-
Notifications
You must be signed in to change notification settings - Fork 489
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
grep 3.6 update is a downgrade to grep 3.0 #2678
Comments
grep > 3.0 breaks a lot of existing things such as compilers and libraries |
Original issue is quickly closed imo, downgrading to grep 3.0 should be considered a temporary solution grep is my most used tool whether I'm on Windows or Linux. I'm not affected by the issue and I find this is a pity to downgrade grep just because a part of users can't get things working as before. It just lacks a patch and issue should be reopened |
@class101 it's not that some users have something broken. Grep > 3.0 broke compiler totally. |
Add an alternative instead of grep use ugrep in PKGBUILD MSYS-packages. When you do this, use utf by default in python. All packages from msys pakages will need to be rebuilt. In mintty you have everything in utf but the default output for python is ansii. And python does not distinguish between upper and lower case letters like in linux, but only does everything like in windows. Add the missing patch files from cygwin to python. |
Rebuild gcc-10.2.0 from msys. With such check results. 😎 |
@piotrvv what you want to say with this? Grep broke MINGW compilers, not MSYS. You will be able to build compiler but it broken. Read carefully issues that we point |
I guess we are dealing with a problem in msys grep. At the end of the line. Which causes problems when building in mingw. I don't think this is just a problem of the mingw themselves. The suggestion is to compare the outputs of grep and ugrep. |
@piotrvv during build MINGW compiler we always use MSYS tools |
I know you use them. But I think a lot of collaboration problems are posed by the grep itself in the msys packages. When you use grep and ugrep you will see for yourself how many problems there are with the msys2. Utf and ascii are mixed up in many places and the end of the windows and linux lines are a big contributor to the problems of grep itself. Msys packages has python in ascii mode by default. |
ugrep keeps telling me i have an asci line, but the mintty terminal still turns utf. When you build the package in msys tool mintty it sees utf but grep goes silly. For the msys packages, it's best to move everything to utf in mintty. |
Sync lines to utf in python perl ruby on the output in the mintty terminal |
You build packages in msys tools but will forget about the default package options in msys packages. |
By default, msys packages uses 3 different default languages. |
@piotrvv I don't understand what you mean. |
Perl has an output in a utf terminal by default. Python defaults to ascii. Ruby has a different default language by default. Try to check something with grep. |
You need to synchronize the default languages in ruby and python perl. Because minty will display them but grep will have problems with mixing different default languages on output. |
Due to the lack of synchronization of these languages, you have problems when autootols does something. Meson and pacman dont understand. |
Due to the default different languages in mintty we have a compilation problem. |
@piotrvv i don't mind about this tools. Concrete mingw GCC commands output parsing wrong. Please read all bug reports about this issue before next discuss. |
I have your msys-tools on my server and they are following the changes. I have a terminal in many places I see problems with proper encoding in the terminal line. You forget about the correct language settings in many packages. |
@piotrvv original issue is not about encoding, but encoding problem maybe as separate issue |
Alexpux wrote:
Except that grep works fine on e. g. my Linux box where I compile most things from source as-is. This is more a problem of the msys toolset that should be fixed rather than insinuate that "grep |
@rubyFeedback if you want new version of grep, you can create your own package, for example grep-new, that replace grep. It will be not overwrites by pacman. |
If you can fix it either within grep or autotools (this is where this bug manifests) we would really appericate. We don't have knowledge nor time to take care of it.
Except Linux uses sane defaults for the text files, this is not the case on Windows hoverver.
So what were we supposed to do, keep latest grep version and stop building updates for most of the packages because the compiler is broken?
It never did. To downgrade package you have to use Aside from that anyone can use native grep from mingw/ucrt/clang repositories which is much newer. |
I don't think this reason is 100% valid. For example, we recently upgraded a project from jdk8 to jdk11. It worked fine under 8, but some features broke in 11, this does not mean 11 is bugged. |
@class101 You and others don't hear what we tell. Without properly working GCC we cannot build other packages. So this is critical issue with grep and we will not update grep while there are no solution |
But, if you had some dependency that didn't work under 11 but did under 8, you might reasonably hold off on updating to 11 until the dependency is updated to support it. This is the situation with autotools working under grep 3.0 but not 3.6. |
grep 3.6 update is a downgrade to grep 3.0
Describe the issue
When I update MSYS2, I'm presented an update that is in reality a downgrade of
grep to version 3.0
Steps to Reproduce the Problem
pacman -U grep-3.6-1-x86_64.pkg.tar.zst
pacman -Su
1~3.0-3
is available3.6-1
Additional Context: Operating System, Screenshots
Output of
grep --version
withgrep-1~3.0-3
Output of
grep --version
withgrep-3.6-1
The text was updated successfully, but these errors were encountered: