but do you think that we should make minimal changes to import_db.sh, or start copying scripts from musicbrainz-docker to musicbrainz-server/docker (and modify them as needed)?
2020-10-27 30155, 2020
_lucifer
the import_db.sh and create_db.sh files
2020-10-27 30105, 2020
alastairp
yvanzo: yes, I saw that. great.
2020-10-27 30102, 2020
yvanzo
imho, the easiest is just to build a custom image on top of musicbrainz-test-database in CB docker-compose.dev.yml as a first step.
2020-10-27 30114, 2020
_lucifer
i am open to doing it both ways. if the long term goal is to merge musicrainz docker into musizrainz-server, then i suggest we go with latter alastairp
2020-10-27 30153, 2020
alastairp
when we first set this up, we intentionally tried to not build an additional image, in order to prevent duplication
2020-10-27 30113, 2020
alastairp
but for our initial testing perhaps we could do that
it should be a small temporary duplication until musicbrainz-server get fixed
2020-10-27 30158, 2020
ruaok
outsidecontext: the apple mac developer distribtion certificate will expire in a month. when you're around for a moment, lets coordinate getting that updated, please.
2020-10-27 30102, 2020
pristine___
Mr_Monkey: read your comment on the PR. Nice. Next step should be to use regular and solid icons for hover/ non hover.
2020-10-27 30150, 2020
outsidecontext
ruaok: oh, it's that time of the year again :( yes, let's do this. I would be around. I can't exactly remember what we need to do, but let's see
2020-10-27 30122, 2020
ruaok
I was about to run off for lunch -- will you be around in a while?
2020-10-27 30143, 2020
outsidecontext
yes, I'll be here until 4 CET
2020-10-27 30136, 2020
ruaok
perfect.
2020-10-27 30155, 2020
outsidecontext
the important part is to get the certificate renewed and get the certificate files. Setting this up for the Picard builds I can do later
2020-10-27 30156, 2020
Mr_Monkey
pristine___: Yes, there's a bit of work left for styling various things, but it's getting there !
2020-10-27 30118, 2020
CatQuest
[12:59] <yvanzo> Sample data is gathered around three artists: Nine Inch Nails, Kanye West, and Пётр Ильич Чайковский.
2020-10-27 30119, 2020
CatQuest
tsk, no Motörhead? ;)
2020-10-27 30111, 2020
reosarevok
Well, IIRC I suggested those and I like NIN more than Motörhead :p
2020-10-27 30141, 2020
pristine___
Mr_Monkey: so if I use regular emoticons, the only peice of code that will be removed is `stroke`? We will still have to set colors, don't size and stuff, so do we want to really use regular emoticons given that we will have to manage hover on/off state? I mean what is the wise choice here? Wanted to discuss before moving on
2020-10-27 30146, 2020
CatQuest
tsk
2020-10-27 30101, 2020
supersandro2000 has quit
2020-10-27 30134, 2020
supersandro2000 joined the channel
2020-10-27 30139, 2020
Mr_Monkey
pristine___: For me, the FA icons styled with CSS stroke/color don't work, that's my main concern. Sometimes they just end up as a blob (image below), whereas FA already has separate icons styles that allow us to show line/filled up icons.
And these are just defined with `color: #whatever`, no stroke at all.
2020-10-27 30137, 2020
pristine___
Right!
2020-10-27 30107, 2020
Mr_Monkey
Now there are two ways to implement switching from regular to solid icons: 1. in the react component, render either one icon or the other depending on your condition or 2. Render both of them and use CSS to hide one or the other depending on your condition (simply applying a specific css class name that will set `display:none` appropriately)
2020-10-27 30135, 2020
Mr_Monkey
I kiiiind of suspect that I had problem with method 1. before, where there was some flickering/rendering issues, but I think method 2 is foolproof, with the disadvantage that we'll be creating twice as many icon components on the page. That being said, I haven't tried to measure if that would be a performance issue, but I suspect not.
outsidecontext: ok, ready to do the cert dance. can you please generate a Certificate Signing Request (CSR)?
2020-10-27 30118, 2020
outsidecontext
ruaok: hey, sorry. I have to finish something up here and leave afterwards :( can we do it another day? I can prepare the CSR anyway, need to check what info we had to set. It likely needs to match what we have set in the dev center
2020-10-27 30159, 2020
ruaok
yeah, no problem. generate them and them mail them to me when you get the chance and I'll take it from there.
2020-10-27 30129, 2020
outsidecontext
ok, let's do that
2020-10-27 30132, 2020
ruaok
alastairp: you about?
2020-10-27 30108, 2020
alastairp
hi ruaok
2020-10-27 30116, 2020
ruaok
hey.
2020-10-27 30129, 2020
ruaok
just wondering how extensive the logging support in troi should be.
2020-10-27 30106, 2020
ruaok
right now I am planning for a -d flag to enable debugging, which sets the log level to debug.
2020-10-27 30119, 2020
ruaok
normal debug level could be info, maybe?
2020-10-27 30127, 2020
alastairp
sounds good to me
2020-10-27 30155, 2020
alastairp
a nice-to-have, but not urgent would be that the name of the class where the logger is called from appears in the log message
2020-10-27 30156, 2020
ruaok
then the patch and element objects get a log(self, msg) and debug(self, msg) function that logs messages.
2020-10-27 30111, 2020
ruaok
log always shows, debug only when -d is given on the command line.
2020-10-27 30129, 2020
ruaok
name of class -- yes, on my radara.
2020-10-27 30138, 2020
alastairp
yeah, I think that's fine. Is it worth just exposing the logging object? `self.logger.info()`, or whatever
2020-10-27 30111, 2020
ruaok
is it possible to prepend the class name if we just expose the logger object ?
2020-10-27 30106, 2020
alastairp
yeah, unsure about that one. I was looking at things like %(funcname) and %(modulename) in the logging formatting, but neither of those is the class
2020-10-27 30128, 2020
alastairp
one option would be to set the name of the logger in each class to `troi.blah.Class`
2020-10-27 30135, 2020
ruaok
I'm happy to wrap the functions in questions and place them in Patch and Element.
2020-10-27 30159, 2020
ruaok
also, class names or slugs?
2020-10-27 30151, 2020
alastairp
patches have slugs?
2020-10-27 30153, 2020
alastairp
but elements don't
2020-10-27 30118, 2020
ruaok
ah. yes. yes patches have slugs.
2020-10-27 30125, 2020
ruaok
but then lets just go with class names.
2020-10-27 30130, 2020
alastairp
I wouldn't bother adding a slug to Element if it's just as easy to hard-code it in the logger setup/log method
