The Official Status Thread
-
-
But not before he's hatted it ;)
-
A double-hatting? I am intrigued.
-
-
d but I actually edited the picture or I'd have got it much sooner. :P
-
-
-
This needs to be added to the list of available emojis.
-
Also, get thee to some officialesque emoji topic.
-
Status: wondering why a program that takes a folder full of images and converts them to a proprietary texture format pegs one CPU core for 10 minutes when given a folder of 80 images.
-
The hatted goalposts you composed. I would have quoted it, but I had this sneaking suspicion that Discourse would have ruined it and I did not feel like screwing with it.
-
-
Don't make me kick you.
-
Status: Been working on getting our build and unit test environments automated using Jenkins CI. I have so far set up 97 jobs and I'm not even close to having everything into the system!
-
-
I am just posting to claim post no. 18800.
Like a true ninja.
Well, kinda. Jeffings. But that's what the newlevator says!
My green-greenyellowvator agrees!You're both wrong!
-
Currently? Yes.
The plans have been set into motion by a powerful ally...
-
-
Status: why am I getting like a million "Good Post" notifications? If someone dicking around?
-
Status: wondering why a program that takes a folder full of images and converts them to a proprietary texture format
If the format is DDS, Paint.NET can open and save those natively.
PRO-TIP!
-
Because it's shit. You're using Valve tools again aren't you?
-
-
The hatted goalposts you composed. I would have quoted it, but I had this sneaking suspicion that Discourse would have ruined it and I did not feel like screwing with it.
I did it frist! @boomzilla's not original!
-
Status: As I get older, I end up on more conference calls. People told me about back pain, rogue hairs, hangovers, etc. No one mentioned conference calls.
-
Status: As I get older, I end up on more conference calls. People told me about back pain, rogue hairs, hangovers, etc. No one mentioned conference calls.
Not only that, but no one bothers to tell you when that stuff starts:
- Back pain: really started around 29
- Rogue hairs: not sure when they started, but started noticing them at 30
- Hangovers: not a problem since I don't drink. Closest thing I've got is chronic migraines: age 11
- Conference calls: really started at 31.
-
Status: wondering why a program that takes a folder full of images and converts them to a proprietary texture format pegs one CPU core for 10 minutes when given a folder of 80 images.
Must be network activity. MilwaukeePC strikes again!
-
The part I'm complaining about is the "one CPU core" part. If it pegged all 8, I wouldn't have an issue.
-
The part I'm complaining about is the "one CPU core" part. If it pegged all 8, I wouldn't have an issue.
Network activity + busywaiting on I/O, I guess.
You can't italizice a plus sign in this font, apparently.
-
Audacity filters are easily-parallelizable, and yet only one core. Making every Audacity task take far longer than it should.
Because Audacity is open source and therefore shitty.
-
Network activity + busywaiting on I/O, I guess.
That doesn't make sense. In that case, you'd expect intermittent activity strewn around all of the cores. It seems pretty clear that it's a single threaded program that @ben_lubar's working with.
-
STATUS:
I have an INSERT trigger, that coerces my data and fills in some "cooked" fields. Along the way, it also validates the data, throwing error if unable to process.
And I also have a CONSTRAINT on the field, performing the same kind of validation, using the same regex.
I can do with just the trigger.
But it feels kind of smelly, to trust a trigger with my data's integrity.
But... it's the same fucking code. Trigger should work just fine....
But still, a constraint is the proper way to do validation. Trigger is just a gate keeper. Constraint is the guarantee.
But... performance
But... data integrityAgggh!
-
That doesn't make sense.
-
-
The Emoji Only topic is
-
Kill the trigger and supply the cooked field values with virtual columns?
-
Kill the trigger and supply the cooked field values with virtual columns?
What are "virtual columns"?
-
An oracle thing. Basically:
CREATE TABLE EXAMPLE ( FOO NUMBER, BAR NUMBER, FOOBAR AS (FOO + BAR) );
You cannot insert in FOOBAR, but you can select from it.
-
STATUS:
Stupid notepad++ suddenly decided to start dumping its temporary crap on my desktop.
LIKE SHITTING ALL OVER MY DOCUMENTS WASN'T ENOUGH
-
An oracle thing. Basically:
CREATE TABLE EXAMPLE (
FOO NUMBER,
BAR NUMBER,
FOOBAR AS (FOO + BAR)
);You cannot insert in FOOBAR, but you can select from it.
Interesting. I don't think postgres has that. I'm basically calculating things on their way in.
-
It also does the work in the GUI event thread because fuck responsive interfaces.
-
An oracle thing. Basically:
CREATE TABLE EXAMPLE ( FOO NUMBER, BAR NUMBER, FOOBAR AS (FOO + BAR) );
You cannot insert in FOOBAR, but you can select from it.
That's a Computed Column in MSSQL-land; I imagine Postgres has something similar.
-
It also does the work in the GUI event thread because fuck responsive interfaces.
Paging bridget99...
-
What are "virtual columns"?
Create a view which contains the original table's columns and also has additional columns populated with whatever "cooking" you're doing.
The validation is trickier, I'd keep the constraint and ditch the trigger.
-
I could have sworn I have seen this conversation elsethread recently. It's deja vu all over again!
-
@Mikael_Svahnberg You just reincarnated into a slightly different timeline is all.
-
LIKE SHITTING ALL OVER MY DOCUMENTS WASN'T ENOUGH
See what happens when you don't write Windows apps according to the API?!
-
Create a view which contains the original table's columns and also has additional columns populated with whatever "cooking" you're doing.
The validation is trickier, I'd keep the constraint and ditch the trigger.
If I create a view, I no longer need the trigger.
I suppose it can work, but seems to be adding extra complexity, without giving much back.
-
Well if Postgres has computed columns, use those. That's what they're for.
I'm just offering an alternative if it doesn't.
Personally, I hate triggers and I'd always do my best to find an alternative to using one. Triggers are basically "scary action at a distance" in its purest form.
-
This. It may be defendable to put the derived value in a column if it's expensive to compute, but probably it's just taking up extra disk (and tanking your insert performance, if per-row triggers in postgres are as slow as in oracle).
And you're not gaining complexity, you're losing it, even if you use a view. After all, you lost a trigger.
-
Well if Postgres has computed columns, use those. That's what they're for.
I'm just offering an alternative if it doesn't.
Personally, I hate triggers and I'd always do my best to find an alternative to using one. Triggers are basically "scary action at a distance" in its purest form.
You can't add computed column to a table, I'd have to use a view.
Except, since these are auto-generated tables, that'd mean changing like 20 functions that manipulate these tables and worrying about refreshing and storage... not worth the effort.Agree about triggers in general. I can live with these because they are basically a pure function, that just does a simple bit of coercion/validation/processing and isn't doing anything crazy, like executing SELECT-s all over the place.