Split Bolt11InvoiceDescription into a version used with references to
the description or description hash in the invoice and an owned version
of these for when constructing an invoice. The latter is useful as it
removes an unnecessary clone and can be used in a future change
specifying either a description or description hash in larger set of
invoice parameters. Since it doesn't use a reference, it can be exposed
in bindings as well.
When we serialize from a byte array to bech32 in
`lightning-invoice`, we can either copy the array itself into the
iterator or hold a reference to the array and iterate through that.
In aa2f6b47df we opted to copy the
array into the iterator, which is fine for the current array sizes
we're working with, but does result in additional memory on the
stack if, in the future, we end up writing large arrays.
Instead, here, we switch to using the slice serialization code when
writing arrays, (very marginally) reducing code size and reducing
stack usage.
In aa2f6b47df we refactored
`lightning-invoice` de/serialization to use the new version of
`bech32`, but in order to keep the public API the same we
introduced one allocation we could have skipped.
Instead, here, we replace the public `Utf8Error` with
`FromUtf8Error` which contains the original data which failed
conversion, removing an allocation in the process.
In aa2f6b47df we refactored
`lightning-invoice` de/serialization to use the new version of
`bech32`, also reducing some trivial unnecessary allocations when
we did so.
Here we drop a few additional allocations which came up in review.
Now that we don't have to have everything in our entire ecosystem
use the same `std`/`no-std` feature combinations we should start by
untangling our own features a bit.
This takes another step by removing the `no-std` feature entirely
from the `lightning-invoice` crate and removing all feature
implications on dependencies from the remaining `std` feature.
`lightning-invoice` previously had a dependency on the entire
`lightning` crate just because it wants to use some of the useful
types from it. This is obviously backwards and leads to some
awkwardness like the BOLT 11 invoice signing API in the `lightning`
crate taking a `[u5]` rather than a `Bolt11Invoice`.
Here we finally rectify this issue, swapping the dependency order
and making `lightning` depend on `lightning-invoice` rather than
the other way around.
This moves various utilities which were in `lightning-invoice` but
relied on `lightning` payment types to make payments to where they
belong (the `lightning` crate), but doesn't bother with integrating
them well in their new home.
`lightning-invoice` currently has a dependency on the entire
`lightning` crate just because it wants to use some of the useful
types from it. This is obviously backwards and leads to some
awkwardness like the BOLT 11 invoice signing API in the `lightning`
crate taking a `[u5]` rather than a `Bolt11Invoice`.
This takes tees us up for the final step, adding a
`lightning-types` dependency to `lightning-invoice` and using it
for imports rather than the `lightning` crate.
In a coming commit, the `lightning-invoice::utils` module will move
to the `lightning` crate, causing its tests to be included in the
global lockorder tests done in that crate. This should be fine,
except that the `lightning-invoice::utils` module currently holds
the `added_monitors` lock too long causing lockorder violations.
Instead, this commit replaces the legacy monitors-added test with
the `check_added_monitors` test utility.
Use the `hex-conservative` crate directly from `bitcoin` instead of from
`hashes`. Although it makes no real difference it is slightly more clear
and more terse.
The `hex` crate is re-exported by `rust-bitcoin` so we can get it from
there instead of explicitly depending on it. Doing so reduces the
maintenance burden and helps reduce the likelyhood of getting two
versions in the dependency graph.
The `rust-bitcoin` project is working towards making the public API
separate from the directory structure; eventually the
`bitcoin::blockdata` will go away, to make maintenance easier here stop
using the `blockdata` module.
Do not run the formatter, so as to make review easier. This patch was
created mechanically using:
search-and-replace bitcoin::blockdata bitcoin
and having defined
```bash
search-and-replace () {
if (($# != 2))
then
echo "Usage: $0 <this> <that>"
return
fi
local this="$1"
local that="$2"
for file in $(git grep -l "$this")
do
perl -pi -e "s/$this/$that/g" "$file"
done
}
```
Previously, we would require our users to handle all events
successfully inline or panic will trying to do so. If they would exit
the `EventHandler` any other way we'd forget about the event and
wouldn't replay them after restart.
Here, we implement fallible event handling, allowing the user to return
`Err(())` which signals to our event providers they should abort event
processing and replay any unhandled events later (i.e., in the next
invocation).
We previously stated in the docs that the invoice description can be at most `1023`
bytes long, which is wrong. According to BOLT 11 it's at most 1023*5 bits (639 bytes) long.
99aa6e27f6 detected that we had an
undefined feature in `lightning-invoice` called `strict`, which was
used to turn on `deny(warnings)`. It resolved that by adding the
feature to the `Cargo.toml`, but we actually don't need it - our CI
already builds with `-Dwarnings`, so any warnings should be
rejected during CI and there's not a lot of value in having a
(public) feature to do the same.
This uses the newly introduced conditional configuration checks that are
now configurable withint Cargo (beta).
This allows us to get rid of our custom python script that checks for
expected features and cfgs.
This does introduce a warning regarding the unknown lint in Cargo
versions prior to the current beta, but since these are not rustc errors,
they won't break any builds with the "-D warnings" RUSTFLAG.
Moving to this lint actually exposed the "strict" feature not being
present in the lightning-invoice crate, as our python script didnt
correctly parse the cfg_attr where it appeared.
When parsing lightning-invoice HRPs we want to read them
char-by-char, tracking at which offset different fields were. Prior
to this commit this was done first by reading char-by-char and then
by indexing using the byte offset which works for ASCII strings but
fails on multi-byte characters.
This commit fixes this issue by simply always walking byte-by-byte
and rejecting multi-byte characters which don't belong in HRPs.
The `PaymentHash`, `PaymentSecret`, `PaymentPreimage`, and
`ChannelId` types are all small wrappers around `[u8; 32]` and are
used throughout the codebase but were defined in the top-level
`ln/mod.rs` file and the relatively sparsely-populated
`ln/channel_id.rs` file.
Here we move them to a common `types` module and go ahead and
update all our in-crate `use` statements to refer to the new
module for bindings. We do, however, leave a `pub use` alias for
the old paths to avoid upgrade hassle for users.
New rustc beta now warns on duplicate imports when one of the
imports is from a wildcard import or the default prelude. Thus, to
avoid this here we prefer to always use `crate::prelude::*` and let
it decide if we actually need to import anything.
`parse_int_be` is generic across integer types and also input
types, but to do so it relies on the `num-traits` crate. There's
not a lot of reason for this now that std has `from_be_bytes`, so
we drop the generic now and replace it with a macro which is called
twice to create two functions, both only supporting conversion from
`u5` arrays.
`lightning-invoice` was mostly written before std's `from_be_bytes`
was stabilized, so used its own `to_int_be` utility to do int
conversions from `u8` arrays. Now that the std option has been
stable for quite some time, we should juse use it instead.
Preallocate for 8 items in the vec. I chose this value for
1. features
2. description
3. payment hash
4. expire time
5. min_final_cltv
6. payment secret
7. route hint
8. for the memes
While this isn't expected to materially improve performance, it
does get us ahash 0.8, which allows us to reduce fuzzing
randomness, making our fuzzers much happier.
Sadly, by default `ahash` no longer tries to autodetect a
randomness source, so we cannot simply rely on `hashbrown` to do
randomization for us, but rather have to also explicitly depend on
`ahash`.
