Current Server Time: 00:50 (Central Europe)

#apertus IRC Channel Logs

2021/03/17

Timezone: UTC


23:36
comradekingu
left the channel
23:41
mumptai
left the channel
23:52
BAndiT1983
changed nick to: BAndiT1983|away
00:07
lambamansha
joined the channel
01:42
briansun79
joined the channel
01:43
briansun79
left the channel
01:43
briansun
joined the channel
01:54
briansun
left the channel
01:55
lambamansha
left the channel
02:06
lambamansha
joined the channel
03:05
lambamansha
left the channel
03:28
eppisai
joined the channel
04:49
Bertl_oO
off to bed now ... have a good one everyone!
04:49
Bertl_oO
changed nick to: Bertl_zZ
05:03
dcz_
joined the channel
05:35
BAndiT1983|away
changed nick to: BAndiT1983
05:45
vedant16[m]
BAndiT1983: > alternatively create a script based on the current icon creation Makefile (https://github.com/apertus-open-source-cinema/AXIOM-Remote/blob/dev/Firmware/Media/Icons/Makefile) or create a new script (python, etc.) that takes care of this.
05:45
vedant16[m]
* BAndiT1983:
05:45
vedant16[m]
> alternatively create a script based on the current icon creation Makefile (https://github.com/apertus-open-source-cinema/AXIOM-Remote/blob/dev/Firmware/Media/Icons/Makefile) or create a new script (python, etc.) that takes care of this.
05:45
vedant16[m]
Can I add a new recipe in the existing makefile itself ?
05:46
BAndiT1983
hi vedant16[m], of course
05:46
BAndiT1983
by the way, why are there often 2 long messages from you? something wrong with your client?
05:46
vedant16[m]
why isn't there recipe for down_icon, help_icon ?
05:47
vedant16[m]
<BAndiT1983 "by the way, why are there often "> can you send a screenshot, I am using element.
05:47
BAndiT1983
because nobody added it
05:47
BAndiT1983
http://irc.apertus.org/index.php?day=17&month=03&year=2021
05:47
vedant16[m]
ohh, so irc doesn't support editing
05:47
BAndiT1983
no
05:48
vedant16[m]
I edited the message that was sent.
05:48
vedant16[m]
element supports that
05:49
vedant16[m]
<BAndiT1983 "because nobody added it"> should I add it ?
05:50
BAndiT1983
are they required and used somewhere?
05:50
vedant16[m]
not for the task
05:51
BAndiT1983
you can add it, if they are required, otherwise i don't see a necessity at the moment, as the challenge task is about 2bit icons, to smooth them out, so they will be probably replaced when this new feature is done
05:52
vedant16[m]
cool 👌
05:52
vedant16[m]
does your irc render emoji?
05:53
BAndiT1983
yes
06:22
vedant16[m]
the makefile is broken, `convert -density 200 $(TEMP_DIR)/TempRocket.svg -trim +repage -filter point -monochrome -resize 200x200 xbm:- | ./convert_xbm_to_h.sh $@ -`
06:22
vedant16[m]
works
06:22
vedant16[m]
magick apparently isn't shipped as a binary
06:23
vedant16[m]
* changing magick to convert works
06:26
BAndiT1983
it depends on the system one is using, as the makefile was tested i doubt that it's broken
06:26
vedant16[m]
ohh, I am on ubuntu 20.
06:26
BAndiT1983
my system (manjaro) understands both fine
06:27
vedant16[m]
I have installed imagemagick and it cannot find magick binary
06:28
BAndiT1983
as the CI container uses ubuntu, it was building just fine, so maybe something changed in-between
06:39
BAndiT1983
vedant16[m]: looks like magick is the newer call for IM7+ (remember reading some discussion a while ago) -> https://imagemagick.org/script/porting.php
06:40
BAndiT1983
convert is a symbolic link in the newer version, maybe you need to check how old your IM is or maybe ubuntu has some own flavour, that's why i stick to manjaro for a couple of years (linuxmint before)
06:40
BAndiT1983
the summary is under "Command Changes"
06:53
se6ast1an_
good morning
06:55
BAndiT1983
hi
07:01
mumptai
joined the channel
07:07
vedant16[m]
<BAndiT1983 "convert is a symbolic link in th"> I am actually using mint20
07:07
vedant16[m]
<BAndiT1983 "vedant16: looks like magick is t"> I'll update my version
07:08
BAndiT1983
mint is using ubuntu as base, if you're not using debian edition
07:08
se6ast1an_
hi BAndiT1983, great to see so much activity here vedant16[m]!
07:08
se6ast1an_
something completely different as quick distraction in case you didnt see it yet: https://hackaday.io/project/178132-fdm-3d-metal-printer
07:08
se6ast1an_
printing aluminum like a normal FDM printer does
07:08
se6ast1an_
just with a 700°C extruder :)
07:08
vedant16[m]
Hi se6ast1an_ I am working on the qualifying statement :)
07:09
se6ast1an_
yes, I figured
07:09
BAndiT1983
interesting, maybe i can modify my prusa ;)
07:09
vedant16[m]
<BAndiT1983 "mint is using ubuntu as base, if"> yup, that's why said ubuntu initially
07:09
vedant16[m]
<se6ast1an_ "printing aluminum like a normal "> woahh, this is nice
07:14
vedant16[m]
`DrawPixel(x + xIndex, yPos, color);`
07:14
vedant16[m]
in this function, color means the color of the pixel to be displayed, so what color codes should I use for the 4 colors in 2-bit pixels ?
07:15
vedant16[m]
nvm, figured
07:15
vedant16[m]
transparent, light grey, dark grey, black. I'll use ascii codes for these
07:16
BAndiT1983
why ascii codes?
07:17
vedant16[m]
just found utils.h, I'll use RGB565 function
07:18
vedant16[m]
<BAndiT1983 "why ascii codes?"> I meant ascii representation of color codes
07:18
BAndiT1983
please elaborate
07:37
vedant16[m]
https://www.codegrepper.com/code-examples/actionscript/ascii+color+codes+python
07:39
BAndiT1983
yes, i know that, but how should it help in the firmware?
07:40
vedant16[m]
Ah, no it doesn't
07:40
vedant16[m]
<vedant16[m] "just found utils.h, I'll use RGB"> I was just being verbose, and then I found this function
07:51
BAndiT1983
what's the most important topic there is, how to properly convert the images to 2bit depth and in the suitable format, manual conversion through online tool is applicable for first tests, but later we want to automate that, so another way would be a python script which prepares the data and stores it in an array and also has a header, e.g. with width, height, data and format (1bit / 2bit depth)
07:51
BAndiT1983
haven't found suitable conversion in IM, as the .h one is rather rudimentary
08:15
RexOrCine1
left the channel
08:33
vedant16[m]
<BAndiT1983 "what's the most important topic "> Yup, imagemagick one doesn't work. I'll use a .h for now
08:35
BAndiT1983
IM can also convert to header or .c, but the format is plain or so
08:36
vedant16[m]
something very strange
08:36
vedant16[m]
uploaded an image: (496KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/CWZVrQAxFCtzopgDNLYWvjRx/Screenshot%20from%202021-03-17%2015-06-01.png >
08:36
vedant16[m]
see the extreme right column ?
08:36
vedant16[m]
this is an array of the following image
08:36
vedant16[m]
uploaded an image: (20KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/TMsDGzFIRAHxPpipxGfIKdHy/jEsCMABc_400x400.jpg >
08:37
vedant16[m]
zoomed out view of the ascii code actually renders the image
08:37
vedant16[m]
you can see a diamond there
08:42
eppisai
left the channel
08:42
BAndiT1983
yes, it's normal, as values change for every pixel, so you can see the image in the file map, what is your trouble regarding that?
08:43
vedant16[m]
<BAndiT1983 "yes, it's normal, as values chan"> no trouble, just a TIL moment
08:44
eppisai
joined the channel
09:07
RexOrCine
joined the channel
09:24
vedant16[m]
uploaded an image: (1085KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/ldABUuFMCUmsMRmGqbeoSMSd/Screenshot%20from%202021-03-17%2015-54-11.png >
09:24
vedant16[m]
Image pixels seem circularly shifted, try to figure out the issue
09:31
se6ast1an_
but the general image shapes can already be easily seen
09:31
se6ast1an_
very good
09:33
BAndiT1983
we just need a proper compression of 4 pixels into 1 byte, 2 bit depth, so the expected values are from 0 to 3, where 0 is transparent and 3 opaque
09:33
BAndiT1983
would suggest to grab smaller picture with few colors or black and white to be able to check the values manually
09:43
vedant16[m]
<BAndiT1983 "would suggest to grab smaller pi"> ohk, will do that
10:08
Hp_Dopeman
joined the channel
11:18
Hanupratap
joined the channel
11:19
Hanupratap
changed nick to: dopeman
11:21
dopeman
left the channel
11:22
Hp_Dopeman2
joined the channel
11:22
Hp_Dopeman2
left the channel
11:27
Hp_Dopeman
left the channel
11:41
Bertl_zZ
changed nick to: Bertl
11:41
Bertl
morning folks!
11:43
se6ast1an_
good day!
11:43
vedant16[m]
good morning :)
11:46
BAndiT1983
vedant16[m]: in fact it's almost good evening here ;) just joking ist 13:46 (CET)
11:47
BAndiT1983
*it's
12:23
lexano
left the channel
12:26
lexano
joined the channel
13:44
mohitdmak
joined the channel
13:47
lambamansha
joined the channel
14:13
mohitdmak
left the channel
14:38
markuse
joined the channel
14:58
mohitdmak
joined the channel
15:00
pangelo[m]
left the channel
15:00
vedant16[m]
<BAndiT1983 "vedant16: in fact it's almost go"> It was actual evening here (IST) when I replied, it was 6pm xD
15:00
mohitdmak
@vup The network manager is working properly now, and I had made some UI changes using the dummy data you had sent. So should I make a PR for task 1 or both altogether?
15:00
BAndiT1983
vedant16[m]: i know, we had many students from India, so learned to take the time shift into account
15:02
vedant16[m]
CET is still okay
15:02
mohitdmak
Also vup can you brief a little on how the current load of the camera can be made accessible to the frontend for the task 2?
15:02
vedant16[m]
it becomes awful to work when mentors are in PST lol
15:05
BAndiT1983
mohitdmak: do you mean processor load?
15:06
mohitdmak
BAndiT1983 well the task demands a graph component to be created based on the instantaneous load of the camera..
15:07
BAndiT1983
found around 6-10 ways how to get the load in the terminal, as far as i've understood you're already grabbing the wifi networks from NetwrokManager, so it should work in similar way
15:12
mohitdmak
Hmm I will look around the code then. As for the network manager, after booting linux I am able to access the wifi networks and have made some nice UI changes on it.
15:12
mohitdmak
Thanks though :)
15:12
mohitdmak
left the channel
15:12
BAndiT1983
you should look around linux first, but that's my opinion
15:15
BAndiT1983
and by the way, nice UI is just one part of a software, stable base, clean code and thought-out structure are more important factors, speaking out of experience
15:23
lambamansha
Hi, I have forked&cloned the repository and able to run the project locally in my system. Also, understood the flow of the project. I am focusing on task 1 of the challenge ( fancy wifi list ). So what my understanding is that currently simple HTML table is being used , you want it to be replaced with more beautiful table may be material-table which has the sorting functionality based on different keys. Wifinetworks key in state contains the data
15:23
lambamansha
that is to be fed in table.
15:23
lambamansha
Please correct me if I am wrong.
15:23
se6ast1an_
and when you say "nice GUI" we want to see screenshots, very curious!
15:25
lambamansha
@vup is my understanding correct ?
15:27
BAndiT1983
sorting by strongest signal is the most common case, alphabetical maybe too, but seldom seen that
15:28
BAndiT1983
also one should think about which side of the communication, host or client, needs to do what in the process
16:07
pangelo[m]
joined the channel
16:12
lambamansha
left the channel
16:39
markuse
left the channel
17:32
markus_
joined the channel
17:37
Bertl
off for now ... bbl
17:37
Bertl
changed nick to: Bertl_oO
17:37
vedant16[m]
bye
17:37
vedant16[m]
https://github.com/VedantParanjape/AXIOM-Remote/commit/a66c637e201ecbdab801a1727e5dee42ceec3b1f
17:37
vedant16[m]
BAndiT1983: does the implementation look good ?
17:37
vedant16[m]
I still can't figure why image is getting cyclically rotated.
17:39
BAndiT1983
what do you mean by cyclically rotated?
17:39
BAndiT1983
which coding guideline are you following there? -> color_map
17:40
vedant16[m]
<vedant16[m] "Screenshot from 2021-03-17 15-54"> see here
17:40
BAndiT1983
how is the image behaving on different backgrounds? e.g. balck, white, gray?
17:40
vedant16[m]
<vedant16[m] "jEsCMABc_400x400.jpg"> this is my image
17:40
vedant16[m]
<BAndiT1983 "how is the image behaving on dif"> not tested yet. lemme do that
17:41
BAndiT1983
seen the image, but still don't understand what you mean by "cyclically rotated"
17:41
vedant16[m]
umm, the right hand is getting displayed on the left
17:43
BAndiT1983
you mean flipped on the y-axis?
17:44
vedant16[m]
<BAndiT1983 "which coding guideline are you f"> sorry, updated
17:46
vedant16[m]
<BAndiT1983 "you mean flipped on the y-axis?"> ugh, no imagine a image to be array. and you just rotated the array in the right direction
17:46
se6ast1an_
vedant16[m]: The drawing origin (X,Y = 0,0) is located in the top left corner, do you also read the image pixels from top left?
17:46
vedant16[m]
or circularly shifted bits (https://stackoverflow.com/questions/13289397/circular-shift-in-c_
17:48
vedant16[m]
I read sequentially, used lvgl.io's image to .h convertor
17:48
vedant16[m]
does it use some other format ?
17:49
vedant16[m]
like starting of array is the right most pixel
17:50
vedant16[m]
oh, wait. I think I assumed it wrong how the 2-bit pixels are stored.
17:51
BAndiT1983
as we haven't written the converter, we don't know, it was listed as example and starting point, our overal goals ragarding that topic were listed by me today before
17:52
markus_
Hi there, I had this problem earlier today when I implemented it
17:52
markus_
That converter doesn't work properly
17:52
markus_
it stores the pixels in wrong order
17:52
vedant16[m]
<markus_ "That converter doesn't work prop"> ohhh
17:52
vedant16[m]
<BAndiT1983 "as we haven't written the conver"> so, does the solution satisfy it ?
17:53
markus_
I wrote my own in Python instead, I can link to my repo in a second
17:53
vedant16[m]
<markus_ "I wrote my own in Python instead"> ohh, wait. don't. I want to try myself :)
17:53
BAndiT1983
like i said, the converter was listed as reference, but we expect that students provide some solution which fits the task and is maybe suitable for automation, like calling from the makefile
17:54
vedant16[m]
<BAndiT1983 "like i said, the converter was l"> ohk, makes sense
17:55
markus_
I had a question regarding coding style.
17:55
se6ast1an_
hi markus_!
17:55
markus_
Since there was a link to the isocpp faq in the challenge
17:55
markus_
but I don't see all that much usage of the STL
17:55
markus_
so should I prefer using std::array or just plain C-arrays?
17:55
markus_
hi sebastian!
17:56
BAndiT1983
you can see the style in the firmware code, but it should be mainly C++ style, so pixelData and not pixel_data
17:57
BAndiT1983
also the guildeines are in the remote repo -> https://github.com/apertus-open-source-cinema/AXIOM-Remote/blob/dev/CodingGuidelines.md
18:01
BAndiT1983
markus_: generally you should keep in mind, that it's a pic32 MCU, so C++11 or similar, are not that much possible, but to some extent
18:01
BAndiT1983
got some crashes actually, as we had some memory hogs in the code, so the board started to crash
18:03
markus_
alright I see
18:03
BAndiT1983
and to bore people a bit more regarding pic32, the compile, XC32, is based on some old GCC version
18:03
se6ast1an_
thanks for testing and verifying the online converter doesnt work, I removed it from the challenge description
18:04
BAndiT1983
it probably works for LVGL library
18:04
se6ast1an_
probably, but not for us :)
18:04
BAndiT1983
but a nice python script should be possible to do, with 0-3 range etc., done that already to check the difficulty for students
18:05
BAndiT1983
of course, as i can't automate that in the CI build
18:05
markus_
but I imagine std::array should be fine in terms of memory right? Since the size will be known at compile time
18:05
vedant16[m]
https://stackoverflow.com/questions/8707183/script-tool-to-convert-file-to-c-c-source-code-array
18:05
markus_
so no crashes at least
18:05
vedant16[m]
xxd does a great job
18:06
vedant16[m]
directly outputs image as an array.
18:06
BAndiT1983
markus_: you can test it with XC32, it shows nice memory summary for the build, also cutter/radare2 can help to disassemble the resulting ELF file, to check the sizes and sections
18:06
BAndiT1983
vedant16[m]: how do you want to convert images with it?
18:08
vedant16[m]
<BAndiT1983 "vedant16: how do you want to con"> image magic -> xxd -> sed
18:08
vedant16[m]
I will comeup with a script
18:09
BAndiT1983
alright, looking forward to it, but don't forget that it needs also width and height info, maybe also format info like 1 bit or 2 bits
18:10
vedant16[m]
yup, will keep this in mind
18:13
BAndiT1983
python with PIL and templating was not that difficult also, have created a template file with placeholders, python can substitute them easily, just as a tip
18:13
shantanu
joined the channel
18:14
vedant16[m]
I was thinking of a very minimal solution
18:16
BAndiT1983
python is not a problem for the CI build, example: lets say someone pushes a new image into the repo, the CI build start and the makefile recognizes that something new is there, it start python, which is/would be included in the container, and processes the SVG/PNG file, afterwards the resulting .h file is committed automatically to the repo and can be used by other people
18:17
BAndiT1983
manual way is also possible, am just having quite some tasks at work regarding CI/CD, so automation is my focus
18:17
BAndiT1983
off for a moment
18:17
vedant16[m]
ohk, then I'll go forward with python
18:17
BAndiT1983
changed nick to: BAndiT1983|away
18:18
shantanu
left the channel
18:18
eppisai
left the channel
18:19
eppisai
joined the channel
18:20
eppisai56
joined the channel
18:21
eppisai
left the channel
18:21
eppisai56
left the channel
18:21
eppisai
joined the channel
18:59
Hp_Dopeman
joined the channel
19:05
dcz_
left the channel
19:08
Hp_Dopeman83
joined the channel
19:08
BAndiT1983|away
changed nick to: BAndiT1983
19:09
BAndiT1983
back
19:09
Hp_Dopeman
left the channel
19:09
Hp_Dopeman83
left the channel
19:09
BAndiT1983
vedant16[m]: didn't want to stop you from your plan, am just giving advices regarding possibilities
19:09
Hp_Dopeman
joined the channel
19:10
BAndiT1983
as we have some python for the remote programmer and in the future probably a mock service to mimic the camera, so it's one of possible things to use
19:10
eppisai
left the channel
19:17
dopeman_
joined the channel
19:18
Hp_Dopeman
left the channel
19:18
dopeman_
left the channel
19:19
Hp_Dopeman
joined the channel
19:23
Hp_Dopeman96
joined the channel
19:23
Hp_Dopeman
left the channel
19:23
Hp_Dopeman96
changed nick to: Hp_Dopeman
19:26
Hp_Dopeman
left the channel
19:26
dopeman_
joined the channel
19:26
dopeman_
left the channel
19:29
eppisai
joined the channel
19:34
dopeman
joined the channel
19:35
dopeman
left the channel
19:36
Hp_Dopeman
joined the channel
19:36
Hp_Dopeman
left the channel
19:40
markus_
<BAndiT1983> So i tested with and without std::array
19:41
markus_
this is only for passing in the 4 colors to be used btw
19:41
markus_
and it made an 84 byte difference actually
19:41
markus_
in the favor of c-arrays and pointers
19:44
markus_
which seems like way too much considering the fact that it's an array of 4 uint16_t
19:47
eppisai10
joined the channel
19:48
eppisai10
left the channel
19:49
eppisai
left the channel
19:56
AMANSINGH[m]
joined the channel
19:56
AMANSINGH[m]
changed nick to: eppisai[m]
20:02
BAndiT1983
markus_: this is what i've expected, as STL brings a bit more overhead, not that interesting on the desktop, but if it's not optimized for MCUs it makes quite some difference
20:03
BAndiT1983
markus_: which task do you target, by the way, just so we can keep the track of which tasks students want to do, firmware or bootloader?
20:09
markus_
BAndiT1983: I'm targeting the firmware task!
20:10
BAndiT1983
alright, thank you
20:16
markus_
left the channel
20:16
markus_
joined the channel
20:19
BAndiT1983
markus_: you can also use 1 color, the task is about creating proper shades to nicely blend into the background, so the icon edges are not jagged, that's why i was saying that 0 as piel value is transparent and 3 is opaque, in-between 2 shades which take background pixel into account
20:23
BAndiT1983
but you can also use the 4 values, although just 2 would be needed there, for calculations, haven't done any prototype yet, usually i create one for gsoc to compare against the student challenge code and often learn new things and ways
20:37
markus_
BAndiT1983 : okay I see, I had missed the part about blending in the background. Thank you!
20:38
BAndiT1983
no problem, just mentioning it, the goal is also important, but the challenge is mostly about how students integrate and collaborate in the project, also the code style is bigger factor
20:43
eppisai
joined the channel
20:53
aman
joined the channel
20:56
aman
left the channel
20:56
eppisai_
joined the channel
21:00
eppisai_
left the channel
21:01
eppisai_
joined the channel
21:04
vup
mohitdmak: if you are reading the log: the load of linux system can be determined by various ways, for example by parsing the output of the `uptime` command or reading the `/proc/loadavg` file. On how to get this data into the frontend, please take inspiration by how for example the NetworkManager data is read out.
21:06
vup
lambamansha: If you read the log: Yes there are several ways the wifi display could be improved for example making the table more beautiful, by using the material-ui table, highlighting the current IN-USE wifi network or showing the SECURITY values using chips. Be creative :)
21:07
eppisai_
left the channel
21:07
eppisai_
joined the channel
21:10
eppisai[m]
left the channel
21:10
eppisai_
left the channel
21:10
eppisai_
joined the channel
21:11
eppisai
left the channel
21:11
eppisai_
left the channel
21:11
eppisai_
joined the channel
21:12
eppisai_
left the channel
21:15
eppisai_
joined the channel
21:17
eppisai_
left the channel
21:17
eppisai_
joined the channel
21:21
vup
eppisai_: hi, is quassel working for you so far?
21:23
eppisai_
Hi,yea its like a charm,I had logged in from two devices (mobile and laptop) so just loging out of phone to have just one account..
21:23
eppisai_
*logging
21:24
vup
awesome, btw there is also a quassel app for android (and for ios aswell I think)
21:24
eppisai_
oh..! thats awesome! I'll download it. thanks! :)
21:35
BAndiT1983
changed nick to: BAndiT1983|away
22:36
mumptai
left the channel