one of the important aspects of this project is it being possible to
replicate (for both offline real-time collaboration as well as legacy
archiving purposes) all resources.
bugzilla is one major piece of critical functionality that is also a
critical online resource.
about the only offline bugtracker that exists (and is any good) is
was going to half jokingly suggest the impractical (and probably insecure) measure of putting the bugzilla installation in a public git repo -- bugs-everywhere looks much better, though I'd be concerned that we need a working, secure, public https interface for those who don't want to install bugs-everywhere locally and figure out how to use it. I know bugs-everywhere has a interactive web interface, I just don't know if it properly supports adding new users, authentication, etc.
another option is fossil, which is a distributed version control system with a wiki and bug tracker:
(In reply to Jacob Lifshay from comment #1)
> was going to half jokingly suggest the impractical (and probably insecure)
> measure of putting the bugzilla installation in a public git repo --
sigh it'd involve dumping - then purging - parts of the bugzilla postgresql
dump. bleh :)
> bugs-everywhere looks much better, though I'd be concerned that we need a
> working, secure, public https interface for those who don't want to install
> bugs-everywhere locally and figure out how to use it.
i'd want it anyway.
> I know bugs-everywhere has a interactive web interface,
yes. it's... "functional".
> I just don't know if it properly supports
> adding new users, authentication, etc.
wark-wark :) not at all. i believe it's there for "convenience", for
when you're running locally, and want to browse bugs, comment on them etc.
locally then "git push" out to merge with other users.
sigh i kinda promised myself i wouldn't mess about with web frameworks any
more :) however if my arm was sufficiently twisted i could likely throw
something together that's suitable for public use, maybe even (gosh)
using the bugzilla postgres user/pass to do it (!)
hm i wonder if we can justify putting in an NLNet Grant request specifically
for things like this, on the basis that, really, we do need full independence.
we could then find - and pay - someone to do these kinds of things.
(In reply to Jacob Lifshay from comment #2)
> another option is fossil, which is a distributed version control system with
> a wiki and bug tracker:
sigh i'd really like it... if they hadn't decided to reinvent git.
UNIX philosophy: do one thing, and do it well. the majority of
other "project management" systems all allow alternative revision
control to be used. an all-or-nothing approach does not inspire
I spoke to Luke about this by happenstance on Sunday, he says that we should mirgate to Fossil-scm after the Oct 2020 tapeout. Deferring until after then.
Michiel mentioned a protocol for a decentralized github-like protocol, allowing pull requests, issues, etc. Seems really interesting and worth investigating!
It allows users of any ForgeFed-compliant service to interact with other ForgeFed-compliant forge services, without being a registered user of that foreign service, just as if they were. In this way, users that choose to self-host have the additional benefit/responsibility of fully controlling of their own authentication/identityand their own data.
briiilliant. at last. it would mean that anyone could submit a bugreport on github... *without* being forced to choose between "terms and conditions that they do not agree with" and "being entirely excluded from the community".