« first day (4889 days earlier)      last day (22 days later) » 

12:21 AM
or dyslexia
1 hour later…
1:22 AM
You’re just emmental
yeah my brain is full of holes and melting
i dunno how i missed that emmental pun must be all the gin
1:50 AM
You should really switch to vodka martinis
Won't I be too shaken for the morning meetings
2:40 AM
El Diablo
4 hours later…
6:18 AM
Wordle 1,070 3/6*

6:57 AM

Wordle 1,070 6/6*

@PaulWhite Thank you. I always have hard time deciding, whether I should make such basic edits, since as new user, someone gotta review my changes. Which means it's work for two people. But I'll take your approval as encouragement the next time.
4 hours later…
10:29 AM
Name resolution is a mess
Those two statements can run correctly while resolving to different tables
So which principal are you running it under and which schemas are those tables in?
Does it matter? Yes, there are two tables in different schemas, one dbo, one not
Point is, the rules are bizarre
I get that. I've asked because of repro
That's why I enforce two-part naming anyway
    CREATE TABLE MyTable (paul integer NULL);
CREATE TABLE dbo.MyTable (dbo integer NULL);
ALTER ROLE db_datareader
    ADD MEMBER Paul;
    ON OBJECT::dbo.ReadMyTable
    TO Paul;
    EXECUTE dbo.ReadMyTable;
@Zikato Repro ^
Thank you
I mean this one is not that bizarre or unexpected
10:44 AM
I think many people would find it bizarre that the 'same statement' resolves to two different objects depending on whether it is run dynamically or not
i.e. one respects the user's default schema and the other doesn't
And the behaviour is different if you run the statements outside a stored procedure, that is as ad hoc SQL
Perhaps you were expecting the demo to produce buttered toast
That would indeed be surprising
ok, I concede it's a bit surprising. I was expecting the first result to return Paul and second one to return DBO but it's the other way around
I shall consider you surprised then
> There is no ambiguity, as object references are always resolved from the point of view of the procedure owner, not the current user.
Ownership of the procedure is not a factor here
I thought it might behave differently if you impersonate login instead of user, but it's the same
Surprise increases
For added enjoyment, try to find where this is documented
hmm, but when I created procedure Paul.ReadMyTable, both statements did read from the Paul schema
10:59 AM
Of course
@PaulWhite So this is not wrong
@Zikato It is wrong because you're not changing the procedure owner, you're changing the containing schema. You can ALTER AUTHORIZATION on the procedure all you like, that's not the important aspect
oh, I see the distinction
It's all very confusing and bizarre. Arcane would perhaps be a good word
I still don't get how the dynamic SQL resolves to the user schema. I thought dynamic SQL are considered anonymous or temporary procedures
You've nerd-snipped me well
11:05 AM
That's just the name resolution rule for dynamic SQL
Point is the name resolution rules are different
> Object names used with all Data Definition Language (DDL) statements such as CREATE, ALTER, or DROP statements, DBCC statements, EXECUTE and dynamic SQL statements should be qualified with the name of the object schema if other users are to use the stored procedure.
> Specifying the schema name for these objects ensures the name resolves to the same object regardless who the caller of the stored procedure is. If a schema name is not specified, SQL Server will attempt to resolve the object name first using the default schema of the caller or the user specified in the EXECUTE AS clause and then the dbo schema.
But non-dynamic SQL in a procedure defaults to the procedure's schema scope, not the user's default schema
Things that make you want to punch SQL Server in the face
I suppose it sort of makes sense because the dynamic SQL doesn't know it was called from a stored procedure, and which schema that sp is in
But it is unhelpful
I've heard of a shop where they don't specify the schema on purpose. I forgot the details but I think it was multitenant in a single DB and they wanted the default schema routing.

No idea if they use dynamic sql
No doubt there are uses for it
And backward compatibility means it can't be changed
But still
Thank you, I've learned something
to never doubt Paul White I guess
11:13 AM
That's not a good rule. I make mistakes like everyone
But not about parameter sniffing
Or cursors
@PaulWhite In that case prepared to be challenged at every opportunity. There is no middle ground
It's the modern way. Everything is on or off.
Erik's not infallible either, but his rates are reasonable
maybe that's why
Could be one of the local factors in play
11:49 AM
@Zikato SO teams does that
This name resolution thing reminds me of the token perm security cache issues I’ve seen
@ErikDarling What would you say the most common cause and solution is?
12:12 PM
Schema-qualification, I suppose
12:25 PM
@PaulWhite the most common cause I’ve seen is applications doing a whole bunch of impersonation
I’ve heard rumors that a lack of schema qualification can lead to compile locks as well
But I mostly see bad compile locks when things are already schema qualified
a memorable one was the user would log in, hit master, change to a different user, change databases, then run a series of procedures that changed to another user and reverted at the end
I suppose they had excellent reasons
something something java
i found that cache getting up above 10gb at some shops
12:51 PM
That post doesn't mention the access check cache configuration options
AIUI TFs 4610, 4618, and 4621 are the old approach
It's strange the cache should grow so large with the default options shown there
I guess that's part of the problem
Or maybe even the main problem
i'd not heard of them in relation to security cache issues
largest one i ever saw was 24gb
I miss the days when KB articles were detailed and honest:
> The TokenAndPermUserStore cache store does not scale well as the number of entries in the cache store increases.
Accessing that KB directly today results in a redirect to the access check cache documentation page
i've updated the post to include a ling to that docs page, thanks.
also added a note about the TFs being startup flags, because @SeanGallardy yelled at me about that once upon a time
1:01 PM
Are they. I thought they were just normal globals, but that makes sense
iirc he said they were a bit like 8048, where the best results are when used at startup
Well, that's when the caches are created so yeah
Fundamentally, this is just another SQL Server gotcha. Permission checks are expensive, so they get cached. If you arrange things so many, many combinations of permission checks are needed, things may go badly for you. There's no warning you're doing something dumb of course
Not even doing something dumb, something SQL Server doesn't handle well
It's hard to imagine a 24GB cache speeding anything up
you know what a great way to fix the problem is
use sa
skip all sorts of permissions checks
I nearly typed that a few minutes ago
sa = skip all
1:05 PM
That used to be very nuch the norm
@ErikDarling I always wondered what that stood for.
now u no
Ultimately, these places don't even have complicated permissions setups, just huge numbers of users or questionable access patterns
Like, you might have 3,000 users but they don't have 3,000 distinct permission sets
Anyway, I hear caching is hard
yeah, having a lot of different users has definitely been a symptom
now that i think about it, the 24gb people had 25k users
as u do
perfectly normal st8 of affairs
1:10 PM
i don't know how many of them were active
neither did they, probs
entropy being what it is these days
do you think reading comprehension being what it is these days has anything to do with entropy being what it is these days
sounds connected
one never quite knows what is correlation and what is causation tho
and everyone who doesn't know the difference will die
@ErikDarling I probably didn't yell
1:20 PM
it was a stern talking to
It was probably a matter of fact statement
needed more emojis
@Zikato eventually, but that's also "soon", depending on one's perspective
I must have been gone a long time, the tags are all grey and bolded now
well b4 entropy flattens out, anyway
@SeanGallardy Looks lovely, doesn't it
1:26 PM
@SeanGallardy bold words help with reading comprehension
Experts at Stack Exchange worked v hard on getting it just right
The tags are doing the yelling now
I think it's easier to read
makes them easier to see
reading is an exercise left to the user
At the cost of seeing anything else
Turns out the changes to the League Tables are not for the per-site views everyone uses
Could you confirm explicitly that the user lists at xxx.stackexchange.com/users will not be affected at all by this change? At first I assumed that's what you meant by "leagues", and downvoted the post accordingly, but after a more careful reading, it seems you're only talking about the lists that appear at stackexchange.com/leagues, which are checked by far fewer people (I'd almost call them a hidden feature) and therefore would be less of a loss. — Rand al'Thor May 8 at 14:08
That's correct, @Rand, it's just stackexchange.com/leagues. The per-site users page will not be affected. — Slate ♦ May 8 at 15:30
I don't think I even knew the network-wide ones existed
Gr8 PR job then
1:36 PM
finally something i have better seo than
I still can't understand how it's a difficult query to run, but I care much less now
Due to query timeouts, the data are often out of date or inaccurate. Ahem, my rates are reasonable. — Erik Darling 2 days ago
five upvotes
wonder when i'll hear from them
Probably raising the PO as we speak
So, 6-8 weeks
hope my landlord is understanding
the evidence strongly suggests so
1:45 PM
hm yea
in today's video i'll be taking the "avoid subqueries" crowd to task for the foolishness
it's 80 degrees (f) today
good on you
very nearly 27C
i've discovered an affection for vodka martinis
NAME = 'PRIMARY' looks wrong, if you want the primary filegroup then you want FILEGROUP = 'PRIMARY' Curious what you think REWIND, NOUNLOAD is supposed to do to a disk file. NOSKIP, NOFORMAT, NOINIT probably isn't desired with file backups, you usually want INIT with a unique file name. — Charlieface yesterday
Comment relevancy continues to decline
That's up there with the dolphin effort from yesterday
2:02 PM
jacuzzi face upped the ante on that one
@ErikDarling with olives?
@PaulWhite wife gets all the olives, i'm left with a cup of cold vodka
Aug 10, 2023 at 16:57, by Erik Darling
@SeanGallardy they’re so good you have to give them all to the girl you’re eating with
consistent snapshot confirmed
2:30 PM
the first rule of romance is to give a girl your olives
2:46 PM
wife h8s olives; I luv them
maybe you're the girl
who can tell
if we're doing the sitcom math, she works in a shoe store, and you hang out on the couch all day vaping and watching television
that makes her al bundy and you peggy bundy
you're even starting to get the hair
I never watch TV, muppet boy
My wife and I are unified in our dislike of olives.
2:48 PM
okay, you watch a screen
how else would you keep the cricket chat so lively
@JoshDarnell rare to see two tops in a successful relationship
@PaulWhite Cricket chat says otherwise
Are you all retarded
can't speak for everyone else...
I never watch TV
The TV watches you though
2:51 PM
@ErikDarling We defy the odds!
There's no TV in this room
My wife has a small TV in the bedroom
She doesn't watch cricket on it tho
A friend of mine in high school had a shirt that said, "TV is my parents" and I still remember how funny I thought that was.
@PaulWhite i think you're supposed to say r-worded now
@ErikDarling Local factors, I expect
We can't all be quoting high works like The Bundys
well anyway, substitute tv for computer and we've got a modern day match
i think having 10 computers is equivalent to having one television
2:55 PM
Someone said "he planned to unalive the other guy" to me the other day and I almost punched them.
Levels of linguistic torture are becoming unbearable
is he a justice impacted or unhoused person
Degrees of Tiresome hitherto unsuspected to even exist
@ErikDarling 🤢
it's quite aggravating to wake up every morning knowing you were born 20 years too late
3:05 PM
@JoshDarnell maybe you meant "ununharmed" instead of punched
calling bad things nice names doesn't make them good. i wonder what the use is.
perhaps it helps one cope/ignore. easier than doing anything actually useful
@Lamak double minus non unharmed
many declining societies are unable to use an apostrophe correctly
we don't know if it is correlation or causation
or how many killer homeless ex-cons have been affected
at some point in the future, someone may have no idea what any of those words mean
3:20 PM
it's interesting we all kind of silently assumed human progress would always be in the forward direction
I'm currently reading a book about how, big picture, the world is getting better (counter to most people's intuition).
I would say that's true, but uneven. Very uneven in places
a lot of human progress over the next 100 years hinges entirely on space flight and exploration
and Reddit inputs to AI
good to see you buying into the Star Trek vision tho
Factfulness: Ten Reasons We're Wrong About the World – and Why Things Are Better Than You Think is a 2018 book by Swedish physician, professor of international health at Karolinska Institute and statistician Hans Rosling with his son Ola Rosling and daughter-in-law Anna Rosling Rönnlund. The book was published posthumously a year after Hans Rosling died from pancreatic cancer. In the book, Rosling suggests that the vast majority of people are wrong about the state of the world. He demonstrates that his test subjects believe the world is poorer, less healthy, and more dangerous than it actually...
3:25 PM
What the hell is that word
well, i think it's a worthy goal to take the aspects of humanity that work and separate them from all the stuff that doesn't
@PaulWhite A made up one, for deepest irony.
I don't know that a lakh of oxygen and low gravity will necessarily improve the average human
Though come to think of it...
I wonder what the first Moon/Mars wars will look like
3:28 PM
hopefully lasers
Slow ones, like in Star Wars
plasma bolts
i hope i can bring my humidor into space
It'll also be interesting to see which drinks work best up there
3:33 PM
you're a euphemism
u phemism
hey can i train an ai on this chat
3:44 PM
I don't know, can you?
@JoshDarnell do you still have that utility for stmgr
3:55 PM
I hope he hasn't killed it
he would use java
@Zikato v good
josh is prob using rust or go by now
i hear he went to a bootcampathon
got into an accelerator
well, I hope he wore a helmet - unlike some EU leaders
someone 4got to czech their safety gear
4:10 PM
did i post my iranian helicopter memes in here
i can't recall
no, there was just one mention of Woeing
oh okay
i only have two
at the moment
big fan, you say
4:13 PM
don't u dare
you're about to get unfriended
you can see it, can't you
4:26 PM
🫒 🫒🫒
4:37 PM
> Daddy's a failure; he used a subquery
No rant about people solving fragmentation by updating stats, but otherwise 💯
well you have to spread the content out a bit
try to stay on track
> Today's blog post looks at using CLR integration to leverage .NET assemblies from within your TSQL queries.
can we not
4:46 PM
@ErikDarling I think I do, yeah. I'm trying to find the hard drive it was on.
hard drive he says
what year is this
just random zipdrives floating around your office
some iomega jaz drives
maybe he used one of sean's floppies
sean is famous for his floppies
wait, is Sean SAM
that's a hot take
4:56 PM
@ErikDarling Confirmed I do have the code. I haven't tried to run it to see if it still works :) Why do you ask?
it would be funny to train ai on this chat history
That would be great.
5:15 PM
If you ask it about rafts it shuts down.
2 hours later…
7:01 PM
Great memes, Erik!
i'm feeling the memetic energy flow through me
8:03 PM
4 hours later…
11:54 PM
when life doesn't give you olives
or when all your olives gone to the girl

« first day (4889 days earlier)      last day (22 days later) »