Eli Schwartz

> pacman vercmp alpha '' results in -1 (alpha < "") [FS#49093](https://bugs.archlinux.org/task/49093) Submitted fix at https://lists.archlinux.org/pipermail/pacman-dev/2017-December/022211.html At least, empty strings are still considered valid version strings, but passing only one...

> **pacman** `vercmp alpha ''` results in `-1` (`alpha < ""`) [FS#49093](https://bugs.archlinux.org/task/49093) What's the dead end here? `vercmp alpha ` is now a fatal error in the latest stable release....

Great, we're now even closer to "all dead ends lead to aurweb". :D

@orbea no comment on your "polite disagreeing with the choice of meson and being honest that their rude response was not appreciated". However regarding the specifics of your objections against...

- I don't understand how "python is hard to cross compile for $PLATFORM" is intended to be an argument that "projects using meson are hard to cross compile for $PLATFORM...

- You still seem to be hung up on somehow needing to cross-compile python from $BUILD_MACHINE to $TARGET in order to run meson on $BUILD_MACHINE using the $BUILD_MACHINE python and...

Is this just #2324 and #2318 again? For what it's worth, I simply don't understand the initial resolution at all. The testsuite should not be writing into the main project's...

This used to be undocumented and an accidental side effect of leaky python implementation details casting. It's officially supported since https://github.com/mesonbuild/meson/commit/6298e2907082aff0763a09165c6c2841ac869e2d and was backported to 0.60.1 So, > graphene 1.10.8...

> meson is annoying to me as a user (for other software). I have to re-patch it every release. Aside: Why would you need to patch *any* software "every release"?...

> For instance we still have things like python 3.6 and manylinux1 and we don't have any plans ATM to be dropping either. Hmm, the opposite was said in https://github.com/pygame/pygame/pull/3166#pullrequestreview-963104323