Wrong number of snap packages reported

[JacksonMcKay]

Description

Neofetch claims my system has 9 snap packages instead of 8. I suspect all package managers with extraneous output (headers, etc.) are affected. I had a look at the source, notably: https://github.com/dylanaraps/neofetch/blob/175915d0a6a8183cf00030b6e88412cb13284289/neofetch#L1392 where the packages are counted by tot and then ((packages-=1)) is done to eliminate the header line. I believe this only decrements the total package count, not the package count for individual package managers which seems to be set in pac() https://github.com/dylanaraps/neofetch/blob/175915d0a6a8183cf00030b6e88412cb13284289/neofetch#L1311-L1312 with the fragment managers+=("$1 (${manager})"). It's only after all this that ((packages-=1)) comes into play.

Apologies, I don't think my bash-fu is up to scratch to fix it myself 😅

Package managers affected: tazpkg, snap, port, scoop, versions

Neofetch version

Neofetch 6.0.0

Screenshot

Notice how --package_managers off seems to list one less package than the default option

Config file

N/A

Verbose log

https://pastebin.com/w5n5JLJg

[dylanaraps]

Great catch!

This will be interesting to solve.

[A-725-K]

Good afternoon ! I think I could have a solution for this bug, but since this is my first contribution to an open source project I would appreciate a little help to test some of the package managers that I don't have to be sure that my fix work before making the pull-request :smile:

As you can see from the photo, now my number of "snap" and "dpkg" packages is correct. For now the fix is available only on snap but I think my solution could work also with the other package managers as well with few changes.

[JacksonMcKay]

Hi! I'd be happy to test your solution with snap and scoop - could you make a pull request or push your code somewhere where I can get my hands on it?

[A-725-K]

Hi! I'd be happy to test your solution with snap and scoop - could you make a pull request or push your code somewhere where I can get my hands on it?

First of all thank you for your kind help ! If you want to get the code you can clone my fork and switch the the branch called packagenum. I will really appreciate your feedback, especially on scoop since I haven't it on my system and it is the most doubtful change I've made to the code :sweat_smile:

[A-725-K]

I was able to test my fix on both GNU/Linux (Parrot OS) and Windows 10 [scoop and snap] and it actually worked! I'll submit soon the pull request. If anybody is able to test the script with other package managers that were affected by the bug I would appreciate much. Thank you all ! 😃

[JacksonMcKay]

Hi, sorry for the delay! I was able to test your fix with snap on Manjaro Linux, scoop on Windows and port on a Mac and all returned the correct number of packages both for individual package managers and with the --package_managers off option (meaning there was no regression there). Looking at your changes it looks like you've covered all bases and I don't see any reason why there would be issues with the other two package managers mentioned - Nice work!

[A-725-K]

Thank you so much @JacksonMcKay for your help and feedback, it was fundamental and you were very kind ! I hope this will be merged soon 😄

[loannaflip]

similar thing also happens on Bedrock. Neofetch prints the wrong number of portage installed packages (gentoo stratum).

[A-725-K]

similar thing also happens on Bedrock. Neofetch prints the wrong number of portage installed packages (gentoo stratum).

In theory my fix should address also Bedrock. May I ask you to test my solution on gentoo since I'm not able to do it, please? It would be great. Thank you for your cooperation !