Is there a reason we don't support the alternate form (#) flag in kernel printf?

rzezeski: No particular reason. Probably just need.

We can add it and teach gcc about the support for it.

k, I ask because the Chelsio common code uses it in some places. I can just update those places for now, but would be nice to have that support. I'll create an issue.

fenix illumos#17614

FEATURE 17614: add support for alternate form flag in kernel printf (New)

↳ illumos.org/issues/17614

#17586 (fenix?) is another similar case that bit me recently

FEATURE 17586: mdb_printf should support %z (New)

↳ illumos.org/issues/17586

if someone brave wanted to commonize our doprnts

i have a from scratch *printf implementation.. though it doesn't support the i18n bits

i ended up writing it for some UEFI stuff

but i borrowed the idea (definitely not the code) from ast where there's a struct that represents a conversion specification

so each one gets parsed, then the struct is used to do the work

it made it cleaner and easier to follow (IMO at least)

Just make sure there's ample test suite investment here given the high blast radius.