WTF Bites
-
without giving an easy-to-find "go back" option isn't very endearing.
Haven't been using software for very long, have you?
-
uncheck Hide system window titlebar
Here’s a WTF:
It buggers up the tab label colour, but only when the window is focused
Anyway despite apparently being on 115.2.2 I don’t seem to have the outlook display at all. Maybe a distro thing?
-
I love GraphQL. It was a great idea to use it for reports.
-
There are two reports. After many hours of requirement-gathering both of them gained a "group by X" checkbox, but in one report it multiplies the number of rows by the number of Xses (like a
GROUP BY
clause), in the other one it does the opposite. "It wouldn't be intuitive" otherwise. OK.
-
@sebastian-galczynski said in WTF Bites:
I love GraphQL. It was a great idea to use it for reports.
API with crappy errors is API with crappy errors whether it is GraphQL or OpenAPI or SOAP or whatever.
-
@sebastian-galczynski said in WTF Bites:
I love GraphQL. It was a great idea to use it for reports.
API with crappy errors is API with crappy errors whether it is GraphQL or OpenAPI or SOAP or whatever.
But that's the problem with all these magical wrappers. The more wrappers, the more information you lose. In this case, the actual problem was that my query was returning interval instead of integer from the db.
-
@Tsaukpaetra said in WTF Bites:
@pcooper said in WTF Bites:
without giving an easy-to-find "go back" option isn't very endearing.
Haven't been using software for very long, have you?
Oh, I have. None of it is very endearing, though.
-
@sebastian-galczynski said in WTF Bites:
@sebastian-galczynski said in WTF Bites:
I love GraphQL. It was a great idea to use it for reports.
API with crappy errors is API with crappy errors whether it is GraphQL or OpenAPI or SOAP or whatever.
But that's the problem with all these magical wrappers. The more wrappers, the more information you lose. In this case, the actual problem was that my query was returning interval instead of integer from the db.
It seems to me rather like a lack of magic in the wrapper. If it was properly magic, it would either just generate its schema from the query, or do a proper type-check during build. But it seems to be duck typed instead as is therefore summoning ghosts.
-
Just seen in an email "sorry not trying to solution this via email."
If only there were a verb that means "to solve."
-
@error I WASN'T ASKING TO SOLUTION THIS
-
@boomzilla If you're not part of the solution, you're part of the precipitate.
-
@HardwareGeek Don't keep me in suspension.
-
Just seen in an email "sorry not trying to solution this via email."
If only there were a verb that means "to solve."
There is: “to solutionize” - this is the short form.
-
Just seen in an email "sorry not trying to solution this via email."
If only there were a verb that means "to solve."
There is: “to solutionize” - this is the short form.
I kinda like invoke a solutionate instance.
-
@Tsaukpaetra surely that produces a different result if it is solutionate rather than solutionized? What about solutionite, solutionide?
I presume it relates to the amount of hot air in discussion?
-
@Tsaukpaetra surely that produces a different result if it is solutionate rather than solutionized? What about solutionite, solutionide?
I presume it relates to the amount of hot air in discussion?
Invoking the solutionate resolves an instance of solutionerator which compiles a collection of solutionite to perform the solutionizing function, after which the solutionate returns a solventation package which can be used in conjuction with solutionide to decouple the solutionite components into a partial solution.
-
In the end, I think the issue is
unsolvableundissolvableundissolutionable
-
It seems to me rather like a lack of magic in the wrapper. If it was properly magic, it would either just generate its schema from the query, or do a proper type-check during build.
It did check types during build (typescript), but it didn't know what type the query returns. And after compilation it's just JS, so the value with the wrong type was assigned with no error and passed further to GraphQL. AFAIK strongly typed database queries are available only in Rust, and they require a functioning database (with data) during build. In other languages the exception would probably be thrown during some cast or assignment, but still not during build.
-
@sebastian-galczynski in an alternate timeline, people in our industry actually care about API schemas and the solution to your problem is trivial.
-
@Tsaukpaetra said in WTF Bites:
@Tsaukpaetra surely that produces a different result if it is solutionate rather than solutionized? What about solutionite, solutionide?
I presume it relates to the amount of hot air in discussion?
Invoking the solutionate resolves an instance of solutionerator which compiles a collection of solutionite to perform the solutionizing function, after which the solutionate returns a solventation package which can be used in conjuction with solutionide to decouple the solutionite components into a partial solution.
These are all valid solutionings going forward.
-
*reads the last few posts*
Ok guys, stop it now, or I'm rebooting every one of you.
-
I'm pretty sure selling people is illegal.
-
@Zerosquare said in WTF Bites:
*reads the last few posts*
Ok guys, stop it now, or I'm rebooting every one of you.
But then you'll distill out all of the good jokes.
-
@Zerosquare said in WTF Bites:
*reads the last few posts*
Ok guys, stop it now, or I'm rebooting every one of you.
But then you'll distill out all of the good jokes.
Distillation will not produce something that wasn't already present.
-
@HardwareGeek said in WTF Bites:
@Zerosquare said in WTF Bites:
*reads the last few posts*
Ok guys, stop it now, or I'm rebooting every one of you.
But then you'll distill out all of the good jokes.
Distillation will not produce something that wasn't already present.
Sounds like someone's got an excess of saline in solution.
-
@sebastian-galczynski said in WTF Bites:
It seems to me rather like a lack of magic in the wrapper. If it was properly magic, it would either just generate its schema from the query, or do a proper type-check during build.
It did check types during build (typescript), but it didn't know what type the query returns. And after compilation it's just JS, so the value with the wrong type was assigned with no error and passed further to GraphQL. AFAIK strongly typed database queries are available only in Rust, and they require a functioning database (with data) during build. In other languages the exception would probably be thrown during some cast or assignment, but still not during build.
Strongly typed queries are available in many languages if you are using some kind of ORM. And my experience with GraphQL is that the frameworks are usually designed to work with those ORMs, in our specific case with .NET Entity Framework.
Now in this case you have the tables declared as types in the ORM framework, so it does not need an instance of the database to check the types, but you can get a desync between the code schema and the sql schema if you write the upgrades by hand rather than generate them.
Other frameworks do need an instance of the database, but these days it really should be a problem to spin up an instance of any database with the right schema and no or minimal data on the development workstation or build server for the purpose.
-
@Zerosquare said in WTF Bites:
Ok guys, stop it now, or I'm rebooting every one of you.
Also having read the last few posts, I'm wondering if you could settle for a half-assed reboot? Like, one that emphasizes the shutting down phase and then kinda forgets () about the subsequent starting again part?
-
@Tsaukpaetra said in WTF Bites:
@Tsaukpaetra surely that produces a different result if it is solutionate rather than solutionized? What about solutionite, solutionide?
I presume it relates to the amount of hot air in discussion?
Invoking the solutionate resolves an instance of solutionerator which compiles a collection of solutionite to perform the solutionizing function, after which the solutionate returns a solventation package which can be used in conjuction with solutionide to decouple the solutionite components into a partial solution.
These are all valid solutionings
goingforwardizing.TFY
-
Strongly typed queries are available in many languages if you are using some kind of ORM
Oh, of course. That was one of the places where I needed to bypass the ORM, because the query itself consists of a bunch of complicated sub-selects and aggregations, the results of which don't represent entities.
-
WTF Bite: I have an old MS Visio diagram on my local disk that I'd like to look at, but I don't have Visio installed. Figured maybe some other program I have installed can at least view this, so I double click it.
In response, Windows opens
ChrEdge with the URL set tofile:///C:/Abs/path/to/MyDiagram.vsd
. This shows a blank page, with the "Downloads" popup open with a warning:MyDiagram.vsd could harm your device. Do you want to keep it anyway?
Keep Delete
I click "Keep", then the little "Open file" link in the popup. This of course causes Edge to navigate to the same URL it already was, start a new "download", and we're back at the "warning" popup...
Usually I like to indulge in some righteous anger when computers do stupid shit, but this is so absurd that I couldn't help it but laugh out loud.
-
@sebastian-galczynski said in WTF Bites:
It did check types during build (typescript), but it didn't know what type the query returns. And after compilation it's just JS, so the value with the wrong type was assigned with no error and passed further to GraphQL. AFAIK strongly typed database queries are available only in Rust, and they require a functioning database (with data) during build. In other languages the exception would probably be thrown during some cast or assignment, but still not during build.
The problem is that if you put actually strong types in (handling all the various types that could actually turn up) then people throw a strop because their code now has to handle lots of edge cases that their test dataset doesn't have an example of and that might never actually crop up in the production dataset either (but it's hard to be sure).
At a higher level, you've got another concern too: is the sampling of the space that you're using to build the data model to start out with representative of the whole dataset? That's a very messy problem indeed.
-
@ixvedeusi said in WTF Bites:
WTF Bite: I have an old MS Visio diagram on my local disk that I'd like to look at, but I don't have Visio installed. Figured maybe some other program I have installed can at least view this, so I double click it.
In response, Windows opens
ChrEdge with the URL set tofile:///C:/Abs/path/to/MyDiagram.vsd
. This shows a blank page, with the "Downloads" popup open with a warning:MyDiagram.vsd could harm your device. Do you want to keep it anyway?
Keep Delete
I click "Keep", then the little "Open file" link in the popup. This of course causes Edge to navigate to the same URL it already was, start a new "download", and we're back at the "warning" popup...
Following you so far, except my popup was "Hey, use Internet Exploder!"
Also, "Turn onz Activsex!"
Sadly, my document isn't a real Visio document...
Usually I like to indulge in some righteous anger when computers do stupid shit, but this is so absurd that I couldn't help it but laugh out loud.
Yeah that association should not exist if you don't have at least the Viewer installed...
-
@HardwareGeek said in WTF Bites:
I'm pretty sure selling people is illegal.
It’s fairly normal for at least a few Celtics each year to be sold. Mostly for a lot more than $10, but some entirely for free.
-
I'm with you. At least you can beat it back into some semblance of its former self:
I'm quite amazed that their what's new highlights specifically says things like "perfect for people used to modern webmail", "emulates a mobile interface" and, like, do they just not realize that people might be using their product specifically to avoid those things?
I especially like how it says "Upgrade to version 115 and experience the future of Thunderbird! " and yet there is no obvious download link to be found anywhere on that page.
-
@HardwareGeek said in WTF Bites:
I'm pretty sure selling people is illegal.
At one time, AltaVista disagreed.
(NSFW)
-
@HardwareGeek said in WTF Bites:
I'm pretty sure selling people is illegal.
It’s fairly normal for at least a few Celtics each year to be sold. Mostly for a lot more than $10, but some entirely for free.
See also: @GuyWhoKilledBear's Law Celtics.
-
Just seen in an email "sorry not trying to solution this via email."
If only there were a verb that means "to solve."
There is: “to solutionize” - this is the short form.
If only I could shoot you out of a cannon.
-
@Tsaukpaetra said in WTF Bites:
@Tsaukpaetra surely that produces a different result if it is solutionate rather than solutionized? What about solutionite, solutionide?
I presume it relates to the amount of hot air in discussion?
Invoking the solutionate resolves an instance of solutionerator which compiles a collection of solutionite to perform the solutionizing function, after which the solutionate returns a solventation package which can be used in conjuction with solutionide to decouple the solutionite components into a partial solution.
@Arantor look at what you have brought down upon us. Are you proud?
-
@Gern_Blaanston said in WTF Bites:
I especially like how it says "Upgrade to version 115 and experience the future of Thunderbird! " and yet there is no obvious download link to be found anywhere on that page.
Well, if you're upgrading to a new version of Thunderbird the implication is kind of there that you already have a previous version, which is capable of doing the download and install itself.
(There are three clearly-marked download buttons on the site's front page. Getting there from the "what's new" page is problematic, as that link is waaaaaay down the bottom of a a lot of empty space and eleven lines of BIG TEXT.)
-
@Tsaukpaetra said in WTF Bites:
@Tsaukpaetra surely that produces a different result if it is solutionate rather than solutionized? What about solutionite, solutionide?
I presume it relates to the amount of hot air in discussion?
Invoking the solutionate resolves an instance of solutionerator which compiles a collection of solutionite to perform the solutionizing function, after which the solutionate returns a solventation package which can be used in conjuction with solutionide to decouple the solutionite components into a partial solution.
@Arantor look at what you have brought down upon us. Are you proud?
Inciting a little anarchy? Always.
-
Ok, now I have a bizarre dns problem:
me@localhost:~$ nslookup otherhost.local Server: 192.168.5.1 Address: 192.168.5.1#53 Name: otherhost.local Address: 192.168.5.222 me@localhost:~$ nslookup otherhost Server: 192.168.5.1 Address: 192.168.5.1#53 Name: otherhost.local Address: 192.168.5.222 me@localhost:~$ ping otherhost PING otherhost.local (192.168.5.222) 56(84) bytes of data. 64 bytes from otherhost.local (192.168.5.222): icmp_seq=1 ttl=64 time=2.10 ms 64 bytes from otherhost.local (192.168.5.222): icmp_seq=2 ttl=64 time=142 ms ^C --- otherhost.local ping statistics --- 2 packets transmitted, 2 received, 0% packet loss, time 1001ms rtt min/avg/max/mdev = 2.097/71.941/141.785/69.844 ms me@localhost:~$ ping otherhost.local ping: otherhost.local: Name or service not known me@localhost:~$ cat /etc/resolv.conf # Generated by NetworkManager search local nameserver 192.168.5.1 nameserver fdf0:9802:e046::1 options ndots:1
Apparently otherhost can't be reached unless I use the search feature, despite the hostname being resolved? What is going on here? The problem only occurs on my desktop, other hosts in the network (android phones included) can reach both otherhost and otherhost.local the same.
Edit: Ok, I just learned about nsswitch.conf and mDNS, and that '.local' is a magic word for it. Changed to something else, works fine. Thx Apple for infecting normal OSes with your crap
-
@sebastian-galczynski said in WTF Bites:
Edit: Ok, I just learned about nsswitch.conf and mDNS, and that '.local' is a magic word for it. Changed to something else, works fine. Thx Apple for infecting normal OSes with your crap
Oh good you found out about it.
It's one of the reasons I'm begrudgingly looking into how to rename a Active Directory domain from
office.local
to something else slightly more cromulent.
-
@Tsaukpaetra said in WTF Bites:
rename a Active Directory domain from
office.local
to something else slightly more cromulent.why not
office.lan
?
-
@TimeBandit said in WTF Bites:
@Tsaukpaetra said in WTF Bites:
rename a Active Directory domain from
office.local
to something else slightly more cromulent.why not
office.lan
?That would be "slightly more cromulent", yes.
-
@Tsaukpaetra said in WTF Bites:
@TimeBandit said in WTF Bites:
@Tsaukpaetra said in WTF Bites:
rename a Active Directory domain from
office.local
to something else slightly more cromulent.why not
office.lan
?That would be "slightly more cromulent", yes.
That's what I used. Since they introduced the fancy TLDs, there are very few options. The most interesting is
.invalid
-
@sebastian-galczynski said in WTF Bites:
@Tsaukpaetra said in WTF Bites:
@TimeBandit said in WTF Bites:
@Tsaukpaetra said in WTF Bites:
rename a Active Directory domain from
office.local
to something else slightly more cromulent.why not
office.lan
?That would be "slightly more cromulent", yes.
That's what I used. Since they introduced the fancy TLDs, there are very few options. The most interesting is
.invalid
Microsoft recommended best practice is to use your public domain (business.com). Of course, those of us using completely legitimate MSDN keys to run a domain in our home network
need to be committedshould just embrace the future and switch to Windows home with always online MS accounts
-
MSDN keys
You don't need to activate Windows Server (apparently) and Pro isn't that much more expensive.
-
@Gern_Blaanston said in WTF Bites:
I especially like how it says "Upgrade to version 115 and experience the future of Thunderbird! " and yet there is no obvious download link to be found anywhere on that page.
There are three clearly-marked download buttons on the site's front page. Getting there from the "what's new" page is problematic
For a company in the Internet / Web browser business their websites have always been poorly designed shit.
-
Just seen in an email "sorry not trying to solution this via email."
If only there were a verb that means "to solve."
There is: “to solutionize” - this is the short form.
If only I could
shoot you out of a cannoncannonicalize you.
-
@sebastian-galczynski said in WTF Bites:
Since they introduced the fancy TLDs, there are very few options. The most interesting is .invalid
The closest there is now to an "official" internal-use-only name is
.home.arpa
, though I don't understand (nor do I really want to) the committee-thinking that didn't define a similar name more suited for businesses.