WTF Bites
-
Status: Fucking Linux...
-
@Tsaukpaetra There are lowercase hyphens?
-
@PleegWat Yes, _
-
@Tsaukpaetra There are lowercase hyphens?
I'm more concerned about lowercase numbers.
-
We could make the same joke with ₀ through ₉, but ⅰ through ⅻ are actual lowercase numerals.
You should definitely trying using them.
-
-
@Vault_Dweller said in WTF Bites:
@PleegWat Yes, _
That's actually uppercase hyphen. See, Shift- types _.
-
@Tsaukpaetra said in WTF Bites:
Status: Fucking Linux...
Why in the $solar_system do you want uppercase letters in a username‽
-
-
@Tsaukpaetra said in WTF Bites:
Ubuntu
I don't think furry ferret is officially recognized as a number yet.
-
@topspin “Furry Ferret” cannot be an Ubuntu release name, because ferrets are, in fact, furry…
-
@Tsaukpaetra said in WTF Bites:
do it yourself at Ubuntu.com
I'm not sure how I feel about being the go-to source of non-ASCII letters.
-
@Tsaukpaetra said in WTF Bites:
Status: Fucking Linux...
On my work machine, my username is
username@sub.domain.com
. Yes, some programs are extremely unhappy with that. Luckily, I haven't needed to run those. (and docker works.)
-
We could make the same joke with ₀ through ₉, but ⅰ through ⅻ are actual lowercase numerals.
You should definitely trying using them.
Sometimes I'm tempted to make a string to int parser that can understand all those stupid unicode things.
-
I'm not sure how I feel about being the go-to source of non-ASCII letters.
Oh right, because you're the only person I know who uses ą. :ro᷎llěyes: :trõllêybüs:
I actually took it from the username of, um, my good friend, um... gołąb
-
I'm not sure how I feel about being the go-to source of non-ASCII letters.
Oh. Here I was thinking @Zecc tried to create an account for Elon Musk's kid.
-
I'm not sure how I feel about being the go-to source of non-ASCII letters.
Oh right, because you're the only person I know who uses ą. :ro᷎llěyes: :trõllêybüs:
I actually took it from the username of, um, my good friend, um... gołąb
Firefox almost fully rendered that...
-
@HardwareGeek said in WTF Bites:
Why the hell would I want to do that?
It'd own the libs. They hate that.
-
-
@Tsaukpaetra said in WTF Bites:
I'm more concerned about lowercase numbers.
They look like this:
Nope. Those are the small-caps
-
@Tsaukpaetra said in WTF Bites:
Status: Fucking Linux...
Why in the $solar_system do you want uppercase letters in a username‽
I want them to be irrelevent to the condition of existence. Much the same way that periods in a Google account username are irrelevant.
-
@Tsaukpaetra Usernames, like filenames and almost anything in computing, should not be case sensitive
-
@Tsaukpaetra Usernames, like filenames and almost anything in computing, should always be case sensitive
Agreed.
-
@hungrier from the blakeyrat school of thought™
-
tl;dr: Do you want to update the
dc:title
of an asset using the HTTP API? Well, too fucking bad, our API silently changesdc:title
tojcr:title
and the only way to work around thiscritical buguseful feature is by setting up a server-side Javascript workflow tofix"sync" it.Which isn't an option for me because I need to do this with the HTTP API, and creating a new workflow with server-side Javascript would require a new story, with code reviews, etc.
-
@error I interpreted the first part to mean that it would update
jcr
as well asdc
, and thought that the was the "does not sync" part. But this is a much bigger
-
@error I interpreted the first part to mean that it would update
jcr
as well asdc
, and thought that the was the "does not sync" part. But this is a much biggerI had that same sequence of thought.
I don't often see instances of "Instead of updating the thing you told us to, instead we're going to update something completely separate and not tell you about it"
-
tl;dr: Do you want to update the
dc:title
of an asset using the HTTP API? Well, too fucking bad, our API silently changesdc:title
tojcr:title
and the only way to work around thiscritical buguseful feature is by setting up a server-side Javascript workflow tofix"sync" it.Which isn't an option for me because I need to do this with the HTTP API, and creating a new workflow with server-side Javascript would require a new story, with code reviews, etc.
Well then change it in the JCR, DC is a view over JCR.
Wait, no it isn't. Ah fuck I hate Adobe docs. They are written to say as little as possible in however long it takes to get you to give up and hire their consultants.
-
@Tsaukpaetra Usernames, like filenames and almost anything in computing, should not be case sensitive
Exactly. You should be allowed to use the case you want, knowing it won't make a difference.
-
use the case if you want to, knowing it won't make a difference.
How is this, "use"? Those tend to make differences.
-
Alternate post:
@Tsaukpaetra UsERnAmEs, LiKe fILeNamEs aNd aLMoSt aNYthinG iN COmPutiNg, sHOuLd NoT Be cAsE SeNSiTivE
-
Alternate post:
@Tsaukpaetra UsERnAmEs, LiKe fILeNamEs aNd aLMoSt aNYthinG iN COmPutiNg, sHOuLd NoT Be cAsE SeNSiTivE
Closer
-
tl;dr: Do you want to update the
dc:title
of an asset using the HTTP API? Well, too fucking bad, our API silently changesdc:title
tojcr:title
and the only way to work around thiscritical buguseful feature is by setting up a server-side Javascript workflow tofix"sync" it.Which isn't an option for me because I need to do this with the HTTP API, and creating a new workflow with server-side Javascript would require a new story, with code reviews, etc.
Well then change it in the JCR, DC is a view over JCR.
Wait, no it isn't. Ah fuck I hate Adobe docs. They are written to say as little as possible in however long it takes to get you to give up and hire their consultants.
Wow, I thought it was a straightforward issue but you're suggesting a much deeper rabbithole than I want to jump into.
-
use the case if you want to, knowing it won't make a difference.
How is this, "use"? Those tend to make differences.
Case preserving, but not case sensitive, is the optimal way. E.g. Windows will let you name your files with whatever capitalization you want, and keep it, but won't let you have files that only differ in case, and will tab-complete files regardless of the case of what you typed or the filename.
-
use the case if you want to, knowing it won't make a difference.
How is this, "use"? Those tend to make differences.
Case preserving, but not case sensitive, is the optimal way. E.g. Windows will let you name your files with whatever capitalization you want, and keep it, but won't let you have files that only differ in case, and will tab-complete files regardless of the case of what you typed or the filename.
When I see how fixing the goddamn uppercase package names the Windows-based devs introduced shakes out I'll either yell sincerely, or facetiously, about this some more.
Usually, that can be corrected easily on the case-sensitive system, then the case-preserving one goes "they're the same picture" and , then I remember alcohol.
-
@Gribnit Nuget works with any case, what's the problem?
-
@Gribnit Nuget works with any case, what's the problem?
Did you feel that?
-
shouldn't, if there's some weird trick that makes fileists hate you
See, Windows is hopelessly behind on understanding Unicode. Unlike MacOS, which does, mostly, understand it, but chose the opposite form from everybody else and is not form-preserving. That's why most of the others simply punt and abolish everything but lowercase ASCII letters, numbers, and dashes and call it a fortnight.
-
Just got bit by the nastiest bit of SQLite: instead of
CURRENT_TIMESTAMP
which would be nice, I have to useCAST(strftime('%s', 'now') AS INTEGER)
. Usingstrftime
in there is ugly but mostly works without the outer cast… until you are using it in a comparison when it silently does the wrong fucking thing (because it decides the number you pulled from your table of expiry timestamps is obviously before the current time despite really not being so) and you're left wondering through the rest of your code “what the fuck happened?”I don't want to talk about how much time and effort this has cost me.
-
Before going to extremes, many hacking victims try the usual routes to get customer service but quickly find out it seems impossible to reach someone at Facebook to help fix the problem.
"Facebook didn't have a phone number to call. There was no email to email," said Jessie Marsala, who lives outside Chicago and emailed NPR in early July about her situation.
When Marsala got hacked, she tried dialing Facebook's headquarters in Silicon Valley. But that number yields a recording that says, "Unfortunately, we do not offer phone support at this time."
Instead, Facebook tells users to report hacked accounts through its website. The site instructs them to upload a copy of a driver's license or passport to prove their identities. But the people NPR spoke with said they had trouble with every step of this automated process and wish Facebook would offer a way to reach a real person.
Brandon Sherman of Nevada City, Calif., followed a tip he found on Reddit to get his hacked account back.
"I ultimately broke down and bought a $300 Oculus Quest 2," he said. Oculus is a virtual reality company owned by Facebook but with its own customer support system.
Sherman contacted Oculus with his headset's serial number and heard back right away. He plans to return the unopened device, and while he's glad the strategy worked, he doesn't think it's fair.
"The only way you can get any customer service is if you prove that you've actually purchased something from them," he said
(A warning to anyone thinking about trying this — other Reddit users have said they tried contacting Oculus support but were unable to get their Facebook accounts restored. Also, last week, Facebook said it was temporarily halting sales of the Oculus Quest 2, which retails starting at $299, because its foam lining caused skin irritation for some customers.)
-
Just got bit by the nastiest bit of SQLite: instead of
CURRENT_TIMESTAMP
which would be nice, I have to useCAST(strftime('%s', 'now') AS INTEGER)
.More generally, the nasty bit is that sqlite lets one declare column types, but is in fact duck-typed (almost) everywhere, except it sometimes converts the values to the declared types if it works and it feels like it. Don't ever rely on those rules.
As far as timestamps go, the “recommended” form seems to be
datetime('now')
, but storing the timestamps as iso-8601 strings is obviously less efficient then storing them as integers. But you don't need to watch out for strigly numbers.
-
But you don't need to watch out for strigly numbers.
I still need to perform arithmetic (addition, comparison) on them though.
-
@dkf String comparison works correctly on iso-8601 datetimes (they are carefully designed for that) and the functions support addition looking like
datetime(something, '-2 days')
. It would make more sense though, given they already use iso-8601 datetimes, to also use iso-8601 intervals, i.e. the previous would look likedatetime(something, 'P-2D')
.The documentation also suggests this for getting the unix time:
(julianday('now') - 2440587.5)*86400.0
That is already numeric (in fact,
julianday
returns real on its own), and it includes fractional part to available accuracy of the system clock and 52-bit precision (about microseconds).
-
@dkf String comparison works correctly on iso-8601 datetimes (they are carefully designed for that) and the functions support addition looking like
datetime(something, '-2 days')
. It would make more sense though, given they already use iso-8601 datetimes, to also use iso-8601 intervals, i.e. the previous would look likedatetime(something, 'P-2D')
.Having them as strings makes other parts of the overall system worse. I want things numeric as it is both compact and easy to work with in the rest of the application. (I definitely don't need the components of the date.)
The documentation also suggests this for getting the unix time:
(julianday('now') - 2440587.5)*86400.0
How about no? That's even worse; it's now got mysterious constants in it!
-
@Zerosquare hey! I just experienced that very shortly ago! Well, except the Oculus part.
-
@Tsaukpaetra said in WTF Bites:
@Zerosquare hey! I just experienced that very shortly ago! Well, except the Oculus part.
I don't have an account.
-
@Bulb
strikes again
-
@Luhmann you're damn right I am!
Let the strike commence.
-
@Luhmann you're damn right I am!
Let the strike commence.
I am announcing a sympathy strike. I am now on a sympathy strike for the remainder of this post.
-
Status: I successfully executed the
tar
command (probably) correctly the first time!