-
Notifications
You must be signed in to change notification settings - Fork 421
Meeting Notes 2025 04 28
Elias Rohrer edited this page Apr 28, 2025
·
1 revision
- 0.1.3
- Need to do backports and release notes, that’s it
- Small router fixes that seem to be rare as well as one issue where we can hit a panic in handling invoice requests when paying a bolt 12 offer
- 0.2 (https://github.com/lightningdevkit/rust-lightning/milestone/17)
- No updates
- May ship when 0-fee commitments gets done but will depend when the next significant feature ships
- 0.5 (https://github.com/lightningdevkit/ldk-node/milestone/5)
- Development was a bit slow from tnull being off and then traveling, but we already shipped 0.5-rc1 a week or so ago, and will push off 0.5 release tmrw as soon as LDK 0.1.2 lands
- Hopefully will get back to review soonish although will be traveling this/next week too, will get back to it asap
- Will do a release with the new release today, language bindings in a day or two
- Attributable failures
- No progress on LDK side but got updates from acinq and LL that they are working on it
- Acinq has a new PR on it, +300 diff to get the feature in for them. Pretty compact
- Async payments
- Arik: been busy last week but back on it refactoring the trampoline forwarding PR taking advantage of the outbound payment module to get free retries
- Got informed by shaavan that in a PR he’s working on there’s a regression in a trampoline unit test, so taking a look at that as well
- RGS
- Been stable, may take it off the agenda
- VSS
- Dual funding
- Jkczyz meeting with dunxen on wednesday about this, just need review from one outstanding PR from wpaulino. Finishing DF for accepting
- Have a draft PR for contributing inputs out
- Splicing
- Moving along, paused last week for travel
- Jeff’s PR on splice_locked handling should be good for review soon
- Working with leo on figuring out some of the monitor refactors we want to do for HTLC sources when we have multiple commitments to sign for
- Have a design, just need to write the code now. Should unblock monitor work
- Expect to have updates by the next sync
- Wvanlint: looking into memory issues with LDK java bindings
- Previously, people were experimenting with jdmalloc. There was instruction to use ld_preload for that, but didn’t have the expected results
- LDK java bindings: are they statically compiled? Does it include the allocator as well, meaning we can’t switch it out with LD_PRELOAD?
- Matt: don’t think so. See whether it links libc. If you dump the symbols it links against, should see malloc in there, should be dynamically linked
- M: based on previous data that showed consistent slow growth over time, didn’t seem as likely to be mem fragmentation where jdmalloc would solve it, seemed like an actual leak. Should chat offline more