RE: fish shell versioning

[Marcin Gondek <drixter%e-utp.net@localhost>]

HI All,

Did it compile for you?

==cut==
=> Extracting local cargo crates
===> Patching for fish-4.2.1
=> Applying pkgsrc patches for fish-4.2.1
=> Verifying /srv/NetBSD/pkgsrc/shells/fish/patches/patch-Cargo.toml
=> Applying pkgsrc patch /srv/NetBSD/pkgsrc/shells/fish/patches/patch-Cargo.toml
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|$NetBSD: patch-Cargo.toml,v 1.1 2025/12/13 17:47:21 vins Exp $
|
|Do not attempt to fetch sources at build time.
|
|--- Cargo.toml.orig    2025-10-07 20:56:06.000000000 +0000
|+++ Cargo.toml
--------------------------
Patching file Cargo.toml using Plan A...
Hunk #1 succeeded at 32 (offset 1 line).
done
=> Verifying /srv/NetBSD/pkgsrc/shells/fish/patches/patch-build.rs
=> Applying pkgsrc patch /srv/NetBSD/pkgsrc/shells/fish/patches/patch-build.rs
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|$NetBSD: patch-build.rs,v 1.1 2025/12/13 17:47:21 vins Exp $
|
|Introduce illumos support.
|
|--- build.rs.orig      2025-11-13 12:09:28.000000000 +0000
|+++ build.rs
--------------------------
Patching file build.rs using Plan A...
Hunk #1 succeeded at 84.
done
=> Verifying /srv/NetBSD/pkgsrc/shells/fish/patches/patch-cmake_Install.cmake
=> Applying pkgsrc patch /srv/NetBSD/pkgsrc/shells/fish/patches/patch-cmake_Install.cmake
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|$NetBSD: patch-cmake_Install.cmake,v 1.3 2025/12/13 17:47:21 vins Exp $
|
|Use pkgsrc defaults.
|
|--- cmake/Install.cmake.orig   2025-11-01 23:44:37.984452608 +0000
|+++ cmake/Install.cmake
--------------------------
Patching file cmake/Install.cmake using Plan A...
Hunk #1 succeeded at 6.
Hunk #2 succeeded at 85.
Hunk #3 succeeded at 105.
Hunk #4 succeeded at 118.
done
=> Verifying /srv/NetBSD/pkgsrc/shells/fish/patches/patch-src_builtins_ulimit.rs
**************************************
Ignoring patch file /srv/NetBSD/pkgsrc/shells/fish/patches/patch-src_builtins_ulimit.rs: invalid checksum
**************************************
=> Verifying /srv/NetBSD/pkgsrc/shells/fish/patches/patch-src_fork__exec_spawn.rs
=> Applying pkgsrc patch /srv/NetBSD/pkgsrc/shells/fish/patches/patch-src_fork__exec_spawn.rs
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|$NetBSD: patch-src_fork__exec_spawn.rs,v 1.1 2025/12/13 17:47:21 vins Exp $
|
|Prevent mismatched types for POSIX_SPAWN_SETSIGDEF on SunOS.
|
|--- src/fork_exec/spawn.rs.orig        2025-10-07 20:56:06.000000000 +0000
|+++ src/fork_exec/spawn.rs
--------------------------
Patching file src/fork_exec/spawn.rs using Plan A...
Hunk #1 succeeded at 116 (offset -1 lines).
done
=> Verifying /srv/NetBSD/pkgsrc/shells/fish/patches/patch-src_path.rs
=> Applying pkgsrc patch /srv/NetBSD/pkgsrc/shells/fish/patches/patch-src_path.rs
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|$NetBSD: patch-src_path.rs,v 1.1 2025/12/13 17:47:21 vins Exp $
|
|Use statvfs() on illumos.
|
|--- src/path.rs.orig   2025-11-24 15:30:10.536505320 +0000
|+++ src/path.rs
--------------------------
Patching file src/path.rs using Plan A...
Hunk #1 succeeded at 694.
Hunk #2 succeeded at 711.
done
=> Verifying /srv/NetBSD/pkgsrc/shells/fish/patches/patch-src_wutil_dir__iter.rs
=> Applying pkgsrc patch /srv/NetBSD/pkgsrc/shells/fish/patches/patch-src_wutil_dir__iter.rs
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|$NetBSD: patch-src_wutil_dir__iter.rs,v 1.1 2025/12/13 17:47:21 vins Exp $
|
|Struct dirent lacks d_type on SunOS.
|
|--- src/wutil/dir_iter.rs.orig 2025-11-13 12:09:28.000000000 +0000
|+++ src/wutil/dir_iter.rs
--------------------------
Patching file src/wutil/dir_iter.rs using Plan A...
Hunk #1 succeeded at 293.
Hunk #2 succeeded at 302.
done
ERROR: Patching failed due to modified or broken patch file(s):
ERROR:  /srv/NetBSD/pkgsrc/shells/fish/patches/patch-src_builtins_ulimit.rs
*** Error code 1

Stop.
make[1]: stopped making "replace" in /srv/NetBSD/pkgsrc/shells/fish
*** Error code 1

Stop.
make: stopped making "replace" in /usr/pkgsrc/shells/fish

==cut==

Thanks,

--
Marcin Gondek / Drixter
http://fido.e-utp.net/
AS56662

-----Original Message-----
From: tech-pkg-owner%NetBSD.org@localhost <tech-pkg-owner%NetBSD.org@localhost> On Behalf Of Paolo Vincenzo Olivo
Sent: Saturday, December 13, 2025 10:51 PM
To: tech-pkg%netbsd.org@localhost
Subject: Re: fish shell versioning

On 25/12/13 09:45AM, Greg Troxel wrote:
> 
> Having foo and fooN is bad, as it is not immediately obvious which is 
> which and why.  We end up there when a package transitions from "the 
> latest version is ok for everybody so we have one" to "the package has 
> made troubling decisions and there are reasons why the latest version 
> isn't ok".  You may point out that fooN and fooM also are somewhat 
> difficult to interpret, but IMHO the risk of confusion is much lower, 
> as people should take that as a clue to read foo*/DESCR.
> 
> While there's generally a bias against renaming, moving fish to fish3 
> seems reasonable.
> 
> However, the most important step is easy and seems not to have
> happened: reading DESCR of both does not explain any of this.   A simple
> "This is the current release.  It is implemented in rust and thus not 
> available on all platforms." and "This is a previous release which is 
> no longer maintained.  It is, however, available on platforms not 
> supported by rust." would do wonders for  understanding
> 

So, I went down this route, moved the packages as stated above and updated the DESCRs. We now have:

* shells/fish3: old version
* shell/fish: the new rust based one.

I think we can keep fish3 as long as a build failure or a bug doesn't come up.

Regards,
Paolo