If you’ve ever looked at windows desktop application source code (shudder) they have all this wacky shit around event handling in their native C/C++ APIs, which is a lot less clean that what you typically see in other languages / platforms which do similar event handling. It wouldn’t surprise me if that shit is hacky and buggy as fuck on the OS side of that implementation too.
So this would be your ctrl-c press event, and whatever handler code.
Though it’s been like 20 years since I looked at some of this. Maybe it’s better now, but somehow I doubt it.
Because your company makes you use Azure Virtual Desktop to connect to a VM and AVD’s clipboard handling is as unreliable as anything else from Microsoft?
I have to connect to remote PC then from that connect to another remote PC then from that connect to another remote PC and then finally log in. I have to enter the same password like six times. It’s a miracle anything syncs across that mess
In my case it’s mostly that browsers and as such also electron Apps unfocus textboxes when the window is unfocused for seemingly no reason, sometimes without giving visual indication meaning that I need to reselect it to copy it.
My C and V key are literally the most worn keys on my keyboard. The letters are completely worn off, they’re by far my most used keys. That amount of use probably makes the contacts degrade and less reliable.
Windows 10 (I think?) introduced a “fancy clipboard” with history and multi paste and generally made it an “app”. In my experience the fancier it is the more failure points, since it’s no longer just a simple OS memory buffer, it has its own application logic.
Since it’s a global hotkey, the OS has to:
Handle the interrupt
Query the currently active window / application for a “copy” output of the currently highlighted context
Wait for a response
Place that response into a buffer of some sort
In a highly volatile asynchronous UI where the user can alt tab at any time, I imagine often what happens is that you just didn’t wait long enough for the data to actually copy. Pressing CTRL+C would be equivalent in reliability as waiting 2 seconds, but it’s more fun to press the button a bunch
For better or worse, usb devices need to be polled by the computer all the time, instead of delivering interrupts at the moment of input events. I’m baffled by this fact, but am not versed in hardware design well enough to weigh pros and cons. I assume that the polling is handled by the usb controller, though.
Query the currently active window / application for a “copy” output of the currently highlighted context
Place that response into a buffer of some sort
To my vague knowledge, the various GUIs all request content from the owner of the selection at the moment of pasting, i.e. ctrl-v — presumably both to avoid storing large swathes of pictures and such, and to allow conversion to the target format on the fly. However, iirc some environments require defining the available formats when content is copied, so idk how this meshes with that. I’m not a desktop programmer, so not entirely sure. Perhaps it’s just to simplify copying for the programmer of the source app, instead of exposing the whole innards.
In a highly volatile asynchronous UI where the user can alt tab at any time, I imagine often what happens is that you just didn’t wait long enough for the data to actually copy.
Properly, any input event would be delivered to the app (or rather the UI element) that was active at the time of the event — and presumably the OS UI runs at a higher priority for that. However, in practice it seems to be a toss-up as to whether it really works like this.
No that’s not what I mean (plus it doesn’t work most of the time, either). It’s the copying that often fails for me, which I only notice when I paste and realize “wait a minute, these are the previous contents of the clipboard.”
Nah that’s not at all how that hand movement works. It’s always rest on Ctrl and then find C for me. It would also type “C” over the highlighted text, which would be very noticeable.
FACT.
Does anyone here know why it’s unreliable?
It wasn’t always like this.
I swear every version of Windows between 3.11 and XP was 100% reliable. Then it just started… slipping.
What’s weird is that it often works on the second try.
If you’ve ever looked at windows desktop application source code (shudder) they have all this wacky shit around event handling in their native C/C++ APIs, which is a lot less clean that what you typically see in other languages / platforms which do similar event handling. It wouldn’t surprise me if that shit is hacky and buggy as fuck on the OS side of that implementation too.
So this would be your ctrl-c press event, and whatever handler code.
Though it’s been like 20 years since I looked at some of this. Maybe it’s better now, but somehow I doubt it.
Microsoft screwing up something as simple as an event listener is peak Microsoft
Because your company makes you use Azure Virtual Desktop to connect to a VM and AVD’s clipboard handling is as unreliable as anything else from Microsoft?
I have to connect to remote PC then from that connect to another remote PC then from that connect to another remote PC and then finally log in. I have to enter the same password like six times. It’s a miracle anything syncs across that mess
That’s a sensible theory for when this happens at work (it’s Citrix in my case).
I’d swear it happens at home on Windows 10 as well, though. Maybe I am pressing it wrong sometimes. Either way, the meme is spot on for me.
In my case it’s mostly that browsers and as such also electron Apps unfocus textboxes when the window is unfocused for seemingly no reason, sometimes without giving visual indication meaning that I need to reselect it to copy it.
I have unconfirmed theories.
For better or worse, usb devices need to be polled by the computer all the time, instead of delivering interrupts at the moment of input events. I’m baffled by this fact, but am not versed in hardware design well enough to weigh pros and cons. I assume that the polling is handled by the usb controller, though.
To my vague knowledge, the various GUIs all request content from the owner of the selection at the moment of pasting, i.e. ctrl-v — presumably both to avoid storing large swathes of pictures and such, and to allow conversion to the target format on the fly. However, iirc some environments require defining the available formats when content is copied, so idk how this meshes with that. I’m not a desktop programmer, so not entirely sure. Perhaps it’s just to simplify copying for the programmer of the source app, instead of exposing the whole innards.
Properly, any input event would be delivered to the app (or rather the UI element) that was active at the time of the event — and presumably the OS UI runs at a higher priority for that. However, in practice it seems to be a toss-up as to whether it really works like this.
Thanks, that’s interesting - I hadn’t thought about the window messages responding to update the all clipboard
It’s odd to me that the paste is almost never the previous clipboard value - it’s basically always null
Windows.
Because copy and paste started bring over formatting.
If you do control shift V, it removes it and is much more reliable.
No that’s not what I mean (plus it doesn’t work most of the time, either). It’s the copying that often fails for me, which I only notice when I paste and realize “wait a minute, these are the previous contents of the clipboard.”
Probably just because you typed C + Ctrl by accident.
Nah that’s not at all how that hand movement works. It’s always rest on Ctrl and then find C for me. It would also type “C” over the highlighted text, which would be very noticeable.
Yeah, I guess it would type a C.
Maybe your keyboard lacks NKRO and your cat was sitting on 6 other keys? =P
I wish I had a cat. 🐱😍