I have never seen a program segfault and crash more than apt. The status quo is extremely bad, and it desperately needs to be revamped in some way. Targeted rewrites in a memory safe & less mistake-prone language sounds like a great way to do that.
If you think this is a random decision caused by hype, cargo culting, or a maintainer's/canonical's mindless whims... please, have a tour through the apt codebase some day. It is a ticking time bomb, way more than you ever imagined such an important project would be.
I've been using apt regularly on Debian for a long time and never seen it crash or segfault. Very strange that you do. All software has bugs of course, but apt is so heavily used that I expect it gets attention. It just works for me.
I’m very very curious to know what it is you’re doing to experience this: I’ve used Debian and its derivatives for 25 years now. On desktops, laptops, and servers. x86, x86-64, and Arm 64. I have never had a segfault with APT. Not a single time. Problems with dependencies or such a few times, but I don’t recall APT ever crashing on me.
Almost 30 years for me, both on my personal machines and at work.
I went through a period about 25 years ago where apt crashed on my (rather janky) desktop almost every other run, and sometimes left my system in a state so inconsistent that I had to fall back on 'dpkg-reconfigure --force' and the like to fix it.
Turns out that it was due to a bad interaction between a failing stick of RAM and reiserfs' tail-packing feature, which was causing frequent silent corruption in /var/lib/dpkg/status and friends.
I don't think I've seen any similar issues since, across what must be many millions of apt runs I've been responsible for.
Perhaps gp is suffering from some similar underlying problem?
If you think this is a random decision caused by hype, cargo culting, or a maintainer's/canonical's mindless whims... please, have a tour through the apt codebase some day. It is a ticking time bomb, way more than you ever imagined such an important project would be.