I did try to create an entry last time with an unset name. it wouldn't sae any other data i set
2016-04-16 10747, 2016
CatQuest
but i could delete it ine as i did earleir
2016-04-16 10731, 2016
Leftmost
CatQuest, you could save with the alias deleted?
2016-04-16 10741, 2016
CatQuest
yes. no name
2016-04-16 10749, 2016
CatQuest
but that was a longer time ago
2016-04-16 10754, 2016
CatQuest
before the current changes
2016-04-16 10727, 2016
Leftmost
Oh, are you saying you can no longer do that?
2016-04-16 10738, 2016
CatQuest
I did some other stuff.. I don't remember.. I was not paying attention. I'll try this later on
2016-04-16 10754, 2016
CatQuest
i havent tried it again :)
2016-04-16 10756, 2016
CatQuest
i will
2016-04-16 10758, 2016
Leftmost
Ahh, okay.
2016-04-16 10712, 2016
Leftmost
Okay, rika is started with the changes I made.
2016-04-16 10729, 2016
CatQuest
also with things like a fullwith space or messed up character instead of empthy
2016-04-16 10731, 2016
LordSputnik
Great
2016-04-16 10732, 2016
CatQuest
things...
2016-04-16 10743, 2016
LordSputnik
So, can we just ignore any identifiers which have no value?
2016-04-16 10706, 2016
Leftmost
LordSputnik, that's what I'm doing in the client-side code, so that should be fine server-side as well.
2016-04-16 10741, 2016
CatQuest
hey, wil lyo uguys still be here in 4-5 hours time?
2016-04-16 10748, 2016
LordSputnik
CatQuest: yeah probably :)
2016-04-16 10700, 2016
CatQuest
great
2016-04-16 10701, 2016
Leftmost
I won't be fully here, but I'll be around-ish.
2016-04-16 10714, 2016
CatQuest
you'llbe.. leftmost?
2016-04-16 10711, 2016
LordSputnik
So why does submitting a deleted valueless identifier cause a new identifier to be created?
2016-04-16 10744, 2016
Leftmost
My areas of concern for testing: do forms prevent the adding of data that breaks our rules? Do we correctly inform the user when they've entered invalid data? Is there data the user can enter that will break form submission? Does the relationship editor behave as expected?
2016-04-16 10746, 2016
adesh joined the channel
2016-04-16 10723, 2016
Leftmost
LordSputnik, I'm not sure. It doesn't actually succeed, as type_id isn't NULLable. It 500s trying to add it to the database.
2016-04-16 10744, 2016
JesseW joined the channel
2016-04-16 10751, 2016
LordSputnik
But if it's removed with the button, it shouldn't be part of the state, and shouldn't be submitted in the form
2016-04-16 10753, 2016
adesh
does brainzbot automatically runs test on jenkins whenever i add any commit to PR?
2016-04-16 10700, 2016
Leftmost
Aha, I'm misusing the interface (which is a bug in itself): I'm clearing the identifier type, but not removing the identifier.
2016-04-16 10746, 2016
LordSputnik
Ahh
2016-04-16 10759, 2016
Leftmost
It lets me submit the form, though.
2016-04-16 10704, 2016
LordSputnik
So we need to do some validation to make sure there are no unexpected blank identifiers
2016-04-16 10732, 2016
Leftmost
We should, but we should also not allow blanking the identifier type field.
2016-04-16 10738, 2016
Leftmost
Actually, same with publication.
2016-04-16 10711, 2016
LordSputnik
I'd leave the server as is actually - it errors when it gets to the database, but that's fine because the submitted form was invalid - there are many forms which could cause a DB exception, and we don't handle any of them at present
2016-04-16 10749, 2016
Leftmost
Yeah, the server side of that is fine. For valueless identifiers, though, that should fail.
2016-04-16 10725, 2016
LordSputnik
Leftmost: could we simply add a check constraint on the DB that the identifier value can't be empty?
2016-04-16 10733, 2016
Leftmost
Sure.
2016-04-16 10702, 2016
LordSputnik
and handle it nicely on the client with better validation
2016-04-16 10725, 2016
LordSputnik
Leftmost: it won't let me submit with a valueless identifier
2016-04-16 10750, 2016
Leftmost
The client-side correctly prevents it now, but the server-side doesn't do anything.
2016-04-16 10756, 2016
LordSputnik
OK
2016-04-16 10702, 2016
Leftmost
That was one of the fixes I pushed.
2016-04-16 10731, 2016
LordSputnik
it also seems to block null types
2016-04-16 10727, 2016
Leftmost
Oh, I bet that's a side-effect of my changes for valueless identifiers. Before it didn't check identifier validity for form submission at all.
2016-04-16 10702, 2016
Gentlecat
adesh: yes
2016-04-16 10748, 2016
adesh
thats cool. okay
2016-04-16 10726, 2016
Leftmost
LordSputnik, reworking the select component to take allowClear as a prop.
So, with that, I think all that will need to be done is to add the DB check that the value is not empty
2016-04-16 10719, 2016
LordSputnik
and then the identifier bugs should be gone
2016-04-16 10757, 2016
Leftmost
Excellent.
2016-04-16 10739, 2016
kartikgupta0909 has quit
2016-04-16 10703, 2016
Leftmost
LordSputnik, I'm going to assume we also don't want to let the user clear any of the selects in the relationship editor.
2016-04-16 10751, 2016
LordSputnik
Leftmost: good assumption :)
2016-04-16 10739, 2016
LordSputnik
bitmap: seeing an erro parsing entities.json in Constants.pm: ", or } expected while parsing object/hash" - have you (or Gentlecat) ever encountered that before?
2016-04-16 10720, 2016
bitmap
nope, haven't seen that
2016-04-16 10733, 2016
bitmap
I'm pretty sure it's valid json, otherwise MBS would crash for me
2016-04-16 10702, 2016
bitmap
at least when I left in trailing commas before, that's what it did
it sounds like JSON:PP is getting used for you, for whatever reason :/
2016-04-16 10748, 2016
LordSputnik
OK, I'll try to install JSON:XS
2016-04-16 10751, 2016
LordSputnik
This is on rika
2016-04-16 10743, 2016
adesh
Gentlecat: I am unable to get the stylesheets working. "fab compile_styling" yields while running fab "Warning: Command(s) not found: compile_styling"
2016-04-16 10718, 2016
adesh
and even after setting COMPILE_LESS= True in config.py
2016-04-16 10720, 2016
LordSputnik
bitmap: that seems to fix it, thanks ;)
2016-04-16 10736, 2016
bitmap
good :)
2016-04-16 10737, 2016
adesh
it still generates 404 for the css files
2016-04-16 10707, 2016
adesh
any idea what i might be doing wrong?
2016-04-16 10712, 2016
Gentlecat
build system has changed
2016-04-16 10736, 2016
Gentlecat
you need to install node, run `npm install`, and then fab build_static
2016-04-16 10748, 2016
Gentlecat
I didn't update the documentation yet
2016-04-16 10720, 2016
Leftmost
LordSputnik, fix pushed to master. Can you look it over to make sure it looks sane before I send it to rika?
2016-04-16 10707, 2016
adesh
okay
2016-04-16 10752, 2016
Gentlecat
take a look at bootstrap.sh
2016-04-16 10755, 2016
adesh
alright
2016-04-16 10743, 2016
Leftmost
LordSputnik, as far as putting -data-js on npm, I was debating that myself. In some ways it seems reasonable, but it's not a terribly useful package without -sql and it's not what most people will want for interacting with BB anyhow.
2016-04-16 10710, 2016
adesh
Gentlecat: its working now. thanks. btw im getting a lot of error 404 for missing-art.png
2016-04-16 10702, 2016
Gentlecat
where?
2016-04-16 10742, 2016
adesh
on "browse reviews"
2016-04-16 10711, 2016
adesh
shouldn't it intead be "placeholder_disc.svg" and not missing-art.png?
