<buZz>
[ERROR @radicale] @radicale/radicalecal: error code from SyncEvolution remote, status 400: PUT: bad HTTP status: <status 1.0, code 400, class 4, Bad Request>
<buZz>
radicale isnt allowing me to upload unnamed events :P
<buZz>
hmmz, ok , deleting it fixed it , but GUI sync seems to fault while commandline sync goes fine
<buZz>
ahh, same thing on contacts with a 'nickname' but no real name :P
<freemangordon>
sicelo: just send that to the ML, it seems like something simple to be fixed, I just lack the knowledge to do it
Twig has joined #maemo-leste
ceene has joined #maemo-leste
mardy has joined #maemo-leste
<sicelo>
freemangordon: yes, thanks. i came to that conclusion too. at least now i have better logs to share
<freemangordon>
:nod:
xmn has quit [Ping timeout: 268 seconds]
Twig has quit [Remote host closed the connection]
alex1216 has joined #maemo-leste
akossh has joined #maemo-leste
<Wizzup>
buZz: nice @ stand
pere has quit [Ping timeout: 260 seconds]
dev has left #maemo-leste [Disconnected: Replaced by new connection]
dev has joined #maemo-leste
uvos__ has joined #maemo-leste
<uvos__>
sdl_compat dosent work with hildon for several reasons
<uvos__>
one is that it fights with hildon over the window size (if the window is not the right size it just resizes its own window in a tight loop) pretty terrible behavior
<uvos__>
it also fails to init gl when glx is not avaialble
<uvos__>
so rn it would require some work for it work on leste at all
<uvos__>
never mind edge cases
<uvos__>
with createive sdl1 usage
<Wizzup>
aha
<uvos__>
we used to need patched sdl2 beacuse of brokenness in hildon
<uvos__>
wrt fullscreen
<uvos__>
this was fixed
<uvos__>
so afaik now we should be able to use devuan sdl2
<uvos__>
i think we might be doing so
<uvos__>
and the repo is just old
<uvos__>
but its something to check
<Wizzup>
mhm
<uvos__>
buZz: the power button thing
<uvos__>
we might really want to report
<uvos__>
i had to hack around that in charging_sdl
<sicelo>
n00mann[m]: btw, nowadays buying an n900 is not exactly advisable, due to the device's age and lack of resources (RAM in particular), unless perhaps one is an ex-n900 owner, in which case there may be other reasons at play :-)
Twig has quit [Ping timeout: 265 seconds]
<sicelo>
in general, maemo leste actively promotes the motorola droid 4, and there are good reasons for that. if willing to put in the work to port and/or improve other devices, there are other more recent devices to choose from, which are much faster
<buZz>
i hope we can find some nice armv8 with full qwerty someday
<buZz>
the 'gemini' looked a bit promissing, but not really that ergonomic
<buZz>
sicelo: would you call pinephone a more recent device?
<n00mann[m]>
The rf functionality of n900 is pretty cool
<sicelo>
n00mann[m]: yeah i still love n900, even though i have droid 4. runtime stability-wise - seems unbeatable :)
<n00mann[m]>
A universal remote turning appliances in the house
* sicelo
is expecting to get two mint N900 in a few weeks
<sicelo>
buZz: tbh, i think pp is a great device
<buZz>
n00mann[m]: rf? IR :)
<buZz>
unless you ment the FM transmitter
<buZz>
sicelo: i like those armv7 allwinners, but i dont think its 'great' really
<buZz>
its great fun though :)
<buZz>
but it feels more like a pi2 with a fancy case & screen & battery & some sensors
<buZz>
than a actual phone
<buZz>
even though formfactor is totally phone
<buZz>
for some application that wants OpenGL , could i wrap gl4es inside the package?
<sicelo>
buZz: maybe n00mann[m] meant *both* fmtx and ir :-P
<buZz>
so i dont have to do it systemwide like 'apt install gl4es' does
<buZz>
sicelo: hehe , then still IR isnt RF
<buZz>
unless perhaps in the 'everything is waves' idea :P
<sicelo>
although ir is broken in mainline for a while now ... insta-oops
<buZz>
aw
<sicelo>
buZz: IR is RF, if we're being pedantic ;-)
<freemangordon>
sicelo: good reading
<buZz>
:)
pere has joined #maemo-leste
<buZz>
fmtx is such a nice feature though
<sicelo>
freemangordon: yeah, i also finally found some time to read it. seems a good start. just hoping it isn't too late (when other SoC's have taken the mobile linux world by storm)
<freemangordon>
yeah, I wonder how useful it would be
mcafeeshred has joined #maemo-leste
mardy has quit [Quit: WeeChat 3.5]
<uvos>
ir is deffinatly not rf
<uvos>
rf is radio frequency
<uvos>
ir might be electromagnetic radation
<uvos>
but in the radio frequency band it is not
<uvos>
there thats sufficantly pedantic now :P
<buZz>
^_^
<buZz>
cookies for everybody
norayr has left #maemo-leste [Disconnected: Replaced by new connection]
norayr has joined #maemo-leste
<uvos>
sicelo: yeah i saw that
<uvos>
sicelo: its neat, if he keeps at it maybe eventually usefull. that said idk if basing a new driver for just sgx on the old pvrkm + blobs is the best approch vs helping the img guys add sgx support to the mesa driver they are writeing for series6, they dident seem opposed to adding sgx suppot in the mesa ml, it was just not something that makes sense for them finacially.
<uvos>
so maybe you could even get some documentation out of them if you seam serious.
<freemangordon>
uvos: I doubt they will open anything series5, I have the feeling they have contracts that stops them from doing so
<uvos>
then again the guy in the blog is just messing around trying to figure out the instrcution set
<uvos>
so maybe this is his goal even
<uvos>
freemangordon: maybe idk
MartijnBraam[m] has joined #maemo-leste
<freemangordon>
and yeah, they have absolutely no reason to do it, from the financial POV
<uvos>
that aside, for sure they have contractual issues with the code in thair closed driver
<freemangordon>
mhm
<uvos>
they arnt rewriteing the series6 driver from scratch for fun
<freemangordon>
keep in mind that SGX was used in apple devices
<uvos>
i know
<freemangordon>
so I would bet on that one :)
<uvos>
they totaly shot themselves in the foot with that
<freemangordon>
looks like
<freemangordon>
are they still relevant BTW?
<uvos>
no
<uvos>
they have very few design wins
<freemangordon>
like, besides Ti, nobody is using IMG stuff, no?
<uvos>
mediatek
<uvos>
used it some times
<freemangordon>
well, I mean nowadays
<uvos>
yeah nowadays, some reasonably recent chips i think
<freemangordon>
ok
<uvos>
like 2018 recent
<freemangordon>
well, that's relatively new
<uvos>
so still sbc/industrial relevent
<uvos>
sure
<freemangordon>
mhm
<uvos>
but not smartphone relevant
<MartijnBraam[m]>
allwinner has IMG gpus in some socs :(
<uvos>
right them too
<freemangordon>
old ones I guess
<MartijnBraam[m]>
rockchip has it in the "high performance" parts
<uvos>
still its very much fallen of a cliff compeard to sgx times
akossh has quit [Quit: Leaving.]
<Wizzup>
don't those allwinner chips work with the closed driver that we have working?
<uvos>
Wizzup: sure yeah the chromeos driver is for one
<freemangordon>
it has hardcoded 'omapdrm" in it, so I doubt
<sicelo>
well the writer of that blog seems to know at least some part of what they're doing, on short chat with them in a postmarketOS room
<uvos>
we also have the blobs for the chomeos driver
<uvos>
btw
<freemangordon>
sicelo: sure the guy looks smart and knowledgeable, the point is - what will be the use of series5 driver in mease in 2 years from now?
<freemangordon>
*mesa
<uvos>
its for some series6 part
<freemangordon>
there won;t be any functioning hardware left by then
<sicelo>
freemangordon: yes, agreed at the usefulness in that sense (hence i earlier mentioned that it might be too late)
<MartijnBraam[m]>
isn't most of the maemo leste hardware support on sgx530 and sgx540? :D
<freemangordon>
yes, but we don't have much issues with the blobs
<freemangordon>
if ay
<freemangordon>
*if any
<freemangordon>
like, I would prefer foss driver but well, not end of the world either
<freemangordon>
as we have more serious issues than that
<buZz>
the panfrost stuff looks nice
<buZz>
i'd like some system with one of their supported malis
<buZz>
i guess thats what we use on pinephone eh :P
<Wizzup>
no, that is lima
<MartijnBraam[m]>
panfrost for the pinephone pro
<buZz>
ah ok, because its older generation?
<buZz>
do we have a mainline for pinephone pro yet then? :P
<freemangordon>
buZz: was that you that said taht n900 outperforms PP in compiling something?
<buZz>
well, in perception , i can measure now
<freemangordon>
is that braveheart?
<buZz>
eh, the 3GB model i think
<freemangordon>
hmm. ok
<buZz>
yeah the 3GB one, and it was single threaded compile btw
<buZz>
want me to measure it?
<MartijnBraam[m]>
huh the n900 shouldn't outperform a pinephone
<MartijnBraam[m]>
even single threaded
<buZz>
i'll go measure, maybe i was just seeing things
<sicelo>
yeah, i'm quite sure there was something else involved there
<freemangordon>
MartijnBraam[m]: agree, unless he hits the same issue I have on braveheart
<buZz>
at least 'cmake' finishes faster on pinephone :P
<freemangordon>
MartijnBraam[m]: ever heard of something like that?
<MartijnBraam[m]>
I know the pinephone starts doing weird things when you load it heavily
<MartijnBraam[m]>
mostly the gpu driver crapping out when it can't hit timings anymore
<buZz>
they are both on usb power for >24hrs
<xmn>
I would think it possible if the n900 is well optimized, no?
<MartijnBraam[m]>
at least that's better on the ppp
<MartijnBraam[m]>
but the ppp development is still a mess
<buZz>
screen is off , doing stuff over ssh
<freemangordon>
xmn: no way, it is 600 MHz a8
PureTryOut[m] has joined #maemo-leste
<freemangordon>
vs 1.3GHz a52?
<buZz>
a53
<freemangordon>
yeah, a53
<xmn>
haha, I know. But you can get it too 1.1.
<MartijnBraam[m]>
on my benchmarks the n900 a lot slower than the a64
<freemangordon>
which makes sense
<xmn>
but I agree that my PP feel mostly faster then my n900.
<MartijnBraam[m]>
hmm I don't think I saved the single thread test results
<freemangordon>
unless you bench the GPUs
<sicelo>
xmn: n900 is slower than everything nowadays :-)
<xmn>
and a lot fast in many case
<buZz>
its no issue to do this , i'll let em both compile devilutionX again, singlethreaded
<freemangordon>
I am not sure which one is slower :)
<MartijnBraam[m]>
but it's 52 cpumark points for the pp and 1300 for the a64
<sicelo>
freemangordon: its the sgx540 that gives pp's mali competition. not the 530
<MartijnBraam[m]>
buZz: no it's allwinner a64 with a cortex-a53
<MartijnBraam[m]>
to make naming more confusing
<xmn>
yeah, it hurts sometime. I don't do web browser on it anymore. But cmdln and specific apps are still fantastic for my use case.
<buZz>
MartijnBraam[m]: ? yes i know
<buZz>
it was 'a8 vs a52' , i corrected it to a53 ;)
<freemangordon>
:nod:
<xmn>
It's like a horse race, were do we place our bets
<MartijnBraam[m]>
huh I lost a message
<sicelo>
xmn: yes, i still love it too. i love that it's stable and predictable for my use cases too (unlike droid 4, unfortunately)
<buZz>
MartijnBraam[m]: probably the superior almost-irc you're using ;)
<xmn>
oh interesting. Never got the D4.
<MartijnBraam[m]>
bleh
<xmn>
lol
<freemangordon>
sicelo: huh? d4 is still not stable enough?
<sicelo>
not for me, unfortunately
<buZz>
it is for me
<freemangordon>
ah, the battery I guess
<buZz>
oh yeah
<buZz>
it can be better :D
<freemangordon>
anyway, zzz
<freemangordon>
night!
<buZz>
i really want some of those flexpcbs made soonish, and source some fitting new batteries
<buZz>
nn freemangordon
<sicelo>
ever since i disabled charge mode, at least powering up from empty has been better ...
<buZz>
:)
<buZz>
i have a love/hate relationship with it too
<buZz>
its complex
uvos has quit [Remote host closed the connection]
<sicelo>
droid4 modem isn't (currently?) as reliable as the n900's, whether you use modemmanager or ofono. additionally, my isp's make heavy use of 'session' type ussd calls, which droid can't do well, even under android
<buZz>
i can do ~24hr connection to gprs with mosh running quite well
<buZz>
as long as i keep battery >30% or something
<sicelo>
i.e. *xxx#, then you get a response, and you reply back, etc. d4 just fails
<sicelo>
works fine for ussd one-shot requests though (android)
<sicelo>
buZz: will test gprs again then :-)
<buZz>
:)
uvos has joined #maemo-leste
<buZz>
a friend of mine has LTE hardware and can spoof a celltower, he's thinking he might be able to do the '4G' the droid4 supports :P
<sicelo>
last i used it, would work fine first time, then next time starts acting up in different ways (that's why i even looked at modemmanager)
<uvos>
buZz: yeah i knda get the fealing that the d4 is less stable at low voltages too
<uvos>
like if it hangs - wd reboots
<buZz>
hmhm, either the cpu, or the modem
<buZz>
or something else
<uvos>
its allways when the battery is quite low
<uvos>
it seams
<uvos>
maybe there is some hw errata we fail to account for
<buZz>
yeah perhaps, or just something we forget to do on voltage drop, like limitting current even more?
<buZz>
i dont know
<uvos>
the lte thing would be interesting
<uvos>
would be good to know if the lte modem is locked to verizon or not
<buZz>
well, he can spoof being verizon :P
<uvos>
sure but he can also spoof it not being verizon
<uvos>
which is more interesting
<buZz>
it wont work ~5cm from the radio (without legal implications)
<buZz>
but, he can document i guess, i'll ask him about progress soonish
<Wizzup>
sicelo: for my the modem has been super reliable since fmg's fixes
<Wizzup>
but iirc you need ussd or something
<buZz>
afaik ofono does expose some ussd stuff?
<Wizzup>
yes but it likely doesn't work atm.
<sicelo>
yes i definitely need ussd, e.g. to check balance, convert credit into data allocation, etc. seems in many countries this can be done via web sites. no such convenience here
<sicelo>
buZz: yes, ofono has very good ussd support, which works beautifully on n900. as mentioned, the ussd issue on d4 exists even under android, at least when it's the request-response type of ussd session
<buZz>
oh ok
<uvos>
i still find this wierd since ussd stuff worked absoulty fine when i was using prepaied
<buZz>
yeah *101# on AH Mobiel to check balance :P
<uvos>
but i gues the modem somehow trips over sicelo's operators implementation
<sicelo>
buZz: one shot works fine for me on android too. the problem is with 'session' ... i think those aren't used in US/EU much
<buZz>
sad :(
<sicelo>
here they're used for almost everything, because - feature phones ;-)
<sicelo>
but nice to know gprs is stable now. i'm already thinking of some extra use for the droid4
<sicelo>
's gprs ;-)
<buZz>
:)
<buZz>
i bet with a decent usb powersource you can run it as server for ages
<buZz>
add wireguard or one of the other vpn methods we have
<sicelo>
heh, i would only use broken phones as servers ... why waste a useful phone ;-)
<buZz>
¯\_(ツ)_/¯
<sicelo>
someday hope to get a better battery for it (d4)
dev has left #maemo-leste [Disconnected: Replaced by new connection]
dev has joined #maemo-leste
dev has left #maemo-leste [#maemo-leste]
dev has joined #maemo-leste
sunshavi has quit [Ping timeout: 252 seconds]
<buZz>
yeah some 1w battery :)
<Wizzup>
off mode will contribute more than a battery
<buZz>
lets hope :)
<buZz>
does d4 have offmode too?
<Wizzup>
the hw, yes
<buZz>
i thought that was only n900
<buZz>
ah
<buZz>
cool :)
<buZz>
either way, 1w without charging would be bliss
elastic_dog has quit [Ping timeout: 264 seconds]
elastic_dog has joined #maemo-leste
<uvos>
my bionic usualy gets 1w
<uvos>
well more like 5 days
<uvos>
bit its fairly close
<Wizzup>
you have a special battery :)
<Wizzup>
my worn bionic gets me 2 days max
<buZz>
ok guess i was seeing things ; PP real 31m51 user 28m1 sys 2m24 , n900 real 76m50 user 65m23 sys 6m16
<buZz>
even the ratios seem to align :P
<sicelo>
poor n900 :p
<buZz>
still , just ~2x faster , not 20x like cpumark said
<uvos>
thats compileing?
<uvos>
probubly io bound
<uvos>
raw cpu perf should be way more than 2x
<uvos>
d4 is allready around 3x on just one core
<buZz>
yeah , make -j1
<uvos>
ok
<uvos>
yeah the n900's cpu is pittifully slow
<uvos>
but the other devices arnt slow but as mutch of a ratio
<uvos>
s/but/by
<uvos>
ie ram/gpu/io
<buZz>
still, imho quite impressive for n900 vs pp
<buZz>
i'll try d4 later
uvos__ has quit [Ping timeout: 260 seconds]
uvos has quit [Ping timeout: 265 seconds]
<Wizzup>
buZz: I mean this comparison isn't too interesting no?
<buZz>
i'm just doing it cause freemangordon asked
<buZz>
was*
<buZz>
the d4 is now just fascination by me
<buZz>
would it be ~40m or ~60m :P
<buZz>
Wizzup: this isnt really work :P its just 'time make -j1' alt-tab back to series
<Wizzup>
;p
<buZz>
i mean, would be funny if its faster than pp :P
<buZz>
d4 real 29m user 18m sys 2m7 , omg, it -is- faster than pp
<Wizzup>
i/o matters a lot here
<buZz>
totally, but io matters a lot for user experience too
<Wizzup>
what I am getting is that it could be kernel related
<buZz>
oh, yeah can be, its just 'current state' i guess