cli. There was a python script. But that python script is disconnected from the database | 14:22 |
dsc_ | sunshavi: yes I'd also want this, for now I'm focussing on fundamentals but eventually we can support such features | 14:39 |
sunshavi | nice. Now I am using an xdotool script that send the text to the conversation window and press the send button (with xdotool) also on freemantle | 14:41 |
dsc_ | lol :D | 14:42 |
dsc_ | yeah, there will be an API eventually | 14:42 |
dsc_ | 'our conversations' is made with Qt, so its easy to spawn up a HTTP server (for e.g. a JSON API) in the same process, or we provide CLI flags for it, or ... <insert solution> | 14:44 |
dsc_ | whatever the implementation, it will be trivial | 14:44 |
dsc_ | actually, a websocket server would be fancy (and more suited than push/pull) | 14:46 |
dsc_ | I mean, more suited than polling | 14:46 |
dsc_ | so with Python you could connect to ws://127.0.0.1:8395 (conversations) and issue commands like 'sendSMS' or 'listContacts' etc. | 14:47 |
dsc_ | but for now not relevant as I'm still implementing basic features | 14:48 |
Wizzup | telepathy could also just be the api? | 14:48 |
dsc_ | Wizzup: the issue is rtcom registration right | 14:48 |
Wizzup | no, conversations will get all tp messages and handle them | 14:48 |
Wizzup | and log them | 14:48 |
dsc_ | hmm | 14:49 |
dsc_ | direct communication with TP would be fine but I think we should create a small wrapper python script for it, so people dont get exposed to dbus internals and lose all hope in humanity | 14:53 |
dsc_ | just a simple api that allow people to do `send_message(protocol_type, recipient, message_body)` | 14:53 |
dsc_ | it might even already exist | 14:54 |
Wizzup | tp scripts for this exist fwiw | 14:58 |
Wizzup | but yeah | 14:58 |
sunshavi | Rigth. That would let us send sms's from ssh when phone ir charging some meters away from us | 15:04 |
Wizzup | there are scripts to do this with tp | 15:07 |
Wizzup | with my new incoming tp code we'll just find it | 15:07 |
Wizzup | we can test it later | 15:07 |
Wizzup | dsc_: yes @ wrapper script btw | 15:33 |
Wizzup | let's look at it once we have a mvp | 15:33 |
dsc_ | yupyup | 15:36 |
sicelo | Pali: nice re:u-boot + n900. i saw the recent commits | 19:23 |
Pali | Recently I sent patches which extend CI tests for UBI booting. | 19:23 |
Pali | But UBIFS on top of UBI is still broken on n900 and qemu, no idea why. | 19:24 |
sicelo | i have a question regarding bq24150 - what exactly is `charge_current` for in sysfs? it's writable, but i don't seem to notice any specific differences | 19:24 |
Pali | Interesting is that UBIFS is working fine on other bigendian powerpc board (with Freescale P2020 CPU). | 19:24 |
Pali | So soemthing has to be terible broken in u-boot. | 19:25 |
Pali | charge_current? let me search for it | 19:25 |
sicelo | https://github.com/torvalds/linux/blob/master/drivers/power/supply/bq2415x_charger.c#L620 | 19:25 |
sicelo | re: UBIFS, yes i read the comments. btw, recent kernel spams dmesg with nonstop `MTD device '%s' is NAND, please consider using UBI block devices instead`. but looks like this can be resolved by disabling some MTD stuff in config | 19:27 |
sunshavi | Pali: I also see the messages on the u-boot mailing list. And Tony answer | 19:28 |
Pali | https://www.ti.com/lit/ds/symlink/bq24150a.pdf | 19:32 |
Pali | page 30 | 19:34 |
Pali | It configures charge current sense voltage based on resistor and sysfs charge current mA value | 19:34 |
sicelo | yes, i can see it. it's the effect of adjusting this value which i'm struggling to understand | 19:38 |
sicelo | i.e. what's the benefit of using 950 instead of 650, for example | 19:39 |
Pali | sicelo: I was looking into old notes to finally confirm that it specifies current for constant charging current mode (CC) | 20:45 |
Pali | so for CC chaging, if you use higher current, battery should be charged more quickly. | 20:46 |
Pali | this is what is called nowadays "fast charge" or similar option in laptops. | 20:46 |
Pali | I'm not sure what happens with these BL-5J batteries if they are charged with higher current, how it decrease lifetime. | 20:47 |
Pali | And I really do not remember what should be the default value. I have not touched this part for years. | 20:47 |
Pali | I found some old API notes: http://maemo.cloud-7.de/bq24150-sysnode.spec.txt | 20:48 |
Pali | default values are in DTS: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/arch/arm/boot/dts/omap3-n900.dts#n749 | 20:57 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!