I'd prefer #NetBSD: https://schmonz.com/2024/06/07/small-arms/

Staged latest shairport-sync for #pkgsrc. Builds on NetBSD, #macOS. Normally I'd commit, wait for evbearmv6hf-el binary package, forget.

Trying something new today: https://cdn.netbsd.org/pub/pkgsrc/current/pkgsrc/doc/HOWTO-use-crosscompile

Anyone running my arm64 macOS binary packages from https://pkgsrc.smartos.org/install-on-macos/ on Sonoma 14.5 or newer may want to upgrade to the newer package sets that I've now built:

$ sed -i -e 's/12.3/14.5/' /opt/pkg/etc/pkg_install.conf /opt/pkg/etc/pkgin/repositories.conf
$ pkgin -f update
$ pkgin upgrade

I'll get new bootstrap kits and a proper announcement done soon, but this is all you need to do if you already have the 12.3 packages installed.

The 14.5 SDK is now required by some software.

#pkgsrc

@ermo

I was pointing out the poor state of the #pkgsrc doco in #NetBSD a little while ago. This is the area where #FreeBSD is equally bad with #pkgng.

FreeBSD's pkg-create(8) manual page assures us that +MANIFEST files must have "file" entries one per file. But the actual code for the pkg command only looks for a "files" object with an array of files.

Making manifest and "build-info" files for these tools is not a case of doing what the doco says to do.

https://github.com/freebsd/pkg/blob/main/libpkg/pkg_manifest.c#L114

Dear #NetBSD people:

pkg_create(1) takes files with its -f, -B, and -b options. But there is zero documentation anywhere for what is supposed to be in the latter two.

Please document this, preferably on that manual page alongside where the contents of the first file are documented.

#pkgsrc

The pace of software is accelerating, and it's no longer possible to produce binary package repositories that are as backwards compatible as they used to be.

I'm shortly going to have to bump my macOS package sets up to a baseline of 14.5, as the current target of 12.3 is no longer sufficient for certain C++20 features like std::ranges::sort which is now required by print/poppler.

First bulk build is in:

https://reports.pkgci.org/Darwin/14.5/arm64/20250724.1420/meta/report.html

I was kinda hoping it'd fix more tbh e.g. numpy.

#pkgsrc #macos

I should commit the other changes in my local pkgsrc tree... #pkgsrc

Toot from pkgsrc/www/firefox-141.0 under NetBSD/amd64-current... #NetBSD #pkgsrc

Nice optimisation now available for anyone using my macOS binary packages from https://pkgsrc.smartos.org

Packages that require fortran support now only depend on the smaller gcc-libs package (7MB) rather than the full gcc package (300MB).

Once you've "pkgin upgrade"ed, remember to "pkgin autoremove" to clean up the now-unused gcc package.

This brings it in line with the SmartOS package sets.

Continuing to push back against bloat one step at a time. Next up, ghc.

#pkgsrc #macos #illumos

@rubenerd I use Radicale and have maintained it in #pkgsrc. I had to write my own authenticator/authorisation to give granular ro/rw ACLs, but it works well with @davx5app and #Thunderbird

My local change for pkgsrc/lang/gcc14 under NetBSD/i386 9 works fine. #pkgsrc #NetBSD

Hmm... pkgsrc/lang/gcc14 on NetBSD/i386 9 is broken... I will import a patch from pkgsrc/lang/gcc12... #pkgsrc

I really wish Apple would stop breaking things.

Latest Command Line Tools breaks yacc, m4, etc.

When you accept the installer it prints a completely bogus estimation time.

Then when it finishes, yacc, m4, etc, are still missing and you go around the cycle again.

This really isn't difficult Apple. With a market cap of over 2.5 trillion dollars, I'm pretty sure you can afford to employ a single #pkgsrc developer. They would find these bugs within minutes before release.

#Apple #macOS

Alt...

Screenshot of Command Line Tools installer displaying "Downloading software. About 40 hours, 21 minutes remaining."

Just to give you an idea of how much software GCC 14's default of -Werror=implicit-function-declaration breaks, here are two pkgsrc bulk build results:

GCC 13.3.0: 24835/28450
GCC 14.3.0: 2589/28450

https://reports.pkgci.org/SmartOS/upstream/trunk/20250630.2248/meta/report.html

https://reports.pkgci.org/SmartOS/upstream/trunk/20250701.2248/meta/report.html

I also had to fix up a bunch of things before the build would even start.

This could take a while. And then the fun starts all over again with GCC 15's changes...

#pkgsrc #c

Of course I made a mistake when tagging the #pkgsrc release 😑 I can't wait for CVS to go away.

@futurebird How about an Amiga 4000 running #NetBSD running nanotodon?

This machine happens to be one of the machines that compiles m68k #pkgsrc binaries for NetBSD.

I love that these machines are still quite useful in 2025, whether running NetBSD or AmigaOS. There was just an update to AmigaOS in March, and I took this machine apart fo install new AmigaOS 3.2.3 ROMs (with Kickstart 47.115).

It’s getting harder and harder to find modern programs for classic Mac OS or even PowerPC Mac OS X, but people are keeping SSL, ssh, usable browsers and email clients, games and all sorts of other things up to date on Amigas.

It finally finished in just under 66 hours.

And now to do it all over again because lang/rust got changed during a freeze.

#pkgsrc

Still going...

pkgsrc-nb0# ps -o etime= -p 9270
2-02:48:30

50 hours and counting...

#pkgsrc #netbsd

@Tubsta @stefano The #pkgsrc Dovecot maintainer plans to update mail/dovecot2 to 2.4. I think it would be better to add a separate clashing mail/dovecot24 package. If the main package goes to 2.4, I may choose to maintain a dovecot23 package instead

Bulk builds on #NetBSD now even slower thanks to www/chromium.

Currently been building for 27 hours with no end in sight.

Nothing else comes even remotely close, second longest being libreoffice at a mere 11 hours.

#pkgsrc #bloat

Toot from pkgsrc/www/firefox-140.0b9 under NetBSD/amd64-current ... #NetBSD #pkgsrc

Busy week upgrading all VMs to #NetBSD 10.1 to allow switching to HVM from PV (as #XenServer and @xcpng have dropped support for PV and unfortunately do not support PVH either) which means moving to uEFI booting rather than pygrub loading the kernel. Just waiting for the next #pkgsrc quarterly release to move the remainder over with nice fresh packages.

lang/pear has to be one of the most frustrating pieces of software to package correctly.

Pretty much every avenue I try to go down ends up with a big fat "nope".

Current status is unpacking, patching, repacking, and then running .phar files, but that appears to only half work (the stub runs but then it can't include support files, even though they are in the archive).

Feels like they go out of their way to make this as difficult as possible.

#pkgsrc

MXkill is an X/Motif based interface to ps, with support for signals and regex parsing.
Now available on pkgsrc.

#x11 #motif #retrotech #pkgsrc

Alt...

mxkill window, showing output for`ps eaux | sort +0 -1', with all running processes indexed. A "Regular Expression" bar at the bottom allows to parse regex. Motif widgets, with a theme mimicking IRIX desktop on mwm.

I have just enabled debug option for pkgsrc/lang/perl5. It enables -DDEBUGGING and perl -D. #pkgsrc

Toot from pkgsrc/www/firefox-139.0.1 under NetBSD/amd64-current. #pkgsrc #NetBSD

If you have missed The NetBSD Foundation 2025 Annual General Meeting you can read more and find logs here!:

https://blog.NetBSD.org/tnf/entry/agm2025

#netbsd #pkgsrc

@rohare em was the first editor I ever used, so I updated the sources from v6 to modernish (the last argument to seek(2) originally used options to work in blocks, not bytes, for example). I added em to #pkgsrc a while ago, so it should be avalable for all platforms that pkgsrc supports

pkgin 25.5.0 released.

Includes some important fixes to SUPERSEDES and CONFLICTS support, and "upgrade" will now update any core package tools first.

Should correctly handle the recent php-X.Y -> phpXY-X.Y renames, and fix some other corner cases.

https://github.com/NetBSDfr/pkgin/blob/master/CHANGES.md#version-2550-2025-05-19

#pkgsrc #pkgin

A wild #blog post appears!

I revived pkgsrc on AIX.

(Yes, I already sent a PR to NetBSD.)

https://briancallahan.net/blog/20250516.html

#linux #unix #bsd #freebsd #openbsd #netbsd #dragonflybsd #aix #ibm #solairs #illumos #pkgsrc #opensource #freesoftware

You've probably seen the local root privilege escalation vulnerability in GNU screen(1):
https://www.openwall.com/lists/oss-security/2025/05/12/1

The note there suggests that #NetBSD ships with a vulnerable version of screen(1). This is incorrect: NetBSD includes screen(1) in #pkgsrc as a _possible add-on package_ you can choose to install. It does _not_ include screen(1) in the base system.

Next version of pkgin will check for upgrades to any installed packaging tools first.

https://github.com/NetBSDfr/pkgin/commit/e5849d8f52215b36a8f8ed6d80e31c15178cff81

Wishing I'd done this sooner so that upcoming fixes for SUPERSEDES would be automatically handled :( Oh well, hopefully will still prove to be useful in the future (or maybe not!)

#pkgin #pkgsrc

NetBSD 10.x kernel MATH_EMULATION
#NetBSD #linux #RunBSD #pkgsrc

https://mezzantrop.wordpress.com/2025/02/04/netbsd-10-x-kernel-math_emulation/

@yrabbit That's for `rsync-ssl`.

(In general, if you have the #pkgsrc tree extracted around you can probably grep for `REPLACE_BASH` and see possible scripts that needs bash :))

Yet to hear anything positive about GNU screen 5.x.

Drops support for #Solaris / #illumos (now has a hard requirement on openpty()), and breaks various use-cases on #NetBSD.

Switching all my builds back over to misc/screen4.

#pkgsrc

A gentle reminder of how important pre-caching expensive mk variables is.

Time to run `bmake pbulk-index` in pkgtools/digest on a stupidly fast Mac Studio.

Uncached:

real 0m4.856s
user 0m1.828s
sys 0m1.731s

Cached:

real 0m0.088s
user 0m0.052s
sys 0m0.025s

This is why your scan phases are so slow.

#pkgsrc

In dreckly we only integrate commits after they have been shown to cause no regressions.

To do that we have automated testing for the following platforms:

macOS (arm64 + x86_64, ppc coming soon!)
Ubuntu
Cygwin(!)
CentOS
NetBSD
SmartOS
FreeBSD
OpenBSD

This gives us a good spread of platforms, compilers, and architectures, and we've already fixed many bugs as a result.

Every commit should always improve a project and move it forwards, not sideways or backwards.

#dreckly #pkgsrc

#pkgsrc 2025Q1 will be the last branch to support several platforms listed below,
due to bitrot and lack of use. If you'd like to speak up, please use the
following threads:

#MirBSD https://mail-index.netbsd.org/tech-pkg/2025/03/12/msg030607.html
GNU/kFreeBSD https://mail-index.netbsd.org/tech-pkg/2025/03/11/msg030600.html
#Interix https://mail-index.netbsd.org/tech-pkg/2025/03/08/msg030572.html
BSD/OS https://mail-index.netbsd.org/tech-pkg/2025/03/17/msg030646.html