Commit Graph

723 Commits

Author SHA1 Message Date
Franck Royer
9ded728879
The first message is actually a quote or rate request 2021-02-12 17:05:06 +11:00
bors[bot]
c117885dad
Merge #183
183: Introduce binary for nectar r=D4nte a=D4nte



Co-authored-by: Franck Royer <franck@coblox.tech>
2021-02-11 22:19:38 +00:00
Franck Royer
ee43125bdd
Add nectar binary 2021-02-11 15:24:38 +11:00
bors[bot]
66d9c7c2cb
Merge #182
182: Segregate CLI specific modules r=D4nte a=D4nte

To prepare for the introduction of nectar modules.

Co-authored-by: Franck Royer <franck@coblox.tech>
2021-02-11 04:08:55 +00:00
Franck Royer
fac5b59d17
Remove unnecessary specification of the types 2021-02-11 10:34:44 +11:00
Franck Royer
7805a6d684
Move cli specific modules under cli folder
nectar will also have its own folder with a separate config.
2021-02-11 10:34:44 +11:00
Franck Royer
901c9e89c9
No evident added value of having separate Seed structs
Hence, reducing complexity of the codebase. Note that the seed will be
used by both nectar and the cli whereas the config mod will be different
so this changes helps with the next step of having a dedicated config
module for each binary.
2021-02-11 09:57:17 +11:00
Franck Royer
7d392c3086
This files contains Command and Arguments structs 2021-02-11 09:47:42 +11:00
Franck Royer
45dccb8be2
Rename the config struct Config 2021-02-11 09:38:29 +11:00
bors[bot]
fadbf1638a
Merge #177
177: Nectar r=D4nte a=D4nte

Re-work Alice do be the Service Provider (nectar).
This means that Alice's event loop runs all the time and processes swap requests.

- We currently use a hardcoded fixed rate
- The binary for nectar is to be done in follow-up PR
- We removed the message acknowledgement feature as it created issues with other work and some work would have been needed to link a ack to the correct swap on Alice side.
- Because we removed the acks, we had to remove a number of restart tests. This can be re-introduced at a later stage once a decision on how to best test the restarts can be done
- We remove the punish test as it is a nectar (alice) feature and current focus is on the CLI (bob).

Co-authored-by: Franck Royer <franck@coblox.tech>
2021-02-10 05:43:34 +00:00
Franck Royer
eb39add5ff
Fix typo 2021-02-10 16:42:55 +11:00
Franck Royer
6e6dc320b4
Alice event loop now handles the creation of new swaps 2021-02-10 16:42:55 +11:00
Franck Royer
15eb9a2fe4
Remove punish test
The punish test needs re-work due to the fact that Alice runs continuously

Currently focusing on the CLI (Bob), so we can re-introduce this test
once we want to ensure that nectar (Alice) punishes.
2021-02-10 16:42:55 +11:00
Franck Royer
3bc8b58b6a
Remove Bob restart tests after communication
The test do not work without acks as we stop the event loop as soon
as a message is considered as "sent" when actually the event loop
and swarm may not have yet sent the message.

The ack allow to avoid this issue as the message was considered "sent"
only once the other party sent a response. However, the ack brings
other issue so a review needs to be done to select the appropriate
solution.
2021-02-10 16:42:55 +11:00
Franck Royer
fd9f633a77
Remove Alice restarts tests
Current focus is on CLI UX. Fair amount of change needs to happen to
cater for Alice (nectar) restart scenarios.
2021-02-10 16:42:55 +11:00
Franck Royer
bfc19d5628
Remove acknowledgements processing
We are aware of issues of timeouts when waiting for acknowledgements.
Also, to properly supports acks in a multiple swap context, we need to
revert to doing event processing on the behaviour so that we can link
leverage the `RequestResponse` libp2p behaviour and link the messages
requests ids to swap ids when receiving an ack or response.

Acks are usefully for specific scenarios where we queue a message on the
behaviour to be sent, save as sent in the DB but crash before the
message is actually sent. With acks we are able to resume the swap,
without ack, the swap will abort (refund).
2021-02-10 16:42:55 +11:00
Franck Royer
cc8b855117
Make it possible to clone a handle
This will be used for new swaps.
2021-02-10 16:42:54 +11:00
Franck Royer
1b2be804ed
Remove unnecessary channels
`alice::swap::run_until` will be called once the execution setup is
done. The steps before are directly handled by the event loop,
hence no channels are needed for said steps: connection established,
swap request/response & execution setup.
2021-02-10 16:42:54 +11:00
Franck Royer
f5ca5faabf
Process execution setup failure similarly to other failures
By merging it in the failure event of the root behaviour.
2021-02-10 16:42:54 +11:00
Franck Royer
4ade5df0e5
Remove unnecessary impl block 2021-02-10 16:42:54 +11:00
Franck Royer
60e0b9382c
Introduced from float API for Monero quantities 2021-02-10 16:42:54 +11:00
Franck Royer
b5b990257a
Move Behaviour in dedicated module 2021-02-10 16:42:54 +11:00
Franck Royer
87be9aeb2a
Prepare separation of event loop initialisation to swap initialisation
As for Alice, the event loop will be started with the program and will
be the one starting swaps (`run_until`) based on libp2p events (swap
request).
2021-02-10 16:42:54 +11:00
Franck Royer
530b9b2ea8
Remove possible mix up of timelocks when using State0::new 2021-02-10 16:42:53 +11:00
Franck Royer
f35ed436ce
Allow EventLoop to hold a database ready to pass to alice::Buider
The `EventLoop` will use the `Builder` interface to instantiate a
`Swap` upon receiving a `SwapRequest` and successfully doing an
execution setup.
Before this change, the `EventLoop` would have to hold the path to the
db and re-open the db everytime it wants to construct a swap.

With this change, we can open the DB once and then hold a
`Arc<Database>` in the `EventLoop` and pass it to new `Swap`s structs.
2021-02-10 16:42:53 +11:00
Franck Royer
39a46baa2c
Preemptively box cancel tx to avoid size difference in enum 2021-02-10 16:42:53 +11:00
Franck Royer
69363e43a3
Preemptively box encrypted signature to avoid size difference in enum 2021-02-10 16:42:53 +11:00
Franck Royer
fd084b764d
Move generation of keys inside State0::new
The event loop will now use this function so I want to simplify its
usage to avoid having to instantiate too many items to use it.
2021-02-10 16:42:53 +11:00
Franck Royer
788445964a
Move main.rs to cli.rs to prepare for nectar binary 2021-02-10 16:42:51 +11:00
Franck Royer
311ba74cd6
Remove sell-xmr from CLI 2021-02-10 16:41:38 +11:00
bors[bot]
639b3169d8
Merge #179
179: Watch only tmp monero wallet r=da-kami a=da-kami

When initializing Monero wallet, try to open temporary wallet, if opening errors then try to create it. If creating errors then fail with message asking the user to configure the wallet RPC.
If opening / creation succeeds get the block_height to verify that connection to wallet works, then proceed to swap. 

Note: RPC error handling is hacky, but I don't see the value of investing time into that at this point. The problem is, that we don't deal with the json rpc errors properly for calls were we don't serialize the result into an object (because the response is empty). Eventually we should switch to using https://github.com/thomaseizinger/rust-jsonrpc-client but that does not support parameters `by-name` yet, see: https://github.com/thomaseizinger/rust-jsonrpc-client/issues/20

Co-authored-by: Daniel Karzel <daniel@comit.network>
2021-02-10 04:09:11 +00:00
Daniel Karzel
56a59bc41c Remove newline in use statements 2021-02-10 15:07:37 +11:00
Daniel Karzel
0a0bce1552 Rename temporary wallet to blockchain monitoring wallet 2021-02-10 15:06:26 +11:00
Daniel Karzel
1e29433bd2 Open or create temporary Monero wallet upon wallet initialization 2021-02-09 16:34:50 +11:00
Daniel Karzel
dac4443bbd Add functionality to open monero wallet through rpc 2021-02-09 16:34:50 +11:00
Daniel Karzel
62605a318a Add CreateWallet trait to expose create_wallet 2021-02-09 16:34:50 +11:00
bors[bot]
27bb9498d5
Merge #170
170: Cancel and refund commands r=da-kami a=da-kami

I plugged the cancel and refund logic into the current state/state-machine logic of the swap.

## Follow ups  (out of scope)

We might want to record issues to be tackled later, since we are on a tight time budget :)
Please let me know what you think @D4nte @rishflab 

### Problems with `ack` after sending a message

Alice was waiting forever when awaiting the `ack` from bob when sending the lock proof in case she runs into a dial error. It seems the `acks` can cause the program to hang. This is a severe problem that we most probably will encountered in production at some point. For this PR I wrapped the `ack` of Alice upon sending the `encsig` in a timeout to work around this problem, see 7463081f88 - but **we might want to consider to remove all `ack` message. I don't see much value in them if we don't have a resilient retry strategy.**

### Do not require Monero wallet for cancel/refund

The cancel/refund commands don't require a monero wallet.
In this PR we re-uses the builder which requires the monero wallet as well - and we check for the monero balance upon wallet initialization, so the command will fail if no monero wallet is started.

### Save Alice connection info in Bob DB

Save Alice's peer-id/address in DB: It's cumbersome for the user to lookup those details again.




Co-authored-by: Daniel Karzel <daniel@comit.network>
2021-02-09 02:40:47 +00:00
Daniel Karzel
53fcd9224c Give finding the refund tx priority over waiting for the punish transaction to be finalized.
This was introduced due to a CI run, where Bob included tx_refund, but Alice had waited until T2 had expired,
and then went for punishing Bob instead of refunding.
Weirdly, Alice's punich transaction did not fail in that scenario.
2021-02-09 12:13:43 +11:00
Daniel Karzel
86290649e7 work in review comments 2021-02-08 21:20:58 +11:00
Daniel Karzel
c930ad84a4 Add --force flag for cancel and refund 2021-02-08 21:19:38 +11:00
Daniel Karzel
02f8eb7f18 Add test for cancel/refund before timelock expired 2021-02-08 21:18:37 +11:00
Daniel Karzel
2d5d70d856 Timeout for Alice waiting for ack for sending transfer proof
If dialing Bob fails Alice waits for the acknowledgement of the transfer proof indefinitely.
The timout prevents her execution from hanging.
Added a ToDo to re-visit the ack receivers. They don't add value at the moment and should be removed.
2021-02-08 21:18:37 +11:00
Daniel Karzel
c9adbde5d5 Add test for Bob's manual cancel and refund 2021-02-08 21:17:56 +11:00
Daniel Karzel
ad2aefc2a5 Refund Command 2021-02-08 21:16:32 +11:00
Daniel Karzel
0600a7048e Cancel command 2021-02-08 21:11:39 +11:00
bors[bot]
ec661178f4
Merge #173
173: Use libp2p-async-await to improve API of execution setup phase r=D4nte a=D4nte



Co-authored-by: Franck Royer <franck@coblox.tech>
2021-02-07 22:17:11 +00:00
Franck Royer
dee8c1e335
Remove scopes 2021-02-08 09:16:31 +11:00
Franck Royer
a48823a665
Re-organise Alice state to be more coherent with the msg sequence 2021-02-08 09:16:31 +11:00
Franck Royer
eefb1b3b16
Use event_process = false for Bob
As we do not process the event, we can just implement the needed `From`
traits.
2021-02-08 09:16:31 +11:00
Franck Royer
a7b89e2fe4
Use event_process = false for Alice
As we do not process the event, we can just implement the needed `From`
traits.
2021-02-08 09:16:29 +11:00