I'm interested in following along with this as well. Just learning about
WebAssembly's potential at the moment. It does seem that a currently missing component
is to harness fedora's infrastructure to build wasm targets and distribute wasm
artifacts. Capturing the steps necessary to build and rebuild libraries is going to be
necessary for automation and keeping current with security bugfixes. Distributing those
updated artifacts to infrastructure-or-parties building packages which rely on them is
going to be necessary to avoid wasteful duplication of effort.
I just tried to build proj, which failed because it needed sqlite. I found a fork of a
two-year-old release of sqlite that's been adapted to function in a peer to peer
network called fluent as well as compile to webassembly. There's been 31 releases of
sqlite in the interim, many of them during the year that the author of the fork was
committing bugfixes to the frozen original code. As this was an exercise, I quit here
because I learned all I needed to.
I may go a bit further than just saying that such an effort requires the use of
fedora's existing binary artifact infrastructure. You may need to create a centralized
place to maintain forks of the upstream repositories which contain branches dedicated to
the web assembly ports. The patching strategy currently used by rpmbuild may not be
enough--I'd expect you'd need a full DVCS hub. You'll want something you
control that facilitates transferring the web assembly support back up to the upstream,
hence the "D".
Please do add me to the list. I may not have time to do much more than think on problems,
but I'm interested in following your progress.