IRC channel logs
2025-09-20.log
back to list of logs
<stikonas>there isn't anything controvertial there, so I'll merge it once CI passes (unless you merge it first) <stikonas>next I think I need to bump snapshot of tinycc-0.9.26.. <fossy>Googulator: yeah, I would like optional modules at some point, but there's a few tricky considerations there <fossy>especially ensuring different combinations of optional modules don't change checksums of the final packages <fossy>any particular reason for RPM? <matrix_bridge><Andrius Štikonas> Yeah, it will be harder to keep build checksums valid if we have different build options <matrix_bridge><Andrius Štikonas> Probably everything optional should come after core stuff <matrix_bridge><Andrius Štikonas> Yeah, I'm also confused how rpm hrlos here <fossy>yes for sure, one model i was considering is having optional packages BUT instead of having the user selecting particular optional packages, the user would select an optional "add-on manifest" which encodes the list optional packages (and then the checksums are linked to the optional manifest) <matrix_bridge><Andrius Štikonas> In principle dependencies can be added fairly readily <matrix_bridge><Andrius Štikonas> But for core path with few divergencies (except for kernel bootstrap) they don't bring that much benefit... <matrix_bridge><Andrius Štikonas> tsort can be used to resolve dependencies... <fossy>that's along the lines of another alternative, just using dependencies for everything and then allow the user to specify a set of final pacakges <fossy>but i don't know how beneficial that is <fossy>because that certainly makes some things a bit more annoying (particularly keeping track of dependencies and optional features that depend on what other packages are available) <lanodan>Maybe could some paths could be extra scripts? That said maybe they could be third-party quite like the scripts/documentation dedicated to bootstrapping a specific distro.