-
Notifications
You must be signed in to change notification settings - Fork 5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
WTF This piece of cmake code is not ported to Qt 5 #24385
Comments
Qt 5 is no more supported on Linux. If you need that, you can port that code on your own. This is not the only thing you would need to port to Qt 5 in 3.7, though. |
As for #24014, you need to either build with gcc or wait until clang would add the required C++20 feature |
@ilya-fedin you must understand that this is a regression and this is not correct!
Also, if clang does not include some functionality of the C ++ 20 standard, then there is an important reason for this, at least speed degradation. Also some C/C++ features included in GCC are against other standards like IEEE etc and some functions are marked experimental. |
This is not an experimental feature and C++20 is specified as the minimum required standard |
If clang doesn't introduce something, it's clang's problem, all officially supported configurations are building fine, everything else could be fixed by PRs from people who is interested in such configurations. |
cmake not contains property MAXIMUM_CXX_STANDARD https://cmake.org/cmake/help/latest/search.html?q=MAXIMUM_CXX_STANDARD
MAXIMUM_CXX_STANDARD only in https://github.com/desktop-app/cmake_helpers/blob/master/init_target.cmake#L7 It's your problem as a developer, you shouldn't do regressions. This is a regression. |
grep the file... |
You couldn't really require something on github. Either you PR or shut up and wait until someone else would take time and interest and solve your problem. |
|
Grep the file itself lol |
@ilya-fedin So, good error dolution!
So, this is good error solution! |
Also, I don't really agree with that matter. If something is not (first-class) supported, it's clear it will regress. And it's the job of the one who is interested in supporting this to fix this. |
You didn't just break the code in one place, you broke the code in two places. |
And? |
None of the officially supported configurations regressed. Everything else is community-supported and it's community's job to fix these configurations. |
@ilya-fedin this is a complete regression.You have broken the build capability. |
If you don't really support something, you don't regress on that (as it's not supported in the first place). |
@ilya-fedin I commit to a lot of projects and the changes just won't be accepted into any major projects if there is a regression. |
If it's a regression to something supported then yeah. If it regresses in something that is not really supported and no one checks such configurations then no one would care about that. |
You just surprised that clang to tdesktop as niche platforms to some projects. Now you know. And I hope you will stop require anyone do anything on github, no one works for you here and no one will do anything just because you think that's how it should be. |
I do not demand anything, just normal developers usually try to solve such problems, and not shift them to the community. I personally have no time to understand the tdesktop code, as I am busy with other projects, and studying the code takes time. And as far as possible, I send patches if some package stops being built by clang. If you do not have enough hands, which is in many open source projects, then you need to look for students and engage in mentoring, for example, through gsoc. For any normal project, it's considered good practice to pay attention to bug fixes. |
Well, it sounded like that.
I guess you usually talk with people from community who is interested in supporting clang. I'm from community, I'm not interested in that though (right now at least). |
Can I ask, how you build tdesktop for darwin/macos? It is official compiler is clang. |
Можно я на русском? Вроде, все свои. Я тоже заинтересован в поддержке сборки через clang. Если я это починю, расставив по месту вызов |
AppleClang uses Qt6. |
strange, I have same error(in another file) with qt6/linux |
AppleClang seem to handle such code just fine |
Sounds really strange, cause apple clang have same frontend that of stock clang. |
Apparently it's not really the same |
@pg83 most likely to be accepted. But honestly, I'm already tired that after each release, the compilation breaks down and I have to spend my time fixing it. Even chromium doesn't take me that long. I don't know why they are prevented from using the latest version of clang and clang-tidy static analyzer. Could identify some of the potential errors at the development stage. There are plenty of errors in telegram, and gcc won't always point them out. At least if clang can compile, then other compilers usually can too. But this does not work for other compilers. I want to note the problem occurs only with qt 6, with qt5 it compiles successfully. This is hardly a compiler problem, in my opinion, gcc just compiles code where clang thinks that we can get unpredictable behavior or there are ambiguous interpretations of the C++ standard. |
@h0tc0d3 AppleClang is based not on the latest clang, I think. There is current development flow, where MSVC is used on Windows, AppleClang on macOS and GCC on Linux. All of them have their problems and that way you get the most portable code. This is not an error in Telegram, this is an error in clang, as pointed out by range-v3 author. It can be worked around somehow, I guess, so PRs are welcome. |
Please fix error and replace range v3 code with stdlib #24014
The text was updated successfully, but these errors were encountered: