The Official Status Thread
-
i thought baked included an oven. TIL
-
Ok, I think they meant boiled. I guess? You know, dump them in water?
No, I CBA to go and fix it.
-
i thought baked included an oven
It does; the article just doesn't state so explicitly.
@Onyx said:Ok, I think they meant boiled. I guess?
Reading on, that is indeed the case.I have a Wikipedia account; I could correct it… eh, can't be arsed
-
The fun of translation...
In Croatian, boiling and cooking are the same word. You can get the true meaning from context, of course.
-
Status: Opted to do nothing of value this Saturday. I went to bed at 10:00 PM last night and woke up at noon today. It's 5:00 and I've already eaten twice, read a third of a novel, watched a ton of pointless videos on YouTube, tried and failed to play Fractured Space, and am now running out of ideas on how to consume the rest of the day in a manner that accomplishes nothing.
-
kerbal?
Status: trying to find a good spring mvc tutorial. my google-fu is failing
-
now running out of ideas on how to consume the rest of the day in a manner that accomplishes nothing.
Attempting to post here today seems to consume a lot of time while accomplishing even less than usual.
-
Being frustrated is not a favorite pastime of mine.
Just found a TAOFLEDERMAUS video where he tests shotgun slugs with frozen vienna sausage as the projectile...I am amused for the next few minutes.
-
Status: my desktop ram decided to pass away, no crash, simply dissapeared from the bios. i think it's time to upgrade =/
-
Status: so it looks like i'm at least partially to blame for the 5xx storms today thanks to https://servercooties.com
-
You weren't the one who suggested t/1000 though; that was me
-
partially to blame
this is a big steaming pile of blame here and i'm not accepting all of it!
-
Well, yes, that's what I was owning up to; if anything, I'm neck-deep, you're only waist deep ;)
-
I'm quadrupedal, that's the same depth for me. :-P
-
And still, we were doing, what exactly? Effectively refreshing the page?
Ok, yes, that might have broke Discourse, and I wasn't helping by running my own testing instance as well.
But we broke it by... refreshing a page?
-
From what @sam was saying, loading any part of that topic (especially as we were loading the end) was querying and sending over 40000 post IDs
-
Dischorse was way broken before that was added to the site.
-
But we broke it by... refreshing a page?
not just any page.
/t/1k has broken the site how many times now?
-
At least twice
-
Any one of those factors could, in theory, cause an improvement in performance.
Irony.
-
there was the auto close thing.... and this one (still maximum of polling every 9 seconds, average would be about 12 seconds based on the data, that's not exactly DoS rates!)
i think there was another one....
-
40000 post IDs
Ok, ok,
OFFSET
is a bit of a painful operation. That's why you create an index on a function that calculates offsets frompagebatch number and...I'm dreaming again, ain't I?
-
Didn't we find out that they send the full list of post IDs awhile ago? As in many many months ago? On our own?
Because I'm sure we did. I'm just not going to try Discosearching from iPad.
-
You don't need to - just avoid
OFFSET
and use wherepost_number > start_of_batch
andpost_number < end_of_batch
assuming there's an index on post_number.
-
Oh, I forgot about
post_number
. That updates on post delete / move, I guess?In that case, yes, you can skip the function. In a more general case, you'd want a function.
EDIT: You might still want one, actually: filter by user. Though effectiveness is questionable there, would require testing.
-
Status: my desktop ram decided to pass away, no crash, simply dissapeared from the bios. i think it's time to upgrade =/
"Warning: Only diagnostic tests 1 through 64 can be run with main memory disconnected." -Frank Durda IV
-
The weird thing is that not only does purple merge into itself
Weird, but can be explained by a non-fast-forward (perhaps forced?) merge from one branch into another that both point to the same commit.@ben_lubar said:but black branches from and merges back into blue before blue gets another commit.
Normal. GitHub for Windows and GitHub.com only use non-fast-forward merges. Something about "making the DAG look pretty" or "not breaking git bisect" or "breaking git bisect" or something.@ben_lubar said:And purple merges into blue twice.
Normal. Someone merged from a branch before pushing, realized their mistake, pushed, and merged again.@loopback0 said:You don't need to - just avoid OFFSET and use where post_number > start_of_batch and post_number < end_of_batch assuming there's an index on post_number.
How do you consistently get apagechunkful of posts though?
-
How do you consistently get a pagechunkful of posts though?
On here, specifically,
post_number
(which is different thanpost_id
) should always be ordered and without gaps.It's the second number in the URL leading to the post. They get "fixed" and updated on moves and deletions.
In general, you'd write a function that calculates it and index the calculated values.
-
-
post_number
updated on
deletions
Breaking all links to a post? That sounds like a bad idea.
-
that's why it doesn't happen.... mostly.
once a post gets post number 42 in topic 8943 no other post will get that post number ever. the post can be jeffed to a new topic where it will get a new post number, but the posts won't bump up to fill the gap.
so only links to posts that have been jeffed will be broken.
that's why there's the unofficial rule about deleting posts in /t/1k
-
Status: Fuck the Edmonton Oilers.
-
Status: 200 OK
-
That... all makes sense, actually. Don't try to figure out Discourse at 2AM I guess (excuses, I'm good at them!).
So yeah, that would mean they probably offset it, and from what little code I saw, I wouldn't be surprised if it was done Ruby side...
-
Status:
==================TIME (ORIGINAL)=================== ⁞ eval list [catch {switchfail xyz} msg] $msg: "1 {couldn't compile regular expression pattern: quantifier operand invalid}" time list [catch {switchfail xyz} msg] $msg: 10.716427217496962 microseconds per iteration ⁞ ======================COMPILING===================== ⁞ ==================TIME (OPTIMISED)================== ⁞ eval list [catch {switchfail xyz} msg] $msg: "1 {couldn't compile regular expression pattern: quantifier operand invalid}" time list [catch {switchfail xyz} msg] $msg: 8.177307145865734 microseconds per iteration ⁞ =====================COMPARISON===================== ⁞ -------- list [catch {switchfail xyz} msg] $msg -------- Acceleration 31.05%
Bwahahahahahaaaa! Even with all the overheads of non-compiled parts in the test harness it's still much faster, and this is error-path code…
-
Status:
I just realized I have two YouTube tabs opened, both of them with the same video. Both of them reached that video due to the "Autoplay" feature.
I think I need more variety in my music choices...
-
-
Status: losing at Words With Friends. QUO IS A WORD DAMMIT!
-
That wouldn't be yamikuronue/testing in yellow would it? I had a detached head situation...
-
-
maybe? let me go find it again.
yep.
followed by a PR commit by @Onyx
Found another earlier one:
followed by a PR merge by me. so looks like that's something that happens with a PR sometimes? maybe it's because github does it rather conservatively?
-
Wait, which line is mine? I did have a small tiff with Git at one point that required some resets, don't really remember what I did in the end.
-
not sure
you can play with the graph yourself though ;-)
-
-
Another thing I do: commit, then revert, then revert again, because I keep thinking revert is rollback but it's not, it's undo the changes and commit again, and then I panic.
-
Status: losing at Words With Friends. QUO IS A WORD DAMMIT!
The Windows Phone version is really dodgy. It wouldn't allow you to play a Z under any circumstances. Last I knew zebra was a valid Scrabble word.
-
I'm on android :/
-
-
Nostalgia'ing while I procrastinate on yard work:
Yes, and horse-drawn carriages are still nice for casual trips in the park or sightseeing, but we don't use them to do actual fucking work. If the transportation world worked like the IT world, you'd see a third of the cars on the road were horse-drawn, full of delusional people inside yelling, "this is so much faster! This is so much better! You should totally learn this!"
-
God! I remember that thread. That test-pattern shirt! That fake fur!