The Official Status Thread
-
@Tsaukpaetra said in The Official Status Thread:
I need to change my Mac address??!? They know you can't usually alter that, right???
https://www.howtogeek.com/192173/how-and-why-to-change-your-mac-address-on-windows-linux-and-mac/
-
@TimeBandit said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
I need to change my Mac address??!? They know you can't usually alter that, right???
https://www.howtogeek.com/192173/how-and-why-to-change-your-mac-address-on-windows-linux-and-mac/
The network cards in question aren't using IP protocols or in fact most of the Microsoft networking stack. Wouldn't work.
-
Status:
PowerShell, why the f
would you display
Inherited
as the header for the Format-Table output of a cmdlet if the actual property (for, oh, you know, filtering on) isIsInherited
?!?!?
-
@Tsaukpaetra said in The Official Status Thread:
The network cards in question aren't using IP protocols or in fact most of the Microsoft networking stack. Wouldn't work.
NetBEUI? Or is it some secret-sauce protocol?
-
@Cursorkeys said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
The network cards in question aren't using IP protocols or in fact most of the Microsoft networking stack. Wouldn't work.
NetBEUI? Or is it some secret-sauce protocol?
It installs some eBUS filter driver on the card. Windows only tries to send packets and never receives any.
-
status: laughing internally because the Ginger bitch doesn't like it when I make licking noises against her.
I guess she just doesn't like being insinuated she's dirty.
Which is strange because she also hates baths...
-
@pie_flavor I see we're in
mysql_real_escape_we_mean_it_this_time()
territory here!
-
Status: Just wrote a DllMain that examines the process name and does a bunch of library cleanup stuff during DLL unload, but if and only if a certain third-party application owns the process. Seems like a fairly mild WTF on the surface, but it goes much, much deeper. Ideally, this cleanup code would always run, but it crashes if called by our own applications from an access violation occurring deep within another third-party library. Somehow, being called by an application that is not our own prevents that access violation from occurring. And if I remove all cleanup code, it goes the other way. Our applications terminate just fine, but then this third-party application crashes on shutdown.
-
@mott555 Does your own application violate the contract somewhere?
-
@PleegWat said in The Official Status Thread:
@mott555 Does your own application violate the contract somewhere?
No idea. Every theory I've investigated on why this is happening has proven wrong. Now I've given up on understanding the problem and I'm hacking it until it goes away so I can get a fix to a customer.
-
@Cursorkeys said in The Official Status Thread:
@Tsaukpaetra I'm sure you know already; but you can stuff an x16 graphics card in anything down to a x1 lane, if the socket has a open end, and it'll still work (at reduced bandwidth) in spite of looking ridiculous hanging over the socket.
It should work, but on Intel boards those small 1x slots are likely to be routed through PCH (ex. Southbridge), which has a rather narrow bus to the CPU + it's shared with all other non-PCI-E devices (namely, storage and networking). Depends on how much data the graphics card moves. CT imaging is not gaming, sure, but if it wants gigabits of stuff per second, it will not look good at all.
I don't recall specifics of AMD boards.
-
@topspin said in The Official Status Thread:
Shit, I'm going to break everything.
Don't worry about it. You're using git; everything's already broken.
-
Status:
ModemStack/src/modem_instance.c:160:5: note: expected 'struct modem_instance_t *' but argument is of type 'struct modem_instance_t *'
-
@Cursorkeys said in The Official Status Thread:
Status:
ModemStack/src/modem_instance.c:160:5: note: expected 'struct modem_instance_t *' but argument is of type 'struct modem_instance_t *'
Java? Is that you?
-
@mott555 said in The Official Status Thread:
Status: Just wrote a DllMain that examines the process name and does a bunch of library cleanup stuff during DLL unload, but if and only if a certain third-party application owns the process. Seems like a fairly mild WTF on the surface, but it goes much, much deeper. Ideally, this cleanup code would always run, but it crashes if called by our own applications from an access violation occurring deep within another third-party library. Somehow, being called by an application that is not our own prevents that access violation from occurring. And if I remove all cleanup code, it goes the other way. Our applications terminate just fine, but then this third-party application crashes on shutdown.
I've never had the need to write a DllMain myself so I'm not sure how relevant the following is, but the title seems relevant:
(I think later articles strengthened "nothing scary" to "almost nothing at all")
-
@Tsaukpaetra said in The Official Status Thread:
@topspin said in The Official Status Thread:
Status:
$ git branch list $
$ git branch --list list * master
Shirley you wanted to create a branch named
list
, right? Why else wouldn't you specify a flag in a mixed syntax like that?Exactly!
And it's easy to remember, too! How do you list your stashes again?
git stash list
.
Where's the--list
there? Hooray for consistency.
-
@topspin Instead they should take a page out of
rm
's book. Typinggit branch list
should have the same result asgit branch --list
, and if you want the old behaviour, you would need something likegit branch list --yes-i-really-want-to-create-a-branch-called-list
-
@hungrier said in The Official Status Thread:
@topspin Instead they should take a page out of
rm
's book. Typinggit branch list
should have the same result asgit branch --list
, and if you want the old behaviour, you would need something likegit branch list --yes-i-really-want-to-create-a-branch-called-list
Or take a page out of their own book:
git stash [push] ...
Calling git stash without any arguments is equivalent to git stash push.Make
list
a subcommand (or whatever these are called) and usegit branch [create] <branchname>
to create a branch, wherecreate
is only optional if the branch name doesn't conflict with keywords.
Or at least something that's consistent.
-
@hungrier said in The Official Status Thread:
Instead they should take a page out of rm's book.
Yes.
rm -f git
-
@Cursorkeys said in The Official Status Thread:
Status:
ModemStack/src/modem_instance.c:160:5: note: expected 'struct modem_instance_t *' but argument is of type 'struct modem_instance_t *'
I'd made a circular reference by including a pointer to a
modem_instance_t
in a function pointer withinmodem_instance_t
, that's what seemed to be doing it.
-
@topspin said in The Official Status Thread:
git branch [create] <branchname>
I never use that.
I use
git checkout -b <branchname>
to create a new branchI only use
git branch
to list branches
-
-
@TimeBandit said in The Official Status Thread:
I only use
git branch
to list branchesI apparently only use it to create list branches.
-
@topspin said in The Official Status Thread:
@mott555 said in The Official Status Thread:
Status: Just wrote a DllMain that examines the process name and does a bunch of library cleanup stuff during DLL unload, but if and only if a certain third-party application owns the process. Seems like a fairly mild WTF on the surface, but it goes much, much deeper. Ideally, this cleanup code would always run, but it crashes if called by our own applications from an access violation occurring deep within another third-party library. Somehow, being called by an application that is not our own prevents that access violation from occurring. And if I remove all cleanup code, it goes the other way. Our applications terminate just fine, but then this third-party application crashes on shutdown.
I've never had the need to write a DllMain myself so I'm not sure how relevant the following is, but the title seems relevant:
(I think later articles strengthened "nothing scary" to "almost nothing at all")
This DLL interacts with some hardware, and some driver-y stuff needs to be properly closed and shutdown or else it may not work for the next application that uses it. Sometimes. Depending on the application. Apparently. Because some crash at shutdown if the driver shuts down cleanly but are fine if you leave things hanging, and some crash at shutdown if you leave the driver hanging but work fine if you do cleanup.
-
Status: Azure's Linux VM helpfully provides you with ephemeral storage, in the form of
/mnt
Who can use it? Why, only
root
, of course!Sub-status: Trying to write up an init systemd thing to create folders that are accessible by the target user on boot on the ephemeral storage. sigh.
-
@Tsaukpaetra said in The Official Status Thread:
Status: Azure's Linux VM helpfully provides you with ephemeral storage, in the form of /mnt
Who can use it? Why, only root, of course!You can mount it elsewhere, like in your home folder, where you have write access
-
@PleegWat said in The Official Status Thread:
@PleegWat said in The Official Status Thread:
Status: The power is out.
Power is restored. Was about an hour total.
Better than mine (on Wed). Went out at just after 9a. Back on at 6:30p. (They were replacing an underground vault) As I drove off I saw all the trucks - should have turned around and powered everything down. Oh well. The UPS worked for a little while... (I think I have around 30-60 minutes of power)
-
@TimeBandit said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
Status: Azure's Linux VM helpfully provides you with ephemeral storage, in the form of /mnt
Who can use it? Why, only root, of course!You can mount it elsewhere, like in your home folder, where you have write access
I don't into linux well, but that requires root, doesn't it? And when mounting, I need to mount it for a specific UID/GID or else it will take root's, right? at this point making an init script to make folders seems trivial in comparison...
-
@Tsaukpaetra said in The Official Status Thread:
@brie said in The Official Status Thread:
@Tsaukpaetra I don't think I get it.
See that dangly bit? The card that doesn't appear to be the proper size?
I missed that - I got stuck on the bend face plate...
-
@Tsaukpaetra said in The Official Status Thread:
@pie_flavor said in The Official Status Thread:
@Tsaukpaetra it means the drivers are busted, obviously.
Stupid questions first - you know you've got two different devices open, right?They are two different devices. Identical devices. One loads the driver. The other does not.
Cheap knockoff Chinese cards - and the MAC is the same?
-
@Tsaukpaetra said in The Official Status Thread:
I don't into linux well, but that requires root, doesn't it?
No. You need root to mount it in /mnt because /mnt is owned by root
if you want it mounted at boot, just add it to fstab
https://www.howtogeek.com/howto/38125/htg-explains-what-is-the-linux-fstab-and-how-does-it-work/
-
@Cursorkeys said in The Official Status Thread:
Status: Trying to read some C code that implements a dynamic plugin system. Feel like my brain is trying to escape. I didn't even know you could do some of these things.
Edit: Like this thing, I can't quite work it out at the moment
enum processing_type { UNEXPECTED = -1, UNKNOWN = 0, INTERMEDIATE _ENUM_SIZE_INT32 = INT32_MAX }; typedef enum processing_type (*processor_t)(const char *data, size_t length, void *parent);
Edit2: Oh, right. It's a virtual method kinda thing.
It's just a function prototype. So you can easily do:
processor_t var = (processor_t)GetProcAddress(...) var(data, -1, nullptr);
-
@TimeBandit said in The Official Status Thread:
No. You need root to mount it in /mnt because /mnt is owned by root
But
/mnt
is the mount. They didn't mount it in/mnt
, it flat-out is/mnt
.I'm not sure if that makes a difference.
I'm also not sure if this installation even pays attention to
fstab
, it's pwned by systemd after all...
-
@dcon said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
@pie_flavor said in The Official Status Thread:
@Tsaukpaetra it means the drivers are busted, obviously.
Stupid questions first - you know you've got two different devices open, right?They are two different devices. Identical devices. One loads the driver. The other does not.
Cheap knockoff Chinese cards - and the MAC is the same?
Nope. Apparently a borked installation of some packet filter driver that wanted to attach to all devices.
-
@Cursorkeys said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
Status: If this is the average expected experience expected for medical equipment, I never want to touch it again. Fuck this piece of shit!
I'm kind of surprised you're even allowed to. As a radiation-emitting device isn't the whole thing, including the PC, covered by the CDRH licence?
Edit: I only know the CDRH exists because my huge ion lasers had CDRH stickers, with dire warnings, all over them.
I am 98% certain he shouldn't be touching it. At all.
-
@Weng said in The Official Status Thread:
@Cursorkeys said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
Status: If this is the average expected experience expected for medical equipment, I never want to touch it again. Fuck this piece of shit!
I'm kind of surprised you're even allowed to. As a radiation-emitting device isn't the whole thing, including the PC, covered by the CDRH licence?
Edit: I only know the CDRH exists because my huge ion lasers had CDRH stickers, with dire warnings, all over them.
I am 98% certain he shouldn't be touching it. At all.
I'm not touching the x-ray part. I'm touching the controller part.
-
@Tsaukpaetra said in The Official Status Thread:
@Weng said in The Official Status Thread:
@Cursorkeys said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
Status: If this is the average expected experience expected for medical equipment, I never want to touch it again. Fuck this piece of shit!
I'm kind of surprised you're even allowed to. As a radiation-emitting device isn't the whole thing, including the PC, covered by the CDRH licence?
Edit: I only know the CDRH exists because my huge ion lasers had CDRH stickers, with dire warnings, all over them.
I am 98% certain he shouldn't be touching it. At all.
I'm not touching the x-ray part. I'm touching the controller part.
Same shit. Like, real clinical IT professionals don't even like PATCHING that stuff
-
@Weng said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
@Weng said in The Official Status Thread:
@Cursorkeys said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
Status: If this is the average expected experience expected for medical equipment, I never want to touch it again. Fuck this piece of shit!
I'm kind of surprised you're even allowed to. As a radiation-emitting device isn't the whole thing, including the PC, covered by the CDRH licence?
Edit: I only know the CDRH exists because my huge ion lasers had CDRH stickers, with dire warnings, all over them.
I am 98% certain he shouldn't be touching it. At all.
I'm not touching the x-ray part. I'm touching the controller part.
Same shit. Like, real clinical IT professionals don't even like PATCHING that stuff
And I now join them.
It's a buncha fuckin' files, could kill the devs to provide more information than
Error N° 1025 [Ok]
??? Or a log? Something?Oh yeah, and to update on that: Apparently turning the computer off and turning it back on again can corrupt the Ethernet drivers (or at least, that's what the problem was as relayed to me).
Also that Antivirus software MUST NOT be installed on the computer.
Additionally, there MUST NOT exist a screensaver on the computer.
Because, I guess, at one point back in '96 a screensaver was something that broke their software and now screensavers are forever banned.
-
@Cursorkeys said in The Official Status Thread:
Status:
ModemStack/src/modem_instance.c:160:5: note: expected 'struct modem_instance_t *' but argument is of type 'struct modem_instance_t *'
That's one of the more gnostic errors from C. It's cause is not properly declaring the structure
modem_instance_t
before using it, which results in you ending up with multiple versions of the same structure that only differ in (internal to the compiler, not visible to you) struct ID.The fix might be to do a forward declaration of the struct, especially if you have a circular definition somewhere.
-
@Tsaukpaetra No, you're missing it. It's 'Error No', as in there's no error.
-
@pie_flavor said in The Official Status Thread:
@Tsaukpaetra No, you're missing it. It's 'Error No', as in there's no error.
Except, that's literally the title bar text of the messagebox, but if you want to interpret it that way, you're free to hold that opinion.
-
@Tsaukpaetra 500 OK.
-
Status: wet buggo.
-
@JazzyJosh said in The Official Status Thread:
Embezzle it all.
That implicates there is something to embezzle ...
-
@Jaloopa said in The Official Status Thread:
Status: enjoying some old fashioned front page trolling
Link for posterity
-
Status: my favourite thing about drawing on paper is how easily I can move layers around if something's a little bit too high
-
@dkf said in The Official Status Thread:
forward declaration
TIL, thank you. I've eliminated the circular reference for now, it didn't really need the whole struct anyway.
-
@Cursorkeys said in The Official Status Thread:
TIL, thank you.
It's to do with the declaration context for function arguments not being the same as the declaration context for the whole file and I never understood who could possibly have thought that that was a good idea. It's almost like it is a compiler bug that was elevated to being part of the language specification…
-
You know, for all the tech stuff that seems to break on me for no reason, my old Toshiba HDD and Samsung SSD have been in use for a decade now and haven't given me any problems.
-
Status: I should probably get around to doing some work instead of browsing WTDWTF at the office all day.