« first day (4467 days earlier)      last day (407 days later) » 

12:09 AM
I wouldn't say you are unreasonably difficult
 
 
1 hour later…
1:33 AM
'Morning all.
 
 
1 hour later…
2:52 AM
@PaulWhite thanks, i think
 
3:50 AM
03/28's:
Wordle 647 3/6

⬛⬛⬛🟨⬛
⬛🟩⬛⬛⬛
🟩🟩🟩🟩🟩
 
4:06 AM
Sunrise
 
 
2 hours later…
6:20 AM
Morning
Wordle 648 5/6*

⬛⬛⬛⬛🟨
🟨⬛⬛🟩⬛
⬛⬛🟨🟩⬛
⬛⬛🟩🟩🟩
🟩🟩🟩🟩🟩
 
 
1 hour later…
7:35 AM
@HannahVernon Maybe. It's hard to remember how discoverable it was before I knew how to get to it from the 'impact' box on user activity page summary
I think if you have a non-comment flag declined you get a more direct link in the message
 
8:13 AM
@HannahVernon yeah, it's hard to find if you don't know where it is: Activity -> IMPACT section, "helpful flags" is a link
 
 
2 hours later…
9:48 AM
I've always thought that tiny link was stupid and completely undiscoverable. I keep forgetting where it is and I go on that page every couple weeks.
 
10:00 AM
@HannahVernon By the way, this link works for everyone dba.stackexchange.com/users/flag-summary/current
2
 
10:16 AM
That still isn't very discoverable. You would think there would just be a Comments button on the left along with Answers Tags Badges etc, but no, Someone needs to go back to UX school.
 
It's possible it's deliberately hard to find.
@Charlieface Comments?
Seems like a suboptimal design to me, but what would I know
 
10:52 AM
@PaulWhite Sorry meant to say Flags, but yeah Comments is missing also.
 
11:24 AM
Wordle 648 3/6

⬛🟩⬛🟨🟩
⬛🟩🟩🟩🟩
🟩🟩🟩🟩🟩
Morning
Had a good starting word today. Alas, the sought after word wasn't....
 
11:49 AM
Any GitHub expert know why it shows diffs where there are no diffs?
I've used both manually cloned version and the Edit button in the GUI. The manually cloned repo doesn't show these changes on commit, only when Pull Request is created
My guess was encoding or different line ends. But no warning from SQL prompt on different line endings and for encoding it seems too random
 
@Zikato Yea that's what I was about to suggest or for some reason different whitespace characters? Could throw both into Notepad++ or something and turn on the Show All Characters to double check. Weird.
 
I have autocrlf so it would have to be the original that has the mix. But it's weird that even GitHub gui would change that
well, it was a roundabout way of saying: I blame @ErikDarling
because he's the maintainer of that repo
 
lol
 
12:10 PM
GitHub's diff is garbage, given it's based on the standard file-based diff command. We need a diff tool that actually understands syntax, so it can ignore insignificant whitespace, and insignificant reordering of declarative syntax.
The amount of times I'm looking through diffs of XML and I realize the only changes are attribute="value"</tag> and attribute="value" </tag>
 
Which is that programming language in which whitespace is important?
 
For some reason the XML formatter in VS has changed to add whitespace after the last attribute. So annoying.
@PaulWhite Python off the top of my head. And XML at the beginning or inside node names.
Or for that matter any language inside a quoted string
 
Probably 🐍 I was thinking of
 
Python = 🐍
Whitespace is an esoteric programming language developed by Edwin Brady and Chris Morris at the University of Durham (also developers of the Kaya and Idris programming languages). It was released on 1 April 2003 (April Fool's Day). Its name is a reference to whitespace characters. Unlike most programming languages, which ignore or assign little meaning to most whitespace characters, the Whitespace interpreter ignores any non-whitespace characters. Only spaces, tabs and linefeeds have meaning.A consequence of this property is that a Whitespace program can easily be contained within the whitespace...
 
12:21 PM
@Zikato often that's because of tabs vs spaces
 
But that wouldn't be changed by the github edit gui. I also look at tabs vs spaces in my code editor
 
@PaulWhite cheers!
 
well, adding ?w=1 after the url ignores the whitespace, but I can't force people to do it
 
Haskell was the other one
 
The difference between 0D 0A and 0A 0D and 0D
 
12:24 PM
Anyway, here it is. I'd be grateful if you could take a look @ErikDarling github.com/amachanic/sp_whoisactive/pull/94/commits/…
 
i just looked
 
in that case I'm grateful
 
Oh noes, someone disagrees lol
 
Silliness wrt tabs will not be tolerated!
 
12:25 PM
yeah, well at least now I know to add the ?w=1
 
@PaulWhite wrt? I'm not familiar with that acronym
 
@PaulWhite Converted me months ago to spaces. Mainly due to blogging
 
@J.D. with respect to
 
Ah
 
@ErikDarling What is that same thing?
 
12:27 PM
the inconsistent line breaks
that @Zikato is talking about
from the pull request @Zikato made just now
because @Zikato hit an error converting xml in who is active
poor @Zikato
 
Won't you think of the people with more than 128 nodes in their XML plan?
 
Sounds like a query needs tuning
 
i do in all my stored procedures
can't help adam being selfish
is try_cast available in sql server 2005?
your change might break the documentation
> Version 11.32 - July 3, 2018 (ALL SQL Server versions 2005+)
 
Added in 2012 IIRC
 
yes, yes
 
12:33 PM
Yeah I saw the joke but you know
It appears TRY_CAST is not a reserved keyword but TRY_CONVERT is. Never noticed that before.
 
yep
wild, huh?
 
I like Glenn Berry's approach of having a script per SQL version
 
way out wiggy weird
 
go contribute to glenn berry's repo
 
ha ha ha
Who even uses CAST anyway.
 
12:35 PM
fools
leading comma people
 
COALESCE types
 
be gone
 
I mean why pay price for a backwards compatibility and not use the new features
 
you can do both with dynamic sql
 
I hear that's vulnerable to injection
Time for a cultural joke
 
12:38 PM
everyone keep your clown noses on, i'm not gonna talk about anyone's mom anymore
 
when you're a hammer all the problems look like a dynamic sql
 
that probably makes more sense in Czech
 
well jesus was a carpenter
i bet when he comes back he's gonna love dynamic sql
 
there's not much in the bible about his feats of carpentry is there?
or are there a few verses about a particularly nice cabinet he made?
it's been a while
 
got me
i don't think he even made crosses
 
12:44 PM
made some people quite cross
 
not hard to make italians angry
 
made a few mistakes along the way
but nailed it in the end
 
Wordle 648 3/6*

🟩⬛🟨⬛🟩
🟩⬛⬛🟩🟩
🟩🟩🟩🟩🟩
Sometimes you just get a good start, hey
 
still blew it
 
yeah that totally should've been a two
 
12:47 PM
I was thinking it would have been because my 2nd guess was fairly uncommon
 
@ErikDarling so is that a "no" on the PR?
 
what would be the least problematic line ending?
 
CRLF windows, LF Linux
 
That dialog box is so stupid
 
12:48 PM
I go with typewriter line endings
 
It pops up sometimes though I only ever edit my own code
 
@PaulWhite have you tried being consistent?
 
Humans really can't standardise on anything
 
Not even humans
 
@Zikato i don't know, adam usually surprises me with what he does and doesn't care about.
 
12:55 PM
Adam: I'm pretty sure the whitespace issues are due to Windows-style line endings. I flipped the proc over to Linux-style during the last release cycle but maybe it's not consistent.
 
try_flip
 
seems like linux was a mistake
 
if you say linux backwards you get windows line endings
 
with just a line feed, words on the next line would be horribly indented
I just don't understand Linux at all
it's worse than tabs
 
i would also blame tabs for that
 
1:02 PM
you can't be just feeding lines all willy-nilly, the carriage must be returned
 
If you don't return it, then it's just stealing
 
Wordle 648 4/6

⬜⬜⬜⬜🟨
🟨⬜🟨⬜🟩
⬜🟩⬜🟨🟩
🟩🟩🟩🟩🟩
 
⬜⬜⬜⬜⬜
⬜⬜⬜⬜⬜
⬜⬜⬜⬜⬜
⬜⬜⬜⬜⬜
⬜⬜⬜⬜⬜
🟩🟩🟩🟩🟩
 
1:21 PM
@PaulWhite Tabs > Replication tho?
 
putting multiples of 3-4 spaces is evil
 
tabs are noxious
 
Not as bad as commas at the end of lines though.
 
writing with line feed only
                                      is what happens when the lever
                                                                                    on the on the typewriter is broken
 
1:38 PM
> The reason I post everything publicly is so that you can answer your own questions for stuff like this. I don’t even require an existing database to test – the script creates a new database for you. Please stop asking me to do additional work here and on LinkedIn – the idea of this post is so that you can do your own testing. Thanks for understanding.
 
1:58 PM
@HannahVernon I was hit up yesterday to look at why SQL Server wasn't shutting down, said it just say there for 4 hours doing nothing but won't shutdown. Sounded familiar. Guess what I found on one of the threads.... isitservicebroker.com
 
LOL
I have no analytics or any server side tracking for that site, so it's nice to hear someone mention it
 
I saw it waiting on service broker in notify shutdown and was like... yeah, I know someone that would appreciate this
 
2:32 PM
If you create a trigger on master for the event ALTER_DATABASE, which contains the statement WHILE (1=1) WAITFOR DELAY 23:00:00; and you execute an ALTER DATABASE, will all other possible DDL and DML statements on that database get held up with a Sch-M lock? Asking for a friend...
 
2:50 PM
@Charlieface You need better friends
 
3:04 PM
@mustaccio Impressive!
 
== English == === Pronunciation === === Phrase === asking for a friend (idiomatic, humorous) Indicates that a question is embarrassing by pretending to be asking on behalf of another person. 2016, "Celebrities React to Presidential Town Hall Debate 2016", Just Jared (Oct 9, 2016): How many more of these #debates do we have to suffer through? Asking for a friend. — Eva Amurri Martino (@TheHappilyEva) 2018, "Lakers survive Ball's gaffe to beat Thunder in OT". — LeBron James (@KingJames) "So what really is "in the shooting motion" rule?!?!?" James wrote on Twitter. "Asking for a friend...
 
I know, I was being funny... well I thought I was being funny
 
3:31 PM
@SeanGallardy I thought it was funny 😁
 
I knew I liked you, Josh
 
pertinent to the sp_whoisactive discussion earlier, does anyone know how to configure github so that on commit or merge, some action runs to clean files? ideally looking to normalize line endings, replace tabs with spaces, and probably clean up trailing white space on new lines.
 
looks legit
 
@SeanGallardy Looks funny now that I realize you knew. I guess my response was even funnier.
 
3:46 PM
Mar 14 at 14:59, by Erik Darling
i thought i was funny until i met you
 
looks arent everything
 
what do you have frog eyes
 
@Charlieface JEAGL in proper chat fashion
 
@mustaccio thanks
dumb question: does it support sql? i don't see it listed here
or rather, i see a lot of language specific things, but not sql
 
4:27 PM
@ErikDarling I don't have anything at ready, but I think I should be able to help with those
 
i thought you already did stuff like that
 
yeah but in Azure DevOps, not GitHub Actions. And there are several tiers - you can set the line endings with .gitattributes. Replace tabs with spaces and trailing ws would be a powershell either in githook or in the GitHub workflow
 
@ErikDarling It supports Prettier, which has a SQL plugin. Don't ask me how to put this all together though
Unless you want to get invoiced, that is
 
4:58 PM
@mustaccio you can send the invoice to adam machanic, ha ha ha
 
 
2 hours later…
7:26 PM
Anyone using Row-Level Security in SQL Server?...like it / hate it?
 
8:04 PM
I mean, it's ok
 
as long as you don't use a scalar udf for the security function it's tolerable
perhaps it's less icky with froid, but i wouldn't chance it
 
@ErikDarling True, definitely still room to make them a performance bottleneck even with single statement TVFs. E.g. I'm doing one right now that references a recursive CTE that's a little iffy at the moment.
 
yes you always seem to find the most obtuse way to do things
 
😆 obtuse is subjective if we aren't talking triangles.
I think I might've found a bug with RLS. Seems like if you reference an object that has an RLS predicate against it, inside of a different TVF used for a separate RLS predicate and policy, that second policy doesn't get applied.
Will probs post a question when I get a chance to create a clean set of repro scripts.
 
Look at this guy, making a repro and scripts and being super helpful
#1 customer right there
 
8:19 PM
lol I try when I can.
 
I think it's more you can when you try ;)
 
@J.D. I don't use it, but seen enough to know it isn't really a security feature, more of a convenience for querying the correct rows without having to add extra code.
Don't expect it to stand up to someone with permission to write raw SQL queries, or to able to accurately time query performance. There are loads of side-channel attacks.
@J.D. Sounds interesting, what was it roughly?
 
@Charlieface 100% this.
 
8:37 PM
@SeanGallardy The real bugbear is that it has side-channel attacks even with indirect webserver calls to parameterized queries. Then again, I often see login pages with no side-channel protection for index lookups on usernames and password hashes, so it's the least of our worries.
 
8:50 PM
@Charlieface Fo' sho'. Even worse, I've worked with databases that stored logins with plaintext passwords in a field called Password lol. And there were some higher level credentials that really shouldn't've been stored in there. (Fortunately not where I'm at these days.)
 
@J.D. And here I am today trying to make head or tail of AWS authentication procedure, which uses SHA256 hashing no less than 6 times one after the other, to no apparent reason.
 
@Charlieface To try to be succinct: 1. Create an RLS predicate on a TVF on a Table. 2. Create a View on that Table. 3. Create a separate RLS policy and predicate on the View using a separate TVF that selects from that Table with a WHERE clause. 4. When selecting from the View, I noticed that WHERE clause in that second RLS predicate's TVF isn't being applied.
 
9:07 PM
On a completely unrelated topic, is there a way to actually stop automatic restarts from Windows updates in Windows 11? I legit had it restart 5 days in 1 week for updates lol.
 
CMD window for /L %n in (1,0,10) do (timeout /t 55 & shutdown -a) cancels the shutdown every 55 seconds
 
hah, outside the box solution
 
I actually use that often because our endpoint management forces a restart once a week, and I'm often running stuff overnight.
0
Q: Can we use SQL Server STRING_AGG() in queries using GROUP BY ROLLUP ()

Keith FearnleyTL;DR: Can we use STRING_AGG aggregate function in a query using GROUP BY ROLLUP(...)? I am converting old code that used the STUFF( (SELECT...FOR XML) ) approach, so that it uses STRING_AGG instead. I cannot seem to use STRING_AGG when using ROLLUP, as I get the error Aggregate functions that ...

@SeanGallardy Why on earth can't STRING_AGG merge sub-aggregates?
 
9:50 PM
Sean doesn’t not dilly nor dally with query business
You’re better off asking Ol’ Debuggy
 
@Charlieface Haha what Erik said, unless it's super interesting and has more to do with the OS side. I would venture a guess of "someone didn't program it in"
 
@ErikDarling he's not here tho
 
@PaulWhite au revoir
 
@ErikDarling In what accent does Debuggy rhyme with White
 
10:09 PM
Maybe something Caribbean
Anyway, his real name is Pablo Blanco. Don’t tell anyone outside chat though.
 
I think there was some mod interferences in the star elections, I no longer see my Tabs > Spaces item in the list.
 
@ErikDarling In SHA256 for Charlie: bc4a63fcc4ab5731d3906e2debdd4fcd4ec698037d535ecf905dd1ee989f350e
@SeanGallardy Btw I starred you on that, since it was a very sensible remark, that somehow survived removal.
 
@SeanGallardy mods aren’t always wrong
 
@ErikDarling Mods were a mistake
@J.D. This man gets it.
 
@SeanGallardy I knew I liked ya for a reason.
 
10:23 PM
@ErikDarling I believe the word is "ever"
 
I seriously don't understand why someone would press the spacebar 4 times when they could press tab once
 
who presses 4 spaces? not i. I simply hit tab et voila. 4 spaces.
2
 
then why use spaces if you already press tab!?
 
because spaces aren't fucked
 
"I press tab because I prefer spaces"
lol
 
10:24 PM
no seriously though, spaces never show up weirdly in any editor
tabs make stuff look very weird all the time
 
I use two editors, notepad and VS. Never have any issues with tabs. I could see if you're using 50 different editors but also then you're doing it wrong :D
 
a pet peeve is opening project code and seeing stuff that just doesn't line up
 
I hear you there
 
what if you have 50 different devs using 50 different editors?
then what, huh?
 
Just wait until I open the can of worms that is curly braces
Clearly you have no standards so your devs probably suck
 
10:26 PM
I love me some curly braces in all the right places
2
 
L O L
 
@SeanGallardy don't tell them that
 
OoO curly braces and parenthesis being inlined drives me crazy
 
they are so important they deserve their own line
 
I'm all about:

MyFunc(stuff)
{
}
 
10:27 PM
with a CRLF at the end
yah
 
None of this:

JFCWTF(stuff){
}
 
A la the latest C# style I've been seeing where they're unaligned.
 
@SeanGallardy 100000% true
 
We have some common ground, curly braces
 
devs hate pressing any extra ~~extra~~ keys
 
10:28 PM
lol
 
god I hate chat markdown
6
it's almost as bad as tabs
 
@HannahVernon I don't think I saw a consensus on where Replication falls into the mix though.
 
10:47 PM
@HannahVernon Three hyphens/dashes
It puzzles me Sean can be so smart about some things while being monumentally dumb about others e.g. tabs
Like it's intellectually impossible to see the difference between pressing the tab key and having tab characters embedded
How would that even be possible. It's just madness
Maybe this is what working with Availability Groups does to a person
 
tabs are landmines
just hanging out, formatting some code
delete three spaces, delete the fourth and your cursor jumps back a full tab stop
now you're an odd number out and you have to hit space again
 
Worse than replication
 
set yourselves free, sheeple
 
@PaulWhite cool, I respect it, even though I disagree.
 
Ok they're equally bad
 
11:01 PM
Thanks for the cheese tonight btw
 
Looks Photoshopped
@CommunityNotes
 
what is this about
the point really comes across
 
Looks like a line feed carriage return problem
 
v good
 
11:13 PM
@PaulWhite actually legit went on some grilled burgers tonight. P solid.
 
@ErikDarling omg I use that all the time. Plus I use this for tabs\spaces at the end of lines: (\t| )+\r\n and replace with \r\n
I should do a blog post with all the regex replaces I use
 
@HannahVernon you should. I do a lot of cleanup like that.
 

« first day (4467 days earlier)      last day (407 days later) »