Shd/pointer slot check #52
Conversation
| @@ -40,7 +42,27 @@ impl PointerCache { | |||
| self.pointer_map.insert(ptr, Some(addr)); | |||
| } | |||
|
|
|||
| pub fn ensure_up_to_date_ptr(&self, ptr: &ShelleyAddressPointer) -> Result<()> { | |||
| pub fn update_block(&mut self, blk: &BlockInfo) { | |||
There was a problem hiding this comment.
Now we know what the start slot is, do we still need to capture it here? I guess it might be a good idea to check that what we thought it was actually is the case, though!
There was a problem hiding this comment.
There are two parts of the module:
- Replaying already collected pointer cache.
- Creating pointer cache (e.g., for test network/sancho network). We do not know eras boundaries in that situation, and need to collect them.
This code is used only in second situation (in handle_deltas call, which is used only in stateful init).
|
|
||
| pub fn ensure_up_to_date_ptr(&self, blk: &BlockInfo, ptr: &ShelleyAddressPointer) -> Result<()> { | ||
| if ptr.slot > blk.slot { | ||
| // We believe that pointers cannot point forward |
There was a problem hiding this comment.
Suggest a warn! here so we can at least know this has happened
There was a problem hiding this comment.
The idea is that we've already collected the necessary info about pointers, and here we just replay the collected info. A pointer cannot point to future -- so if it happens then our database is up-to-date in respect to this.
There are some forward pointers in blockchain: and we need to skip errors about them (that's the reason for these changes).
|
|
||
| if let Some(conway_start_slot) = self.conway_start_slot { | ||
| if ptr.slot >= conway_start_slot { | ||
| // Conway epoch slots cannot be referenced |
There was a problem hiding this comment.
Do we really need to have warnings about incorrect pointers in blockchain? They will be shown again and again each invocation. Another idea: we can collect the errors and record errors that should be suppressed.
| return Ok(()); | ||
| } | ||
| } | ||
|
|
There was a problem hiding this comment.
Is this test now redundant given we're testing against pointing forward above?
There was a problem hiding this comment.
This test remains reasonable: if the database for pointers is not up-to-date, then we should write about that.
|
All understood, missed the different phases. |
|
Oops. |
No description provided.