Current Server Time: 04:26 (Central Europe)

#apertus IRC Channel Logs

2019/01/25

Timezone: UTC


00:54
siddhantsahu
joined the channel
00:54
siddhantsahu
left the channel
03:04
futarisIRCcloud
joined the channel
05:15
futarisIRCcloud
left the channel
06:01
BAndiT1983|away
changed nick to: BAndiT1983
06:04
BAndiT1983
vup2, this was done in the early days of prototype and kept, haven't investigated this area further as we had other things to foucs on
06:25
futarisIRCcloud
joined the channel
06:57
Bertl
off to bed now ... have a good one everyone!
06:57
Bertl
changed nick to: Bertl_zZ
07:12
BAndiT1983
changed nick to: BAndiT1983|away
07:17
BAndiT1983|away
changed nick to: BAndiT1983
07:53
se6astian|away
changed nick to: se6astian
08:16
LordVan
joined the channel
08:35
futarisIRCcloud
left the channel
08:57
se6astian
changed nick to: se6astian|away
08:58
se6astian|away
changed nick to: se6astian
08:58
se6astian
left the channel
08:58
se6astian|away
joined the channel
08:58
se6astian|away
changed nick to: se6astian
10:35
futarisIRCcloud
joined the channel
11:43
vup2
BAndiT1983: ok so possibly removing the systemd specific code should be no problem
11:43
BAndiT1983
if the functionality will not change, then yes
11:44
BAndiT1983
http://0pointer.de/blog/projects/journal-submit.html
11:44
BAndiT1983
this was one of tutorials i've used while learning about systemd daemons
11:45
BAndiT1983
if you have examples about your proposed way, then i wouldl ike to read more into it, as most systemd docs are wall of texts
12:12
dan_
joined the channel
12:13
dan_
left the channel
12:41
LordVan
left the channel
12:54
futarisIRCcloud
left the channel
13:19
futarisIRCcloud
joined the channel
14:23
aombk2
left the channel
15:29
BAndiT1983
changed nick to: BAndiT1983|away
15:47
siddhantsahu
joined the channel
15:52
se6astian
changed nick to: se6astian|away
15:52
siddhantsahu
left the channel
16:18
BAndiT1983|away
changed nick to: BAndiT1983
16:34
futarisIRCcloud
left the channel
17:02
RexOrCine|away
changed nick to: RexOrCine
18:01
se6astian|away
changed nick to: se6astian
18:15
se6astian
BAndiT1983: the general goal is to have all functions from *.c files also in the corresponding *.h file I assume correct?
18:42
se6astian
changed nick to: se6astian|away
18:42
BAndiT1983
not all, just the ones reuqired by other modules
18:45
se6astian|away
changed nick to: se6astian
18:46
BAndiT1983
found what patch-1 is, probably a branch created in my fork by gitkraken
18:47
se6astian
ah right
18:47
se6astian
so are you available now?
18:47
se6astian
should we do the merge?
18:48
BAndiT1983
yep, am on it, so my repo is conflict free first
18:48
BAndiT1983
please check the underscores in front of vars, like _main_menu
18:49
BAndiT1983
my long-time favorite merge and diff tool is p4merge, am glad that it also works under linux
18:52
se6astian
will work on underscores now, good idea
18:54
BAndiT1983
you can leaven underscores for fonts there for now, till we change the names
18:59
BAndiT1983
se6astian, let us create some basic code guidelines, this would prevent most merge conflicts, e.g. avoid underscores in front of vars, braces on new line, C lower-case naming for methods and vars like set_white_balance()
19:01
se6astian
removed underscores for fonts already
19:02
se6astian
why leave them?
19:02
se6astian
lets add the rules to readme?
19:03
BAndiT1983
very good, yes, let's place them there, but we should move it to wiki later
19:06
BAndiT1983
so, merge done, now i have to adjust the files manually, as it's very cumbersome to overview all the conflicts
19:08
se6astian
braces on new line - I dont do that
19:10
BAndiT1983
still don't understand how people tend to place braces on same line, luckily my company doesn't follow this for java code
19:16
BAndiT1983
is E1_rot now E1_pos?
19:17
se6astian
fine for me
19:19
BAndiT1983
the change is not mine, i suppose its yours, that's why i'm asking what we should use, as the code is using E1_rot, but globals.h contains E1_pos
19:20
BAndiT1983
will change to E1_pos where it complains, according to globals.h definition, can be renamed later, jsut want to get it to buildable state
19:23
se6astian
must have been a mix up then, both names are descriptive
19:24
BAndiT1983
oh wow, i can see menu switching with buttons 7 and 8
19:25
BAndiT1983
will commit my merge, afterwards you can check it and start to complain ;)
19:25
BAndiT1983
what were your action about imgui submodule? as it is in conflict state, but i cannot recognize why
19:28
se6astian
I manually added it
19:28
se6astian
git submodule add ....
19:28
BAndiT1983
ah, probably gitmodules is conflicting, as it was already in my repo
19:32
se6astian
I guess you can just remove it and merge
19:33
BAndiT1983
seems to be merged now, will check what the pull requests against master repo will show
19:34
BAndiT1983
ok, they are even at the moment, please check it also from master
19:34
BAndiT1983
11 commits are in my fork
19:35
BAndiT1983
i mean 11 commits which can be pulled
19:37
se6astian
great, lets do it
19:41
se6astian
done
19:43
BAndiT1983
could you check if it switches the menus now?
19:47
se6astian
will do
20:08
se6astian
merging my own changes now
20:08
se6astian
well soon
20:37
se6astian
BAndiT1983:
20:37
se6astian
you here?
20:56
Bertl_zZ
changed nick to: Bertl
21:34
se6astian
about to leave now, current code doesnt compile
21:35
se6astian
may I suggest you install mplabx so you can also verify this
21:35
se6astian
the refactoring tools are also quite good, currently there are a few variables that got renamed in the wrong way
21:35
se6astian
will continue tomorrow
21:45
se6astian
changed nick to: se6astian|away
21:55
BAndiT1983
se6astian|away, will check it
23:11
vup2
For c++ / formatting i can recommend clang-format, takes a little effort to setup, but (especially when calling it from a commit hook) it help tremendously with consistent c++ / c formating with almost no effort (after setting it up)
23:11
BAndiT1983
se6astian|away, please paste the build command from your IDE
23:12
vup2
I think it would be useful to include a clang-format config file in the control daemon and also the remote
23:13
vup2
That will also probably help the gsoc students
23:16
BAndiT1983
we can think of something in that direction, but we need some unified way to execute formatting
23:19
vup2
BAndiT1983: regarding the systemd changes: i have unfortunately not found any simple tutorial / example, i am just extrapolating from the way inetd works and StandardInput=socket being the option that should make the systemd way work with inetd
23:22
BAndiT1983
have to check that topic again, as i know that systemd supplies the socket through file descriptor, there is a part in the source for that, but it wasn't tested thoroughly yet
23:22
BAndiT1983
i can disable journal functions and check if stdout will land in journalctl
23:23
BAndiT1983
or will just check inetd sources first
23:23
BAndiT1983
http://0pointer.de/blog/projects/inetd.html
23:25
vup2
For the formatting i think a pre commit hook would be the best, a possible way to share these is described here: https://www.viget.com/articles/two-ways-to-share-git-hooks-with-your-team/
23:25
BAndiT1983
this is also helpful (adding as reference for later) -> https://unix.stackexchange.com/questions/313513/detect-if-program-should-log-to-systemd-journal-or-stdout
23:26
BAndiT1983
already collided with commit hooks at work ;) but they can be useful if used correctly
23:27
BAndiT1983
but i doubt there is a possibility to run clang-format on github, maybe before CI and on successful build upload formatted version, if there are changes
23:28
vup2
The current systemd service has the line 'StandardError=journal' which should mean stderr gets logged to the systemd journal
23:29
vup2
Yeah running clang-format in ci and commiting a formatted version from the ci should be possible, but probably unnecessary if we find a way to reliably share the git hooks
23:30
BAndiT1983
git hooks involve a server which can run tools, or am i wrong
23:30
BAndiT1983
systemd logging to journal is fine, but maybe we should go the way of outputting through systemd to log files, one for acces and one for errors, like nginx
23:31
BAndiT1983
as the info about received commands and values is rather important for bug fixing and finding errors
23:32
vup2
oh sorry that wasn't clear on my part, is wasn't talking about something like github webhooks, but about the local git hooks, git support running a local file on some actions like before a commit by just placing it unter .git/hooks/pre-commit
23:33
vup2
i am not sure if systemd provides a way to log to a external file, but doing it on our own doesn't seem that hard
23:34
BAndiT1983
ah, you mean while adding a commit locally, it would also work, although i prefer to execute formatting often manually on the whole directory, at least in intellij i use the shortcut very often
23:34
BAndiT1983
https://stackoverflow.com/questions/37585758/how-to-redirect-output-of-systemd-service-to-a-file
23:35
vup2
oh i also prefer running the formatting myself, but i think the git hooks would help other people (and also prevent someone from forgetting to run the formatting)
23:37
BAndiT1983
also correct
23:37
vup2
ok so rsyslog seems to have systemd integration, interesting
23:38
BAndiT1983
lets discuss formatting hooks it with se6astian|away, when he is online again, but this would definitely be an improvement
23:39
vup2
but i mainly brought up the topic of systemd because i want to get rid of any systemd dependency :)
23:39
BAndiT1983
archlinux uses it, that's why it was chosen
23:39
vup2
yes, but we are thinking of moving away from archlinux
23:40
BAndiT1983
if we can start the daemon through different mechanism, also no objections from my side
23:40
BAndiT1983
really?
23:40
vup2
yes
23:40
BAndiT1983
my only problem with it, is the constant update flood, rather problematic when developing and building on camera
23:40
BAndiT1983
is it a plan for micro or also normal one?
23:41
vup2
while i like archlinux, archlinuxarm seems to regularly have strange problems and is really that well maintained
23:41
BAndiT1983
what would be the alternative distro?
23:42
vup2
if there are no objections from anybody i think it would also make sense for the beta
23:42
vup2
we liked void the best
23:42
BAndiT1983
as long as the scripts and apps are working
23:43
vup2
the scripts and everything should continue to work
23:44
BAndiT1983
found docker container, if i will find some time, then i would test daemon build there
23:44
BAndiT1983
at the moment i'm battling the build scripts of xc32 compiler, docker container is ok, but their make scripts are totally off
23:44
vup2
that is one reason i think void is better than for example alpine, void still uses glibc as opposed to musl with alpine uses, with could couse some troubles
23:45
vup2
nice, we probably will build a test image with void this weekend to test if all the other stuff is working
23:45
BAndiT1983
using alpine mostly for docker or virtualbox, till now everything worked, but i can't say anything about bare metal support
23:46
vup2
well using musl as my main libc i can tell you if you a lot of stuff needs patches to work with musl
23:46
vup2
s/if you//
23:47
BAndiT1983
to be honest, not that deep there yet in musl / libc holy war topic
23:47
vup2
apline has most of the stuff patched, but the even the control daemon does not compile on musl at the moment (because flatbuffers doesn't compile)
23:47
BAndiT1983
but hope to push the progress while my vacation in february, away from web development in java
23:47
vup2
yeah i also have nothing againts glibc, but i thought it would be fun to test musl a bit
23:48
BAndiT1983
ah, right, flatbuffers was the problem last time
23:48
vup2
yeah i can image more interesting / fun things than java web development :)
23:48
BAndiT1983
https://pkgs.alpinelinux.org/package/edge/testing/armhf/flatbuffers
23:50
BAndiT1983
flatbuffers is also included in void packages -> https://voidlinux.org/packages/
23:50
BAndiT1983
nice, was caring it in submodules, as it's not always available in distros
23:53
RexOrCine
changed nick to: RexOrCine|away
23:53
vup2
yes the changes that broke flatbuffers on musl are newer that their latest release
23:53
vup2
(which is what alpine uses)
23:54
BAndiT1983
i know, but alpine edge is not bad for slim docker containers and dev environments in virtualbox
23:54
BAndiT1983
at least for most tasks
23:55
vup2
while i am also a fan of using submodules for dependencies it think it could be useful to check for a system installation first and if that is not found fall back to using the submodule one
23:56
vup2
yes i also really like alpine, but i think using something which uses glibc could save some unnecessary troubles
23:56
BAndiT1983
of course, also no objections there
23:57
BAndiT1983
good point about distro packages, will check how to use cmake to do the job for daemon
23:57
BAndiT1983
nevertheless, off for today, good night, see you
23:57
BAndiT1983
changed nick to: BAndiT1983|away
23:59
vup2
yeah heading off for today as well, good night everyone