Documentation of Missing Features.

[zephraph]

I'm going to go ahead and start a list of features I find that are missing from reql. This will be updated as we go. Please let me know if I missed anything!

Stream

Not Implemented

r (rethink object, may not be desired)

• [ ] dbCreate
• [ ] dbDrop
• [ ] dbList

  Database

• [ ] tableList

  Table

• [ ] indexList
• [ ] indexRename
• [ ] indexStatus
• [ ] changes
• [ ] update
• [ ] replace
• [ ] delete (_Note: Exists, but is currently delete)
• [ ] sync

  Sequence

• [ ] innerJoin (Note: It looks like both of these functions are implemented in a single join method)
• [ ] outerJoin
• [ ] offsetsOf
• [ ] union (?)

[neumino]

Thanks for putting that up @zephraph.

So I have recently started to work again on reqlite. I'm sorting out a few things first. but I should be clear to push in 2-3 weeks. I don't remember all the details, but I can provide a list of things that are currently not available in my branch

Stuff that I should be able to finish for next release

• [x] geospatial operation -- baring rounding errors + distance point to line/polygon
• [ ] changefeeds -- That's requires some plumbing
• [ ] Syncing to disk -- I would actually use it for stuff like rethinkdbdash/thinky
• [ ] Fake system tables

Things I would like to do

• [ ] Use a B-trees instead of a hash table to index documents -- That would actually help a bit for lazily evaluating streams.
• [ ] Actually build B-trees for secondary indexes

Things I'm not sure

• [x] r.js with a lambda function -- I have a simple r.js command,
• [ ] r.http because it is the only asynchronous operation and is a pain-ish to do.

Things I don't plan to do (but I would merge a PR if someone does it)

• [ ] Real clustering operations -- I don't plan to implement them for now as you can't cluster reqlite
• [ ] Lazily evaluating streams
• [ ] r.range() followed by a method that keep the stream type -- because it would require a decent chunk of work/refactoring and is kind of out of the scope of reqlite for now.

As far as I know, all the rest work. There might be a few assertions missing for types, but that's all. The test framwork was reworked a bit. The test are basically now running queries against reqlite and rethinkdb and compare the output.

[zephraph]

/cc @Slava

@neumino, I'm really wanting to see baked in support for rethinkdb in meteor so I'll do everything that I can to facilitate this library. Feel free to assign me tasks if you would like me to investigate or work on certain tasks.

[Slava]

I think we need more tests on the queries. At this point I am not even sure if basic operations like "add" and "orderBy" work w/o bugs. Fixing the existing tests would be a good start. Also, @neumino, are there tests we can adopt from the main rethinkdb repo?

[neumino]

@Slava -- Yes, for now I wrote the queries myself to make sure that all the code paths were tested. I plan to just pull all the tests for the JS driver and dump it in reqlite.

[neumino]

Opened a lot of issues to track done what's left. Closing this meta issue.

[neumino]

@Slava -- Could you check that _saveOriginal still works for your need? It was a big merge ;-)

[neumino]

@zephraph -- if you want to help now, there are tons of things to grab.

[zephraph]

I see that, haha. I'll look through the open tasks.