got me thinking about the possibility of losing trust in open source software.
Mine was the horrible node.js platform. I come from Sphere, a Game development engine with an embedded JS interpreter to add code to actions (triggers/collisions/etc) and so I love ECMA-script, despite it's shortcomings.
But when I saw the amount of libs you need to pull in to run somewhat simple programs... it's an audit nightmare.... quite buggy too (things might have improved, the project was new)
I also had fights with the Perl community. I use only vanilla perl, as no server can import stuff from the internet/CPAN ... let alone compile it (no compiler on servers). Pure-perl libs are ok. On the plus side, my code works on older unix machines.
And then... Python came, and we can pull in a lot. Maven came, and we can pull in a lot... (but with security analyzer tools before we can build a release).
And for ansible... Galaxy has a lot... some are bad quality, or not flexible enough, so you give up and roll your own (then again, maybe I like flexibility too much). And then they release a new version of Ansible, and the thing you needed is right there in the core... grr so much wasted time.
I've burned myself too with proprietary software. TurboPascal4 (student release) , when exiting the program, wrote the FAT to your floppy. So If you had swapped the software disk with your code-disk... your code was now only readable with lowlevel disk block sector readers (Thanks PCTools5!).
And a new version of MSVC actually made code NOT compile. And I never found the bug, just weird abends when running a release-mode; while debug modes ran fine (but too slow) so I suspect a race condition somewhere. The version of MSVC after that did not grok the old project. It was abandoned we could never replicate the old toolset.