« first day (3675 days earlier)      last day (29 days later) » 

2:47 AM
@Wildcard My objection was not about the "special file" term -- that was just terdon misconstruing my words in order to "win" the argument. Or whatever. (calling a /dev/fd/X path a "special temporary file" is atrociously innacurate, but that answer is so bad that I didn't even noticed it).
What that answer actually says (in the part I had quoted) is that in a command like "paste <(yes) <(yes) | head" first the "yes" commands are run till their end, their INFINITE outputs are SAVED in some kind of files, and after all this, the "paste" command is run with their names as arguments. Which is of course an utter absurdity.
And the first part of the answer is not much better, containing all kind of naive folklore about "files being connected to the STDIN stream" and a lot of ridiculous hedging terms.
 
3:05 AM
cool story
 
3:16 AM
That description is more like a heredoc than process substitution, though one could imagine a reading of "saves" in relation to a FIFO that implies continuous+blocking and of "then" that refers to the order commands are started
The invective speaks more to the speaker.
 
3:30 AM
Perhaps Chuck Norris was involved in shell development, so there's no issue iterating two infinite files and then pasting their output
 
3:43 AM
@AndrasDeak I see someone is up early. Or up late.
 
4:03 AM
@FaheemMitha yes...
 
@AndrasDeak One of the above, at any rate.
 
4:44 AM
@MichaelHomer "The invective speaks more to the speaker" Like what? You cannot help trying to make it personal, and bring up that "paint them as aggressive, not knowing their place" routine. As if that ever worked over the internet.
@MichaelHomer "one could imagine a reading of saves in relation to a FIFO" Yes, you can equivocate ad nauseam. And no, here-docs don't have to be implemented with temporary files, either. Yeah I'm "aggressive" and "atacking" you, again.
 
5:14 AM
@user431397 We have never previously spoken.
Or have we?
In any case, I still agree that it is at the very least poorly phrased
 
 
6 hours later…
11:33 AM
Can a RO or mod please unstar that message of mine? It's not exactly factual so I'd rather it wasn't so prominent
 
12:22 PM
@AndrasDeak done
 
Thank you
In other news, there's paranormal activity in my laptop. My usual cp -uav command that I use to pull photos from my sd card to the hdd started to copy over files already there, which should be prevented by -u. I checked the timestamps with ls and they seemed to match to the minute. Checked the files and their md5sum is the same, so it's not about files having changed. Ended up using cp -uavn.
I'm wondering if somehow daylight saving messed with something, because that still seems more likely than cp changing behaviour.
 
12:45 PM
Weird. Clutching at straws here, but does stat give the same timestamp as well? I think it gives more precision than ls.
 
I wouldn't expect already existing files to be affected by DST, neither on my hdd nor on the sd
@terdon I can check
 
Really doubt it, of course.
Probably something to do with how/where the timestamp is stored. Maybe it's being displayed the same due to your locale settings but is somehow different under the hood?
 
Ah, no, I got confused. The ls dates are one hour off. Ugh.
 
Ah, that makes more sense!
PEBKAC strikes again :)
 
Sorry :) I probably looked at a file that was copied over... Or the fact that I changed DST on my camera since changed everything.... ugh :D
$ ls -l DSC_0677.JPG
-rw-r--r-- 1 user user 10149351 Oct 27 12:05 DSC_0677.JPG
$ ls -l /media/user/NIKON\ D3200/DCIM/100D3200/DSC_0677.JPG
-rw-r--r-- 1 user user 10149351 Oct 27 13:05 '/media/user/NIKON D3200/DCIM/100D3200/DSC_0677.JPG'
ugh
Why the heck does changing DST affect existing files? Is that not weird?
this will also mess with my backups
And the problem started before I switched off DST on my camera. So entirely due to debian's time switch.
Yeah, no, I don't think I'm wrong here. If the photo was originally taken at a given datetime, that should stay that way.
 
12:56 PM
@AndrasDeak presumably, the timestamps are stored as seconds since the epoch and then displayed depending on the locale settings.
 
Nah, that would make sense...
So the issue is that my hdd has ext4 and the sd FAT which presumably stores them absolutely?
 
I would guess it's something like that, but I am really just guessing. I'm sure you know more about how this works internally than I do.
 
1:08 PM
Anyone following the SSC tuatara?
 
@terdon why would you think that? :D
In any case it might be easiest to just pull the affected 700 photos again and back them up again next time...
 
@jesse_b That's one hell of a fancy lawn mower.
The SSC Tuatara is a sports car designed, developed and manufactured by American automobile manufacturer SSC North America (formerly Shelby SuperCars Inc.). The car is the successor to the Ultimate Aero and is the result of a design collaboration between Jason Castriota and SSC. Initially powered by a 6.9-liter twin-turbocharged V8 engine, the capacity of the engine was later reduced to 5.9 liters in order to allow the engine to have a higher redline of 8,800 rpm. SSC had stated that the power output would be rated at 1,350 hp (1,007 kW; 1,369 PS) or 1,750 hp (1,305 kW; 1,774 PS) on E85 fuel...
 
@terdon It's hard for me to imagine that speed. At 316 MPH you are moving almost as fast as a 45 caliber bullet travels
That's 1 mile in 10 seconds
 
But by the look of the car, you can only drive it on a race court, it's very low.
 
Yeah people drive supercars on the street though, A lot of them have a front lift option to get you over speed bumps and such but I don't think this one has it
316 MPH is 141 meters per second
You can barely see that far
 
1:15 PM
Yeah. That... doesn't really attract me at all. I'd like to go for a ride once, but I wouldn't want to own one.
 
Oh I definitely would never want to go that fast. The driver that did the run said he almost died and he hopes no customer ever attempts that
At that speed not only can you not hit the brakes, but if you let off the gas too fast you will lose control and die
 
Seems pretty pointless, really.
 
It's impressive engineering though
But yeah top speed records are mostly pointless for any car but it's a common trend for car companies to gain recognition. After koenigsegg broke the record he said he has no interest in trying to do it again because he would rather build a useable car, but it put his companies prestige up there with the other hypercar manufacturers
 
Impressive engineering yes, absolutely.
 
2:12 PM
@Tim Thanks. I've already enabled HTTPS on a development machine, but that post might be useful if I ever have to run a Python app on a production web server using HTTPS.
 
Tim
@BlackPanther Nice to know. How did you manage to enable HTTPS?
 
@Tim I had to create a CA and a certificate, then sign the certificate using that CA, make chrome trust the CA, and finally pass the certificate with the key and password to Kestrel. Something along those lines.
@Tim do you have Fedora?
@Tim I'm asking because I'm trying to downgrade a package but dnf outputs Error:Unable to find a match: dotnet-sdk-3.1.302-1.fc31.
What's the correct syntax to install a specific version of a package using dnf?
 
Tim
@BlackPanther I have only used apt. Does Fedora has a GUI frontend to dnf?
presumably: sudo dnf install <package-name>
@BlackPanther Do you have step by step commands for achieving that?
 
2:31 PM
@Tim Not that I know of, maybe one exists.
But I prefer the command line.
@Tim That installs the latest package.
I have something like this:
$ dnf --showduplicates list dotnet-sdk-3.1
Installed Packages
dotnet-sdk-3.1.x86_64                                                                    3.1.402-1                                                                     @packages-microsoft-com-prod
Available Packages
dotnet-sdk-3.1.x86_64                                                                    3.1.101-1                                                                     packages-microsoft-com-prod
dotnet-sdk-3.1.x86_64                                                                    3.1.102-1                           
Can someone delete the above post, formatting is messed up.
$ dnf --showduplicates list dotnet-sdk-3.1
Installed Packages
dotnet-sdk-3.1.x86_64                  3.1.402-1                    @packages-microsoft-com-prod
Available Packages
dotnet-sdk-3.1.x86_64                  3.1.101-1                    packages-microsoft-com-prod
dotnet-sdk-3.1.x86_64                  3.1.102-1                    packages-microsoft-com-prod
dotnet-sdk-3.1.x86_64                  3.1.103-1                    packages-microsoft-com-prod
dotnet-sdk-3.1.x86_64                  3.1.104-1                    packages-microsoft-com-prod
That's better :).
So, say I have the latest package version 3.1.402-1 installed with dnf, but I want to downgrade to version 3.1.104-1, the commands sudo dnf install dotnet-sdk-3.1.104-1 and sudo dnf install dotnet-sdk-3.1.104-1.fc31 both output:
No match for argument: dotnet-sdk-3.1.104-1
Error: Unable to find a match: dotnet-sdk-3.1.104-1
And
No match for argument: dotnet-sdk-3.1.104-1.fc
Error: Unable to find a match: dotnet-sdk-3.1.104-1.fc respectively
How can I downgrade to a specific package version using dnf?
 
Because that's not the package name, is it?
I mean where did dotnet-sdk-3.1.104-1.fc come from?
dotnet-sdk-3.1.x86_64-3.1.104-1?
 
2:48 PM
@Tim I had to search my memory since I archived the solution process. The details of the way I did it is this stack overflow answer and this following message of mine here in the Unix /dev/chat chatroom:
Oct 27 at 0:55, by Black Panther
Sorry, I was testing the latest attempt, and it now works!!! I ran the command openssl pkcs12 -export -out localhost.pfx -inkey localhost.key -in localhost.crt, and now chrome connects to Kestrel!
@AndrasDeak Thanks. I assume the left column shows the package name but they are all the same?
@AndrasDeak From an example in man dnf:
> dnf install tito-0.5.6-1.fc22
Install the package with a specific version. If the package is already installed it will automatically try to downgrade or upgrade to the specific version.
 
@BlackPanther I would think so but I'm 100% guessing. Just looking at unix.stackexchange.com/questions/266888/…
 
`$ sudo dnf install dotnet-sdk-3.1.x86_64-3.1.104-1`
`No match for argument: dotnet-sdk-3.1.x86_64-3.1.104-1`
`Error: Unable to find a match: dotnet-sdk-3.1.x86_64-3.1.104-1`
 
Thanks, it was worth a shot
Last answer says there's even a dnf downgrade
 
Yeah, twas.
I saw that too. I'm quite familiar with dnf install so know exactly what will happen when I give it a command. dnf downgrade, not so much.
 
Yeah
Perhaps there's a "dry run" switch that tells you what it would do
 
3:03 PM
Just checked. Unfortunately, there's no dry run.
But having read the command help, I see that it downgrades a package.
The only problem is, I don't know how to specify which package to downgrade to.
 
@BlackPanther That's not a shocking outcome
 
@BlackPanther any chance it asks you?
 
@AndrasDeak Having tried. If I could do a dry-run first, that would alleviate any concerns.
I don't really fancy cleaning up any mess caused by running a command I don't fully understand. Call me risk averse, but been there, done that.
 
no, you're doing it right
 
3:28 PM
Hey @AndrasDeak I read the following in man dnf:
Downgrade Command
       dnf [options] downgrade <package-spec>...
              Downgrades the specified packages to the highest installable package of all known lower versions if possible. When version is given and is  lower  than  version  of  installed
              package then it downgrades to target version.
It seemed safe to use as there did not seem to be any gotchas, and after reading this redhat example on how to downgrade httpd I realized that I wasn't using the right format for the package name and version string:
Now running dnf downgrade dotnet-sdk-3.1-3.1.104-1 or dnf install dotnet-sdk-3.1-3.1.104-1 works.
There is no dry-run but at least both commands prompt you to proceed or decline.
Based on the output from both commands before they prompt you for an input, it seems that dnf install dotnet-sdk-3.1-3.1.104-1 just calls dnf downgrade dotnet-sdk-3.1-3.1.104-1.
@AndrasDeak are you on Fedora btw?
I mean it looks like dnf install dotnet-sdk-3.1-3.1.104-1 just calls dnf downgrade on the package with the specified version dotnet-sdk-3.1-3.1.104-1.
 
3:47 PM
@BlackPanther no, otherwise I wouldn't be 100% guessing
 
@AndrasDeak An exorcism might be in order.
 
I see, your guesses were fairly on the mark though.
@AndrasDeak Does cp use time stamps to check if a source file is newer than a destination file?
If so, maybe using something like content defined chunking, which restic implements, would be better?
 
4:34 PM
@Wildcard what about gimp?
 
@BlackPanther no idea
 
@Tim As far as I know, in the written form American/Canadian English differ from UK English in the spelling of a few words. I did not know that Australian and New Zealand English differ in any way to UK English other than when comparing accents in spoken English?
@AndrasDeak Using time stamps is the easy and obvious choice, but is not robust. If you called cp using strace maybe you could debug it.
Stack Exchange chat has a different layout to the Stack Exchange site. Is Stack Exchange chat a different website to Stack Exchange?
 
4:50 PM
Not sure what you mean, but probably. All three chat servers are independent too.
 
I mean this layout chat.stackexchange.com/… is different to this one stackexchange.com
@AndrasDeak Do you think those resources would still be useful for someone who has lower intermediate level knowledge of git?
@AndrasDeak Can you think of any reason for making all three chat servers independent? I assume it will be more costly to run.
 
@BlackPanther depending on what that means, less likely
@BlackPanther some dynamics are server-wide. Like flags and mod privileges
 
@terdon Wow, that's fast. Just discovering this manufacturer and car now. I think that's faster than some helicopters?
@AndrasDeak I see, thanks.
@AndrasDeak right, dynamics?
 
Like flags and mod privileges...
Chat behaviour
 
5:06 PM
I see, thanks.
 
 
3 hours later…
7:42 PM
@BlackPanther no idea.
 
@MichaelHomer No, I don't know you. Yet I dare violate your (imaginary) personal space, and address you directly over your personal remarks about me. You better ban or ignore me, since I'll keep doing it whenever I feel like.
So, what does that "invective" says about me?
 
@user431397 spoken like a true troll
 
@AndrasDeak that's fine. better be a true than a fake troll
 
Uh, okay?
 
Right, enough of this. @user431397 if you can't engage constructively and politely here, please stay away.
 
7:54 PM
@user431397 Sorry, I wasn't certain that was who you were.
 
 
2 hours later…
10:09 PM
@BlackPanther The GIMP isn't exactly designed as a drawing program. It depends what you are trying to do, though.
 
10:32 PM
Really more of a general image manipulation program.
 
that kind of comes with the name ;)
 
Oh, yes. So it does.
@BlackPanther A common use is to clean up or otherwise change images to be more to your liking/requirement.
I think it's less used to create images from scratch, though one could certainly use it for that.
And it doesn't do vector graphics, which is kind of a big minus.
 
11:39 PM
I was wondering whether to use PCRE with Python 3 or stick with the old re module. But Debian, oddly, doesn't seem to have a Python PCRE library, unlikely though that seems.
Comments? And would it be reasonable to ask about this in the Python room?
According to Tom Christiansen (who should know), Python doesn't have a PCRE library. Which perhaps isn't that surprising, but is unfortunate.
 
11:59 PM
@FaheemMitha Do you have a good reason not to stick with stdlib re?
 

« first day (3675 days earlier)      last day (29 days later) »