Re: About the intentionally duplicated package wechat-beta-bwrap and wechat-universal and unmaintained orginal package wechat-uos

[Kimiblock Moe]

Hi Aᴀʀᴏɴ,

If you test the old wechat-uos package in a sandbox (assuming you have 
a WeChat account), you'll realize that the old version is lacking a 
good amount of functions such as chat history saving and VoIP. UOS 
called the new "Universal" WeChat an "Reconstruct and Upgrade" 
("升级" is upgrade in Chinese, while "重构" is reconstruct) in 
their blog post provided by 7Ji[a] and their AppStore[b][c] which 
provided a feature list which the old version lacks. This in my point 
of view proves that the new Qt version is an upgrade to the old one, 
and the old version is just keeping there so that users who don't want 
to migrate staying on the old version just fine. (UOS is advertising 
itself "stable", after all)


I personally agree the fact that this is a small breaking update, but 
that's just users' downloaded files stored in a different place. If 
anyone wants the old version, they should package an old 
wechat-uos-legacy-bin instead of expecting wechat-uos staying on the 
old version indefinitely. It'll be more sensible if wechat-uos's 
maintainer(s) introduce a detection logic which alerts old users where 
the old data is located at.


--
Sincerely,
Kimiblock


[a]: https://bbs.chinauos.com/zh/post/17786
[b]: https://postimg.cc/bDt46brm
[c]: https://postimg.cc/1nQbjXFP

Re: About the intentionally duplicated package wechat-beta-bwrap and wechat-universal and unmaintained orginal package wechat-uos

[Aaron Liu]

Ey,

Welcome to the mailing lists! Please read and follow 
our [guidelines][1]. Specifically, try to use plain text emails.


Also, I've noticed that your PKGBUILD includes a lot of packages 
included in "base" as dependencies. Please don't do that; packages 
installed as part of "base" are part of every Arch distro. Additionally, 
packages that don't compile anything and install a precompiled binary 
instead (such as yours) should have a -bin suffix.


Firstly, may I ask what's the difference between the universal version 
`wechat` and `weixin`? If its just a newer version of WeChat, then it 
should remain the wechat-uos package name or prefix. 


As 7Ji has explained in their rather long email, "weixin" uses Electron, 
which is non-native, while "wechat" uses Qt, which is native. The latter 
is a rewrite of the former, and both are maintained.


To reduce this confusion, maybe the latter version's package should be 
named like "wechat-uos-qt" instead.


I also believe that they are keeping the old "weixin" (probably 
unmaintained) because the fact that the so called "Universal" Qt 
version doesn't port over any data from legacy Electron version
Regardless of speculation on the state of the Electron version's 
development, it is still a separate version, and people want both 
versions of the package. Therefore, I agree with 7Ji that the Electron 
and Qt versions should remain separate.


[1]:https://wiki.archlinux.org/title/General_guidelines#Mailing_lists  

--
Cheers,
Aᴀʀᴏɴ



OpenPGP_0x4E85967FC7C436BE.asc
Description: OpenPGP public key


OpenPGP_signature.asc
Description: OpenPGP digital signature

Re: About the intentionally duplicated package wechat-beta-bwrap and wechat-universal and unmaintained orginal package wechat-uos

[Aaron Liu]

Ey,

While I agree with most of your points...

Additionally, the maintainer is ... being unfriendly to both tencent 
and WeChat, phrasing them as "trash" in the startup script[i], ... 
clearly a violation to the Arch Linux Code of Conduct[l] 
How does being unfriendly to a proprietary corporation violate the code 
of conduct? It's not really a political topic, Tencent and WeChat aren't 
"someone" so it does not count as a personal insult, and "respect other 
... projects" only applies to FOSS stuff.


I also see no reason that non-arch distro detection should be removed. 
It doesn't break any submission (or conduct) guidelines.


As a side note, packages that don't build anything and instead install a 
precompiled version should have a -bin suffix.


--
Cheers,
Aᴀʀᴏɴ



OpenPGP_0x4E85967FC7C436BE.asc
Description: OpenPGP public key


OpenPGP_signature.asc
Description: OpenPGP digital signature

Re: Package Maintainer Application - Carl Smedstad

[Torsten Keßler]

Am 03.03.24 um 16:27 schrieb Carl Smedstad:

Hi everyone,

Hi Carl!


My name is Carl, or carsme on AUR/IRC, and this is my application to
become a Package Maintainer.
Thank you very much for your application! I already came across your 
work on the AUR a couple of times and had the pleasure to adopt some of 
your packages and add them to [extra].




For reference, PKGBUILDs for all packages I maintain in the AUR are
available here: https://github.com/carlsmedstad/aurpkgs
As mentioned earlier, they are of high quality and cover a broad range 
of build systems and applications. One minor thing I spotted while going 
through some of PKGBUILDs is the use of CMAKE_CXX_FLAGS, for instance in


https://github.com/carlsmedstad/aurpkgs/blob/d071c9edc238af3819a5fddb6131195ef5df5761/luau/PKGBUILD#L35

If you set CMAKE_CXX_FLAGS then it overrides Arch's defaults. You have 
to explicitly add $CXXFLAGS to the cmake option to append options.



Hopefully that gives a good picture of who I am and where I'd like to
start my journey as a package maintainer, if accepted.
Overall a very strong application! I think you would be a great addition 
to the team.


Cheers,
Torsten


OpenPGP_0x07D06351CA5B31BE.asc
Description: OpenPGP public key


OpenPGP_signature.asc
Description: OpenPGP digital signature

Re: Question about aur package existing in git, but not visible in aur

[Julien Virey]

Hey Aaron,

Thanks for the prompt reply.

I did find the the deletion request:
https://lists.archlinux.org/hyperkitty/list/aur-reque...@lists.archlinux.org/message/YJMFPPNPLLZPP5GGDDBHKXMIVCWZPUEA/

I was also able to push to ssh://aur.archlinux.org/toggl-cli.git without
issue, so thanks for your explanation.

https://aur.archlinux.org/packages/toggl-cli is now available.

Salutations,
Julien


On Sat, Mar 16, 2024 at 3:20 PM Lime In a Jacket (Aaron Liu) <
aaronliu0...@gmail.com> wrote:

> Ey,
>
> If the package doesn't exist but its repository does, it means the package
> existed but was deleted. Anyone can now push to that repository to list the
> package again, but I'd advise you to search for the deletion request first.
>
> Cheers,
> Aᴀʀᴏɴ
>

Re: Question about aur package existing in git, but not visible in aur

[Lime In a Jacket (Aaron Liu)]

Ey,

If the package doesn't exist but its repository does, it means the package 
existed but was deleted. Anyone can now push to that repository to list the 
package again, but I'd advise you to search for the deletion request first.

Cheers,
Aᴀʀᴏɴ

Question about aur package existing in git, but not visible in aur

[Julien Virey]

Hi,

The package toggl-cli is not visible in
https://aur.archlinux.org/packages?O=0=toggl-cli.
But it still exists in ssh://aur.archlinux.org/toggl-cli.git

The package has not been updated since 2016.
commit dd7340e8804776df85fe1eb82a79384e7f1349fb (HEAD -> master,
origin/master, origin/HEAD)
Date:   Mon Feb 22 06:00:17 2016 -0300

I was about to release a new package for this pkgbase (
https://github.com/watercooler-labs/toggl-cli).

Can I request a deletion / reset of ssh://aur.archlinux.org/toggl-cli.git
to use this pkgbase ?

Salutations,
Julien

Re: About the intentionally duplicated package wechat-beta-bwrap and wechat-universal and unmaintained orginal package wechat-uos

[Kimiblock Moe]

Hi 7Ji,

Firstly, may I ask what's the difference between the universal version 
`wechat` and `weixin`? If its just a newer version of WeChat, then it 
should remain the wechat-uos package name or prefix. People from the 
#archlinux-cn group and I tested and realized that WeChat cannot run or 
login without some system files from the UOS system. Therefore it's 
still exclusive to China's UOS. Naming it "-universal" might be a 
little misleading in my opinion. I also believe that they are keeping 
the old "weixin" (probably unmaintained) because the fact that the so 
called "Universal" Qt version doesn't port over any data from legacy 
Electron version "weixin" (Chat history, Pictures, Videos etc.) Causing 
a breaking upgrade which is not ideal for a non-rolling distribution. 
It is advisable to update the original "wechat-uos" and put in a 
.install file to remind users about this breaking change using 
`post_upgrade`. Arch is always a rolling "always up-to-date" system, 
therefore it is suggested that a new package called wechat-uos-legacy 
submitted to the AUR for those who doesn't want to upgrade and 
introduce a detection logic in wechat-uos which reminds old users to 
migrate their data or install a legacy version.


I apologize about the script calling the Tencent WeChat a "trashApp". 
But it's an internal variable for debugging bubblewrap sandbox which 
was never meant to be used or seen by users. Also regarding the 
non-Arch detection logic, I don't see the any reason why not including 
it. This detection prevents the issue which appears on systems with a 
modification package "moeOS" installed, and from my knowledge it never 
broke any Arch user's installation. According to you mail reply, 
unfortunately the wechat-uos-bwrap is broken on your system. But I 
didn't see any reports from you on the AUR comments area (or maybe it's 
missed). If there's an issue with the package, please report so that I 
can improve the quality of this script.


Nevertheless, I still suggests that updating `wechat-uos` and merge all 
other "Native WeChat" packages into it, including wechat-uos-bwrap, 
wechat-beta-bwrap and wechat-universal.


--
Sincerely,
Kimiblock

Re: About the intentionally duplicated package wechat-beta-bwrap and wechat-universal and unmaintained orginal package wechat-uos

[Guoxin Pu]

Hello, I'm 7Ji, maintainer of wechat-universal-bwrap[a] (wrongly typed 
as wechat-universal in Kimiblock's original mail), and co-maintainer of 
wechat-beta-bwrap[b]. leaeasy[c], the maintainer of -beta, is my friend. 
I think it's necessary for me to clear up some misunderstanding and 
introduce the current situation of wechat builds from Tencent that's 
released on the Uniontech OS (an officially backed variant of Deepin 
Linux) appstore front that all these `wechat` packages are trying to 
"borrow" from.



For a long time, WeChat never had a native Linux build. Linux users in 
China that had to use WeChat for different reasons had to use Wine, even 
on the officially backed UOS.



On ~ late 2020, a "native" client that's built using Electron was 
released onto the UOS appstore, and into their Deb library. This package 
was named `com.tencent.weixin` in the deb library, and named `WeiXin` on 
their appstore. This is the binray release which wechat-uos[e] re-packs 
from.



On March 5th, 2024, another native client that's still in Beta testing 
was leaked to public, it was built using native Qt tech stack. This 
package was named `com.tencent.wechat` in the deb library for UOS, and 
named `Weixin (Universal)` on their appstore after it was officially 
relased on March 14th and announced on the UOS tech blog[f]



Please note that these two builds, while both released onto the UOS 
appstore and their deb library, have different deb names 
(`com.tencent.weixin` vs `com.tencent.wechat`) and different store front 
names (`Weixin` vs `Weixin (Universal)`), they're treated as different 
packages both in the deb library and in the appstore front. They're 
therefore two distinct packages, that shall be re-distributed as two 
distinct packages.



On to `wechat-universal-bwrap`,  it bears the name `wechat-universal` 
because it re-packs `com.tencent.wechat` / `Weixin (Universal)`, and 
that's the right name for the package. It has the whole commit history 
of `-beta` and was properly renamed from `wechat-beta-bwrap` in a 
commit[g]. In fact, the whole package was `-beta` renamed, and leaeasy 
agreed to me on this privately. This was re-submitted as a new package 
due to the AUR limitation that packages can't be simply renamed. But 
there's a merge request PRQ#57762[d] filed by myself immediately after 
it was submitted, and no new commits was pushed to `-beta` after the 
rename.



On to `wechat-uos`, it re-packs `com.tencent.weixin` / `Weixin` from day 
1. And as `com.tencent.weixin` / `Weixin` is still maintained in the UOS 
deb library and appstore front and would be maintained as a distince 
package from `Universal` for a long time, it should still re-pack that. 
Switching the source to `com.tencent.wechat` is irresponsible and a bad 
idea for existing users as they use completely different tech stack and 
the data is not possible to switch around losslessly between the two builds.



On to `wechat-uos-bwrap`, it should do whatever `wechat-uos` does and 
provide additional bubblewrap sandboxing, judging from the name. 
However, in reality it diverged from `wechat-uos` and switched to the 
`.wechat`/ `Universal` source, essentially being a completely different 
package not only with sandboxing differences, breaking the assumption 
that is was a tweaked `wechat-uos`. Additionally, the maintainer is both 
packing non-Arch distro detection logic[h] stuff for their moeOS, and 
being unfriendly to both tencent and WeChat, phrasing them as "trash" in 
the startup script[i], and packs broken URLs currently. This PKGBUILD 
never worked on my KDE plasama + Wayland setup, and the unfriendliness 
is clearly a violation to the Arch Linux Code of Conduct[l]



On to `wechat-beta-bwrap`, it borrowed from `-uos-bwrap` heavily for the 
first commit as leaeasy was a new beginner to write PKGBUILDs. But it 
worked from day 1 and never packed non-Arch logics, and huge rewrites 
had landed in the PKGBUILD after the initial commit. It also had a right 
name at the time it was submitted: `-beta` as it re-packs a leaked Beta 
build different from `uos`. leaeasy, the maintainer of this package, 
invited me to become the co-maintainer after I requested to add aarch64 
support, which eventually landed as commit 99160c, and I have since 
actively maintain the package. I knew there was an ongoing merge request 
from it into `-uos-bwrap` and I already replied to PRQ#57379[k] to 
explain why `-beta` should be its own package, not overtaking the 
`-uos-bwrap` name. Those reasons still apply, to why `-universal` should 
be its own package, not overtaking the `-uos` name, and `-uos` should 
not use `com.tencent.wechat` source.



So if this mess really needs to be solved, these packages should be 
handled as follows:


 - `wechat-beta-bwrap` shall be merged into `wechat-universal-bwrap`, 
completing PRQ#57762[d], effectively renaming the package. I'm the 
co-maintainer of `-beta-bwrap`, and submitter & maintainer of 
`-universal-bwrap`. The