deprecate solana-stake-program
#8860
Conversation
|
The Firedancer team maintains a line-for-line reimplementation of the |
this commit backports the stake program excision from solana-labs#8834 but leaves the crate intact with a warning
2501babe
force-pushed
the
20251103_depr_stake
branch
from
2a262c8 to
23a9b3b
Compare
| @@ -1,11 +1,13 @@ | |||
| #![cfg_attr( | |||
| not(feature = "agave-unstable-api"), | |||
| deprecated( | |||
| since = "3.1.0", | |||
| note = "This crate has been marked for formal inclusion in the Agave Unstable API. From \ | |||
| v4.0.0 onward, the `agave-unstable-api` crate feature must be specified to \ | |||
| acknowledge use of an interface that may break without warning." | |||
| ) | |||
| #![deprecated( | |||
| since = "3.1.0", | |||
| note = "\ | |||
| This crate will be REMOVED in Agave v4.0.0 and its name will be taken by the BPF Stake Program. \ | |||
| Users of this crate's public exports must switch to: | |||
| * `stake_state` -> `solana_stake_interface::state`. | |||
| * `points` -> `solana_runtime::inflation_rewards::points`. | |||
| * `id()` / `check_id()` -> `solana_sdk_ids::sysvar::stake::_`. | |||
| * `get_minimum_delegation()` will continue to live at the same path. | |||
| * Some convenience functions such as `stake_from()` and `create_account()` are removed." | |||
| )] | |||
There was a problem hiding this comment.
this is the only change original to this pr
2501babe
changed the title
solana-stake-program removalsolana-stake-program
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## v3.1 #8860 +/- ##
=========================================
- Coverage 83.2% 83.2% -0.1%
=========================================
Files 864 865 +1
Lines 375142 375290 +148
=========================================
+ Hits 312240 312246 +6
- Misses 62902 63044 +142 🚀 New features to boost your workflow:
|
t-nelson
left a comment
t-nelson
left a comment
There was a problem hiding this comment.
ideally this would have gone into master in a way that could be directly backported in full. that is, first as a deprecation + usage removal, then the actual crate deletion independently. so we can just mergify bp the former basically 1:1 with master. too much trouble to back out of master and put it back in in parts?
Co-authored-by: Trent Nelson <[email protected]>
|
ok i made the three prs to revert -> deprecate -> remove but i think we should just land this. we would have to change the commit that |
|
Hey @bw-solana we've got an "expedient" vs "follow the proces" backport debate ☝️. Any stance? |
|
If I'm understanding correctly, these are the 3 PRs that could get merged and backported: Agree with Trent generally speaking. But the fact that the revert now collides with another change makes me think the faster and lower risk thing is to just let this go in as is. The fact Hana went through the exercise tells me this isn't a lazy person whining their way into a bad precedent 😄 |
t-nelson
left a comment
t-nelson
left a comment
There was a problem hiding this comment.
it would appear i've been overruled
|
@grod220 mind giving this one sme approval again? |
6 participants
Problem
we remove the last stake program code in 4.0, to free up the name to be reused
Summary of Changes
make the deprecation unconditional on the unstable api flag (although it was right: not stable!)
i decided to backport the code changes from #8834 to stop using
solana-stake-programin agave because reusing this work as-is seemed a lot cleaner than marking some hundred+ callsites#[allow(deprecated)]. this pr is identical to the 4.0 one, except that we do not delete the stake program cratecloses #8841