so, then all we really need to do is fix up the timezone before we submit data to PG and we're done?
(if the user set the timezone)
alastairp
mayhem: when I looked into it, I think the recommendation was that if you want to store a time as local to the user then you need to store both the timestamp and the relevant timezone or offset as a separate data
lucifer
mayhem: alastairp: not sure, thinking about how we want to handle cases like travelling or when someone moves to another timezone after some years.
yuzie
yes, that makes sense, and then we don't need a column, but only apply the timezone when display to user?
alastairp
lucifer: yeah, we thought about those, but also recognise that they're quite infrequent and not the common case
lucifer: but moving timezone when a user moves, doesn't require us to have a separate column, does it?
if the user needs to shift the listens between date A and date B, by X hours, we can adjust the timestamps.
lucifer
mayhem, sorry not sure how the adjustment would happen. do you mean by the user changing the config?
mayhem
no.
literally adding X hours to timestamp if it needs adjusting
lucifer
i understand that part but where/what would do it? like would the ts writer do it the browser or something else?
alastairp: yup agreed, but we should decide if that's a use case we want to support.
alastairp
lucifer: by altering timestamps/data at ingest stage, it means that once that change is done, all new items should be good (after you move, for example)
for temporary movements, one would hope that eventually clients start sending the timezone info
in which case, that timezone will apply even if the configurerd one is different
atj
are you forgetting about DST?
mayhem
the case I was thinking about was when a user says: Oh, I was on vacation for these 14 days.
Move all timestamps by 2 hours between date X and date Y
lucifer
ah i see, mayhem so lettign the user select the listens they want to alter the timestmap for manually?
alastairp
atj: we have a utc timestamp, and a name of a timezone, so I believe with that we can say "render this timestamp with the correct rules for the tz"
mayhem
atj: I would love to forget about DST, but everytime I do, everyone gets angry that I am an hour late.
lucifer: yes
lucifer
mayhem: yes that makes sense now 👍
mayhem
cool
that said, we've not reached a conclusion: separate column or no?
lucifer
alastairp: yes makes sense. the user changes the config when they move and then the future listens get the timestmap accordingly.
mayhem: my understanding of the timestamp type is that modifying the time forward or back in this way will modify the underlying UTC value, which I don't think is what we want
> When you insert a value into a timestamptz column, PostgreSQL converts the timestamptz value into a UTC value and stores the UTC value in the table
so I think a separate column makes sense
mayhem
ok.
alastairp
and it should be an offset, the realized value of the timezone at the time of the listen
mayhem
is that a real PG column, or a value on the "data" JSONB?
alastairp
so that it solves the DST issue, and the issue of DST dates changing
atj
based on my limited understanding, separate column which is a FK to table of timezones
mayhem thinks the latter
lucifer
alastairp: we can let the user select a custom timezone to be applied for a particular period of listens instead of actual +- hours.
alastairp
lucifer: yep
my gut feel is that it's tied to listen time, and therefore should be an actual column
Freso
<BANG>
It’s United Nations Monday for South-South Cooperation!
(I didn’t find an appropriate song, sorry. :\)
TOPIC: MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | BookBrainz: #bookbrainz | Channel is logged; see https://musicbrainz.org/doc/IRC for details | Agenda: Reviews, Per-project mb_db users (bitmap, reo), summit covid preparedness (alastairp, monkey)
mayhem
alastairp: actual columns is serious work. :(
lucifer
mayhem: we need to process the entire table for some tasks so we can do this at that time.
get the code and all ready but not deploy to prod just yet.
reosarevok
mayhem, lucifer: shhh meeting time
:)
Freso
No mailed in reviews! (I could have sworn I saw one earlier today, but I don’t see it now, so it was probably never there.)
(Also, main machine is still out of business, so sorry for being a bit slow. Current setup is a bit cumbersome to work with.)
Anyway…
reosarevok: Go!
reosarevok
Hi!
I mostly worked on more React porting of small bits of the code such as cdstub and cdtoc templates
Freso
(Others up: bitmap, zas, mayhem, monkey, alastairp, lucifer, akshaaatt, atj, CatQuest, skelly37, Pratha-Fish, Shubh, yellowhatpro, ansh, riksucks, yuzie, Freso — anyone else who would like to give review, let me know ASAP!)
reosarevok
Also reviewed our TT macros and saw we could get rid of a surprising amount of them, which was nice
Still quite a bit to go, but a lot less :)
And small bugs and tickets here and there
Now helping bitmap test the React relationship editor, which is the next big bit of React coming, hopefully
Fin! CatQuest, go!
CatQuest told me that
"I'm healing and tired, will probably continue for a few weeks, been doing some editing (including cleaning that wikipedia-mess report)"
That'd be a list of issues in Wikipedia linkage the wikidata bot found, and I'm really happy CatQuest found some time to help with that :)
Freso: what about your week?
Freso
🙋
I mostly (finished with?) recovered from surgery plus tried to bring my main rig back to life. My working conclusion right now is that it’s the motherboard that’s dead. :( I’ve ordered new parts which should hopefully arrive between Wednesday and Friday, so I should be back up and running then. Had a couple of days where I was feverish but for the last several days I’ve been off painkillers and I’ve started eating hot
and non-soft food now, and it seems fine, so I think
I have more or less recovered from the surgery… so that’s nice at least.
In between struggling with this, I did manage to give blog access to a number of our GSoC students and I have scheduled (the last?) two posts for Tues and Thurs this week. :)
I think that’s fin.
bitmap: Go!
bitmap
hey
last week I fixed an issue causing missing CSS when deploying new containers, which I noticed when putting the new relationship editor branch on test.mb
(but css was missing on prod.mb for like 10 minutes while I debugged the issue and restored it...)
I also spent time testing/reviewing/updating open PRs, and fixed MBS-12566
MBS-12566: Incorrectly warn about bundled/shortened URLs in release editor
bitmap
fin, go lucifer !
lucifer
hi all!
last week, i worked on addressing feedback on existing PRs and finally merged almost a dozen of those.
thanks mayhem and alastairp for the reviews. further, i worked on verifying the canonical data fixes by running it over entire MLHD dataset which also led to wolf going down 😅
finally worked on moving messybrainz data to timescale db and debugging issues with api validation that were discovered as a result of the messybrainz PR.
alastairp: next?
alastairp
hi
mayhem
well done on the debugging and well done crashing a monster!
Freso
(Still up: zas, mayhem, monkey, akshaaatt, atj, CatQuest, skelly37, Pratha-Fish, Shubh, yellowhatpro, ansh, riksucks, yuzie — anyone else who would like to give review, let me know ASAP!)
Pratha-Fish 👀
alastairp
last week I reviewed some of lucifer's and ansh's PRs
I also did some work with Pratha-Fish to finalise his goals for the SoC project (extended until end of Oct), and worked with lucifer on testing his improvements to the canonical tables
I finished text dumps of the metadata and canonical tables, to be integrated periodically into lB dumps
I also worked on saving MB oauth keys in LB so that users can add MB tags directly from the metadata viewer
monkey and mayhem and I had a small talk about covid preparedness for the summit, which we'll discuss after the summit
Pratha-Fish: do you want to speak? I know you're busy with exams this week...
Pratha-Fish
alastairp: sure
Hello everyone :)
lucifer
alastairp: after the meeting? :)
alastairp
lucifer: after reviews :(
Pratha-Fish
last week I worked along with alastairp to figure out how to restructure the MLHD code base
lucifer
:+!
Pratha-Fish
and how to make it more "presentable" and usable for anyone new
Halfway done through the README file. Super happy to share that we'll be potentially writing a series of blogs explaining all the new and interesting findings that we made through the project :D
Currently busy with exams as well. But I'll be back on 14th 👌
alastairp
not potentially, we will!
postive thinking
Pratha-Fish
indeed! Super excited for the blogs 🔥
fin 🐡
riksucks: next?
Freso
Or maybe… zas?
zas
hey
Freso
(Still up: mayhem, monkey, akshaaatt, atj, CatQuest, skelly37, Shubh, yellowhatpro, ansh, yuzie — anyone else who would like to give review, let me know ASAP!)
zas
last week, I reviewed last PRs from skelly37
I also renewed few expiring certs (*.mb.o), did usual upgrades
rebooted wolf that was killed by lucifer ;)
mayhem snickers
plus usual supervision, patches, user support, Picard related stuff
TOPIC: MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | BookBrainz: #bookbrainz | Channel is logged; see https://musicbrainz.org/doc/IRC for details | Agenda: Reviews, Per-project mb_db users (bitmap, reo), summit covid preparedness (alastairp, monkey), summit: extending main topic days to friday (mayhem, alastairp)
fin. mayhem ?
alastairp
mayhem: ^ not sure if you think this is worth mentioning?
mayhem
k
last week was catching up from the vacation and other emails. some convos that were already mentioned...
but, over the summer I worked out that QuickBooks' mail servers have been blacklisted by many providers.