Problem: The latest release seems to have some index problems, which is
frustrating, but might be able to be circumvented with a database
rebuild. The rebuild was something that Ahau funded and I'm excited to
be able to finally expose this to users without asking them to `rm -rf`
random files in their `~/.ssb` directory.
Solution: Replaced unused 'progress' bars (always at 100%) with rebuild
button.
Problem: There were lots of CSS changes in the preview + blob PR and I
didn't take enough time to look at them. It looks like a few of them
made style problems (e.g. extra specing in the 'next' / 'previous' links
on the profile pages) and it made it harder (IMO) to determine that
links are links (since we've been using font weight to distinguish).
Solution: Revert those changes back to the previous defaults. If the
margin glitch was on purpose or anyone feels strongly about
distinguishing links using some other style (color?) then I'm happy to
merge those, but I'd like to fix the extra margin and have *some* way to
distinguish links in the mean time.
Problem: With the new ssb-db@20 we have to explicitly add ssb-private1
to ensure that we can see private messages.
Solution: Add ssb-private1 as a plugin.
Problem: Our tests seem to have been disabled in various places, which
means that type errors, typos, and other small problems were introduced.
Solution: Remove the comments and fix the underlying problems without
disabling the linters.
Problem: I've been worried about merging these version bump pull
requests since we don't have many tests, and I *really* don't want to
merge any catostrophic breaking changes anywhere.
Solution: Add some tests that are more comprehensive, starting by
creating a temporary SSB database and keypair and then editing the
profile, previewing a message, publishing a message, etc., until we have
a handful of small functions actually being tested. This won't ensure
that everything works forever, and it *really* doesn't test replication,
but it should help increase our confidence that test success means that
fewer things are broken.
Problem: I thought it would be a good idea to keep our SSB stack in
another module but it just creates friction and adds another layer to
our stack that has to be independently upgraded and versioned.
Solution: Merge the @fraction/flotilla source code into Oasis so that we
don't have to maintain two separate projects.
Problem: You can't publish email addresses or user@host combos or
anything like that because the regex is super liberal in what it
matches.
Solution: Ensure that the first word in a message is a mention *or*
ensure that there's whitespace before a mention. This also adds a
regex101 link (in lieu of proper testing, for now) which makes this
easier to test and track over time.
See-also: https://github.com/fraction/oasis/issues/466
Problem: Recently there was a PR [0] merged with a quickfix to avoid
some test failures, which is something I've been trying to avoid. While
`process.exit()` works fine, I'm worried that it means we don't
understand what's happening under the hood, *plus* I have the [maybe
unjustified?] worry that it might kill the process during a database
write or something dangerous. It looks like this particular test hang
was caused by both a stream and some number of intervals that were left
open.
Solution: Provide a way to close the stream and intervals in `index.js`
and ensure that we do that before closing the server.
[0]: https://github.com/fraction/oasis/pull/462
after having restored the package-lock.json to that of the master branch, this
PR now uses common-good@2.0.3, which has common-good check as a command.
i ran it manually as $(npm bin)/common-good check, and fixed whatever
it complained about.
we had duplicates and _old_ names in the first attempt,
with this new strucutre only one name per feed is kept (the most recent)
also: tweak regexp some more
should not correctly match end of string and NOT match inside md link
https://regex101.com/r/Uu6iTj/3/
replaces immediatly if there is one match,
but just spews multiple ones.
* add silly in-memory about:name cache
rational is also commented the code. the tl:dr; is oasis already doesnt
use ssb-about or other indexing plugins.
butt for @mentions we need a quick way to lookup names to feeds,
otherwise we block the preview flow with long query times.
Jose Fabio
Supernova
Christian Bundy
Daan Wynen
Alexander Cobleigh
Henry