Current Server Time: 16:18 (Central Europe)

#apertus IRC Channel Logs

2021/05/10

Timezone: UTC


03:27
lambamansha
joined the channel
05:20
lambamansha
left the channel
09:31
se6ast1an
good day
09:34
BAndiT1983
hi
10:02
Spirit532
left the channel
10:02
Spirit532
joined the channel
11:52
Bertl_zZ
changed nick to: Bertl
11:52
Bertl
morning folks!
12:07
RexOrCine
left the channel
12:07
RexOrCine
joined the channel
13:37
RexOrCine
left the channel
14:05
RexOrCine
joined the channel
14:57
Bertl
off for now ... bbl
14:57
Bertl
changed nick to: Bertl_oO
15:09
__m__
joined the channel
15:48
daFred
joined the channel
15:58
RexOrCine
Ah... I see.
15:59
Bertl_oO
changed nick to: Bertl
16:01
Cscar
joined the channel
16:01
se6ast1an
hi everyone, if you are here to join the general assembly please go to the channel as described in invitation email
16:22
daFred
left the channel
16:26
daFred
joined the channel
16:51
__m__
left the channel
16:51
__m__
joined the channel
17:03
bluez
se6ast1an: invitation email?
17:04
se6ast1an
hi bluez
17:04
se6ast1an
we will delay the start of the IRC meeting a bit today
17:04
se6ast1an
as the association is currently holding its general assembly
17:05
bluez
ah, i see.. np :)
17:07
daFred
left the channel
17:08
daFred
joined the channel
17:26
__m__
left the channel
17:39
se6ast1an
Meeting time!
17:39
se6ast1an
bluez: now we are here :)
17:40
se6ast1an
as we are already 40 minutes late lets speed things up here
17:40
bluez
is too :)
17:40
se6ast1an
anyone who wants to report/share/discuss anything?
17:40
se6ast1an
I will give a brief status update of progress since last week
17:41
bluez
will you start?
17:41
se6ast1an
so bluez and metal_dent[m] have finished their internship this week, dont leave us :)
17:42
se6ast1an
I will now prepare the documents this entails
17:42
se6ast1an
other progress
17:42
bluez
don't intend to :)
17:42
se6ast1an
hurray
17:42
Bertl
is glad to read that ;)
17:42
metal_dent[m]
is present but very little to report
17:42
se6ast1an
I have received our magewell hdmi capture device and successfully tested it today
17:43
se6ast1an
it will go to Bertl soon to replace the french research group one loaned to us for evaluation
17:43
se6ast1an
manfred has depaneled the Tele produced PCBs and handed them over to me today
17:44
se6ast1an
the Sensor boards have been handed back to tele with the image sensor sockets inserted for soldering
17:44
se6ast1an
it turned out the drills diameters were done so good this time that no force was required to insert the sockets to the proper depth
17:44
se6ast1an
so next time we can just let Tele do the insertion
17:45
se6ast1an
I will now scan all finished boards on a high resolution flatbad scanner and Bertl will review the placements and boards in general
17:45
eppisai
is present and little to report :)
17:45
se6ast1an
the association has just held its general assembly and the board has been (re)elected:
17:46
se6ast1an
Chairman: me, Vice chairman: Oscar (Cscar)
17:46
se6ast1an
advisory board: Bertl & vup
17:47
se6ast1an
regarding GSoC we have now requested slots for students but Google does not allow us to share the allocation yet
17:47
se6ast1an
slots will be announced by google on 17th
17:47
se6ast1an
May 17 18:00 UTC Accepted student projects announced
17:49
se6ast1an
thats it from my side I think
17:49
se6ast1an
metal_dent[m]: & eppisai please share your very shot report/summary
17:50
metal_dent[m]
hi, i'm preparing my final detailed internship report which i'll soon share with se6ast1an and BAndiT1983|away and also a presentation which i need to do in front of my college panel
17:51
metal_dent[m]
that's it
17:51
metal_dent[m]
and yeah se6ast1an i also won't leave the org so soon ;)
17:51
se6ast1an
great
17:51
eppisai
Hi everyone ! :)
17:52
eppisai
Have just been improving code for keyboard screen (after going through it again,I saw some things can be improved upon and can be made less confusing by following SOLID Principle)
17:52
eppisai
Also, I am moving keyboard to a widget class, and will write unit test for the same and probably complete in a day or two.
17:52
eppisai
Other than that, for my current draft pull requests,
17:52
eppisai
1. Will talk with BAndiT1983 and see if their are changes to be made in current windows build instructions of visualizer.
17:52
eppisai
2. Will see if I can make some progress in stack analysis again(been more than 2 weeks since I last tried). So, any changes I make in firmware will actually be useful, and finally can open/create pull requests for all my firmware/UI changes. Hoping to collaborate in this regard with Andrej and Sebastian, if they have some free time :)
17:52
eppisai
So, Curial part is stack analysis for me now I think, for any changes I make (even small ones like moving GetStringBufferWidth out of painter object) in firmware.
17:52
eppisai
Week has just started! :D
17:52
eppisai
Thought it might be a short meet, so has typed report in wait time :D
17:53
eppisai
*have
17:54
eppisai
*had
17:54
se6ast1an
many thanks!
17:54
se6ast1an
bluez: anything from you?
17:54
bluez
yup
17:55
bluez
so the issue with mapping logic only to SLICELs and not SLICEMs was finally resolved
17:55
bluez
for some reason it required tweaking a weird option, as mentioned by someone in the XIlinx forum, but it works thankfully
17:56
bluez
with that now it's mostly safe to reconfigure our LUTs
17:57
bluez
and more importantly: by combining all other scripts, i wronte the main interface script
17:57
bluez
https://github.com/Swaraj1998/reg-test/blob/master/scripts/reg.py
17:57
bluez
this demonstrates the whole process
17:58
bluez
a command like: python reg.py REG_32_CONST 0 -w 0xdeadbeef
17:58
bluez
will write the value to the specified register instance... and the specified register index (0 here out of [0-31])
17:58
bluez
value can be read with '-r' option
17:58
bluez
so the final implementation is (kinda) complete!
17:59
bluez
now most of the task is to optimize the whole process
17:59
bluez
currently it takes too much time for it to be usable... almost 2 minuites to write and 1 minuite to read ':D
18:00
bluez
most of it is because of the processing of the scripts, with a lot of calculations and i/o and stuff
18:00
bluez
so all that can be optimized a lot
18:01
bluez
i will be working on that.. but for now i will also focus on my final report and the final presentation...
18:01
bluez
so that's it from me!
18:01
se6ast1an
great, many thanks!
18:01
Bertl
well, it's a great proof of concept and I'm glad you finished it just in time for the internship! congrats!
18:02
bluez
yup... it is indeed a 'PoC'.... and yea i'm glad too at least that much is done ':D
18:03
Bertl
looking forward to the final report and the future optimizations!
18:03
bluez
thanks to you ofc... wouldn't have done it surely without ur help ;)
18:03
bluez
yup, me too!
18:03
Bertl
you're welcome!
18:04
se6ast1an
vup/anuejn any news?
18:04
anuejn
has no relevant news
18:04
vup
I have a little
18:05
vup
So I continued working on the control daemon
18:05
vup
this week I mostly focused on the RPC part
18:06
vup
I fleshed out the design there a bit more. The general idea is to have something similar to a directory structure
18:06
vup
each entry in a directory can either be another directory, a function which can be called, or a hook, which causes the daemon to notify you if the hook triggers
18:07
vup
Each directory is given a "type". For example the directory containing the CMK12k functions (like setting exposure time and gain) would have "CMK12K" as a type and so on
18:08
vup
Furthermore each entry can only be accessed with the proper authentication
18:08
vup
for example we might want to allow everbody to change the gain of the sensor, but loading the bitstream should be something only a authenticated user can do
18:09
vup
I am still not sure on the authentication part, maybe we can defer this completely to something like network authentication (you can only access the daemon via network, so if you cannot access the network the camera lives in, you obviously also don't have access to the daemon)
18:10
vup
So I would like input on wether authentication is something you think is useful / necessary or not.
18:10
vup
Finally I also filled out how I currently envision using the daemon from bash / python / etc would look like: https://codimd.niemo.de/Jv9uOS8pTeufU0_MwUsDjA?both#usage-vision
18:10
Bertl
I think it makes perfect sense
18:11
vup
Bertl: having authentication in the daemon itself?
18:12
vup
that is it from we this week I think, not a lot for show, but atleast in my mind the RPC is now a lot more fleshed out :)
18:13
Bertl
we could defer the authentication to pam
18:13
vup
yeah, pam is the first "authentication backend" i would implement
18:13
Bertl
which might allow configuring the level required on an individual basis
18:14
Bertl
i.e. you don't want any, just adjust the settings
18:14
Bertl
you want high security, go crazy!
18:15
vup
what I more mean is, do we want to do handle authentication in the daemon itself at all, or do we want to use some external mechanism, like VLAN
18:15
Bertl
ah, hmm, good point
18:15
vup
(or only having the daemon listen on the wifi network interface, and just having a password for the wifi network)
18:16
Bertl
wifi networks are notoriously weak regarding security
18:16
tpw_rules
my 2 cents: i think it would be good to have something inside the camera itself? otherwise you need a certain amount of control over the network infrastructure. which i know i wouldn't have in my lab for instance
18:16
Bertl
so that might not be sufficient for those folks relying on security
18:17
Bertl
but I wouldn't spend too much time on this aspect, just make sure that some reasonable security can be implemented ;)
18:17
vup
sure, that is kind of my view aswell, I would like to be able to just make the camera accessible on the internet and still be able to restrict who can access it
18:18
vup
Another point is granularity, do you just allow access to the daemon or not, or do you allow access to specific (RPC) functions for specific users
18:19
Bertl
indeed, but if you plan to implement fine granularity for the on-camera case, you likely want to do the same for the 'network' case as well
18:19
vup
yeah of course
18:20
se6ast1an
thanks for the update! maybe more to discuss after meeting
18:20
se6ast1an
but to conclude that one, Bertl would you speak the finishing words as usual?
18:21
Bertl
nothing to report from my side for this week
18:21
Bertl
I spent that little time I had on answering emails and reviewing stuff
18:22
Bertl
so if nobody has anything else, we can conclude the meeting
18:23
se6ast1an
thanks!
18:23
se6ast1an
yes MEETING CONCLUDED!
18:23
se6ast1an
many thanks everyone!
18:23
Bertl
thanks for moderating!
18:27
Bertl
off for now ... bbl
18:27
Bertl
changed nick to: Bertl_oO
20:36
se6ast1an
https://lab.apertus.org/ has cert warning, morrigan is on it
21:21
LordVan
joined the channel
22:21
LordVan
left the channel
23:55
daFred
left the channel