Packages being flagged out of date
I have noticed several packages that have been flagged out of date, but they are not. The stated reason for the flag will be a problem the user has with the PKGBUILD that is unrelated to the program's version or if the package can be built. I did not see anything in the wiki directly addressing this, so, being relatively new to contributing to the AUR I was curious if this is considered appropriate. At least one package also has an orphan request referencing the flag, which seems misleading to me if the flag should only be used for alerting the maintainer and users to a new version upstream. I have purposely not included package names to avoid calling out anyone publicly.
Sep 17, 2023 15:12:00 Jeremy Gust
I have noticed several packages that have been flagged out of date, but they are not. The stated reason for the flag will be a problem the user has with the PKGBUILD that is unrelated to the program's version or if the package can be built. I did not see anything in the wiki directly addressing this, so, being relatively new to contributing to the AUR I was curious if this is considered appropriate. At least one package also has an orphan request referencing the flag, which seems misleading to me if the flag should only be used for alerting the maintainer and users to a new version upstream.
I have purposely not included package names to avoid calling out anyone publicly. The page to flag packages out of date specifically says it's for new updates only not for bug reports and such.
-- Kusoneko GPG: https://kusoneko.moe/gpg.txt https://kusoneko.moe
On 9/17/23 13:15, Kusoneko wrote:
The page to flag packages out of date specifically says it's for new updates only not for bug reports and such.
Yes, I had forgotten that. "Please do not use this form to report bugs. Use the package comments instead. Enter details on why the package is out-of-date below, preferably including links to the release announcement or the new release tarball." There it is, in very plain language. One would think that a person that spends a lot on time helping to maintain the AUR would have seen that. Unfortunately, it appears that some have not, or ignored it.
Sep 17, 2023 16:55:19 Jeremy Gust
On 9/17/23 13:15, Kusoneko wrote:
The page to flag packages out of date specifically says it's for new updates only not for bug reports and such.
Yes, I had forgotten that.
"Please do *not* use this form to report bugs. Use the package comments instead. Enter details on why the package is out-of-date below, preferably including links to the release announcement or the new relea se tarball."
There it is, in very plain language. One would think that a person that spends a lot on time helping to maintain the AUR would have seen that. Unfortunately, it appears that some have not, or ignored it.
Did you need to PGP encrypt a message to the list? -- Kusoneko GPG: https://kusoneko.moe/gpg.txt https://kusoneko.moe
On 9/17/23 14:58, Kusoneko wrote:
Did you need to PGP encrypt a message to the list?
-- Kusoneko GPG: https://kusoneko.moe/gpg.txt https://kusoneko.moe
I only meant to reply to the list. Did something go through incorrectly? I'm a novice with PGP, and I'm not aware of configuring Thunderbird in any particular way regarding it.
Sep 17, 2023 17:06:16 Jeremy Gust
On 9/17/23 14:58, Kusoneko wrote:
Did you need to PGP encrypt a message to the list?
-- Kusoneko GPG: https://kusoneko.moe/gpg.txt https://kusoneko.moe I only meant to reply to the list. Did something go through incorrectly? I'm a novice with PGP, and I'm not aware of configuring Thunderbird in any particular way regarding it.
It was sent to me, PGP encrypted, with the list in CC. -- Kusoneko GPG: https://kusoneko.moe/gpg.txt https://kusoneko.moe
Hello, That is where Jeremy has replied all, which has sent one copy to you (which thunderbird automatically PGP encrypted), and one to the mailing list (which can not be PGP encrypted, so is sign-only). And no PGP signing is not mandatory unless it is an application, or some official matter (if you are arch staff), but apart from this I doubt anyone cares. Hope this has cleared up the confusion. Take care, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@icebound.dev
On 9/17/23 15:23, Polarian wrote:
Hello,
That is where Jeremy has replied all, which has sent one copy to you (which thunderbird automatically PGP encrypted), and one to the mailing list (which can not be PGP encrypted, so is sign-only).
And no PGP signing is not mandatory unless it is an application, or some official matter (if you are arch staff), but apart from this I doubt anyone cares.
Hope this has cleared up the confusion.
Take care, Huh, weird. In Thunderbird Polarian's messages say "Reply List" and Kusoneko's say "Reply All" on the same button.
Hello, It is because my emails go directly to the mailing list, kusoneko sends one directly to you and then CC's the mailing list, which means you will get two copies, one will say "reply all" which will CC the mailing list and email them directly, and then the other email (which is sent back to you from the mailing list), will have "reply list". As I only email the mailing list, and do not send a direct email, you will not have "reply all" as there is not multiple recipients. Hope this makes more sense now :) Take care, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@icebound.dev
You're correct, my client default to reply to sender and cc the list for some reason, I have to manually move the CC to To and remove the extra To in order to reply to list. It's a bit of a pain in the ass to do for every single message, hence why I usually don't bother. Most lists check the headers and if you're already in them won't send you a second copy, so it generally doesn't matter much, I was just confused for a second cause I thought he encrypted the message to the list and that would have been bad. -- Kusoneko GPG: https://kusoneko.moe/gpg.txt https://kusoneko.moe
Hello, What client are you using? I did check for it in your email headers, but it appears you disabled this for privacy (which is a pretty good idea). Take care, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@icebound.dev
I've been using FairEmail as my mobile client. -- Kusoneko GPG: https://kusoneko.moe/gpg.txt https://kusoneko.moe
Ah, I do not think there is any way to disable that for that specific email client. I do recommend K9 mail, however considering it is being mozilla'd, I might not recommend this client for much longer once Mozilla makes a mess of it. I can't really help as I like to type all my emails on my Laptop, I hate typing emails on phone, and I only use it for reading emails, scanning barcodes which are emailed (for example for transit in the UK you get emailed your tickets), or if I MUST, I will reply to short emails. This has got quite off topic, best to try to return to the original topic :) Hope this helps. Take care, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@icebound.dev
Hello, You only flag out of date when there is a new version (https://wiki.archlinux.org/title/Arch_User_Repository#Flagging_packages_out-...). If the flag is invalid, and an invalid orphan request is filed, it is completely fine because the package maintainers (trusted users) will *hopefully* realise and will decline the request. If they are your packages, and you are sure its an invalid flag, you can simply click "unflag" and it will be "back up to date" without any commit. If you got bugs, they should be posted in comments, not flag out of date. I have realised Kusoneko beat me to it, but hey double the support doesn't hurt anyone :P Take care, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@icebound.dev
On 9/17/23 14:08, Polarian wrote:
You only flag out of date when there is a new version ( https://wiki.archlinux.org/title/Arch_User_Repository#Flagging_packages_out-... ). If the flag is invalid, and an invalid orphan request is filed, it is completely fine because the package maintainers (trusted users) will *hopefully* realise and will decline the request.
Yes, hopefully. Although that does seem like a waste of their time.
If they are your packages, and you are sure its an invalid flag, you can simply click "unflag" and it will be "back up to date" without any commit.
Luckily not my packages, people have been polite and used comments.
If you got bugs, they should be posted in comments, not flag out of date.
Agreed, I hope persons that are improperly using out-of-date flags see this thread and are reminded.
I have realised Kusoneko beat me to it, but hey double the support doesn't hurt anyone :P
Appreciated!
On 9/17/23 14:11, Jeremy Gust wrote:
or if the package can be built
Side note - not directed at your circumstance directly (because I have no idea of the specifics), but... I haven't flagged any packages as out of date, but If you have packages that cannot be built any longer -- that is a bit of a problem. I've run across several AUR packages that depend on ancient versions of nodejs or composer, etc.. and the packages are not flagged. They can no longer be built because a dependency no longer exists, but they are not flagged. In these cases, the bottom line is that a package that can no longer be built -- isn't doing anyone any good. So long as your packages build, and you are not behind the upstream release (doesn't apply if you package is specifically for an older version), then generally you shouldn't be flagged. -- David C. Rankin, J.D.,P.E.
On 18/09/2023 04:09, David C. Rankin wrote:
On 9/17/23 14:11, Jeremy Gust wrote:
or if the package can be built
Side note - not directed at your circumstance directly (because I have no idea of the specifics), but...
I haven't flagged any packages as out of date, but If you have packages that cannot be built any longer -- that is a bit of a problem. I've run across several AUR packages that depend on ancient versions of nodejs or composer, etc.. and the packages are not flagged. They can no longer be built because a dependency no longer exists, but they are not flagged. In these cases, the bottom line is that a package that can no longer be built -- isn't doing anyone any good.
This is when you write a comment about your findings, you don't *need* to flag it.
participants (5)
-
David C. Rankin
-
Jelle van der Waa
-
Jeremy Gust
-
Kusoneko
-
Polarian