Drop futures dependency from lightning-block-sync
#2141
Conversation
|
Note, as an alternative, |
|
As a further alternative, we could use sync mutexes with an |
Some how I'd understood that `futures` had reasonable MSRV guarantees (e.g. at least Debian stable), but apparently that isn't actually the case, as they bumped it to upgrade to syn (with apparently no actual features or bugfixes added as a result?) with no minor version bump or any available alternative (unlike Tokio, which does LTS releases). Luckily its relatively easy to just drop the `futures` dependency - it means a new connection for each request, which is annoying, but certainly not the end of the world, and its easier than trying to deal with pinning `futures`. See rust-lang/futures-rs#2733
TheBlueMatt
force-pushed
the
2023-03-fuck-rust
branch
from
8699320
to
78c73b0
Compare
|
Actually, went ahead and did the second because its super trivial. |
|
Grr, also need to replace the futures usages in the async BP. I think its quite doable though. |
TheBlueMatt
force-pushed
the
2023-03-fuck-rust
branch
from
a041fc9
to
cf0e7dc
Compare
|
Ok, removed it from BP wholesale in two commits. Honestly dunno why we were relying on a whole mess of proc macro in the futures crate to implement select lol. |
In general, only one request will be in flight at a time in `lightning-block-sync`. Ideally we'd only have one connection, but without using the `futures` mutex type. Here we solve this narrowly for the one-request-at-a-time case by caching the connection and takeing the connection out of the cache while we work on it.
TheBlueMatt
force-pushed
the
2023-03-fuck-rust
branch
from
cf0e7dc
to
152ac15
Compare
|
That said, I'm open to pushback on the last commit (or the second one) - it does add a trivial use of unsafe, which is trivial, but technically its not required to fix MSRV, it just leaves us open to futures breaking us again in the future. |
|
|
||
| const DUMMY_WAKER_VTABLE: RawWakerVTable = RawWakerVTable::new( | ||
| dummy_waker_clone, dummy_waker_action, dummy_waker_action, dummy_waker_action); | ||
| pub(crate) fn dummy_waker() -> Waker { unsafe { Waker::from_raw(RawWaker::new(core::ptr::null(), &DUMMY_WAKER_VTABLE)) } } |
There was a problem hiding this comment.
what would not introducing this unsafe look like?
There was a problem hiding this comment.
Using another crate which has the unsafe in it instead. There's no way to poll a future in rust with zero unsafe. Now, we could consider coming up with some other way to do no-std timers inside an async context outside of polling a timer future, but I'm not 100% sure what is cleaner for users.
Codecov ReportPatch coverage:
📣 This organization is not using Codecov’s GitHub App Integration. We recommend you install it so Codecov can continue to function properly for your repositories. Learn more Additional details and impacted files@@ Coverage Diff @@
## main #2141 +/- ##
==========================================
- Coverage 91.38% 91.37% -0.02%
==========================================
Files 102 102
Lines 49759 49767 +8
Branches 49759 49767 +8
==========================================
+ Hits 45472 45474 +2
- Misses 4287 4293 +6
... and 2 files with indirect coverage changes Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report in Codecov by Sentry. |
`futures` recently broke our MSRV by bumping the `syn` major version in a patch release. This makes it impractical for us to use, instead here we replace the usage of its `select_biased` macro with a trivial enum. Given its simplicity we likely should have done this without ever taking the dependency.
TheBlueMatt
force-pushed
the
2023-03-fuck-rust
branch
from
152ac15
to
718ab54
Compare
As `futures` apparently makes no guarantees on MSRVs even in patch releases we really can't rely on it at all, and while it currently has an acceptable MSRV without the macros feature, its best to just remove it wholesale. Luckily, removing it is relatively trivial, even if it requires the most trivial of unsafe tags.
TheBlueMatt
force-pushed
the
2023-03-fuck-rust
branch
from
718ab54
to
491100d
Compare
|
Addressed arik's comments and squashed. |
| // not without awaiting, we need a Waker, which needs a vtable...we fill it with dummy values | ||
| // but sadly there's a good bit of boilerplate here. | ||
| fn dummy_waker_clone(_: *const ()) -> RawWaker { RawWaker::new(core::ptr::null(), &DUMMY_WAKER_VTABLE) } | ||
| fn dummy_waker_action(_: *const ()) { } |
There was a problem hiding this comment.
given these are dummies, would it make sense to insert some unreachable!s?
There was a problem hiding this comment.
Not quite, they're still called, but they're no-op's.
Some how I'd understood that
futureshad reasonable MSRV guarantees (e.g. at least Debian stable), but apparently that isn't actually the case, as they bumped it to upgrade to syn (with apparently no actual features or bugfixes added as a result?) with no minor version bump or any available alternative (unlike Tokio, which does LTS releases).Luckily its relatively easy to just drop the
futuresdependency - it means a new connection for each request, which is annoying, but certainly not the end of the world, and its easier than trying to deal with pinningfutures.See rust-lang/futures-rs#2733