hide more file writing details from engine #269
Conversation
Signed-off-by: tabokie <[email protected]>
Signed-off-by: tabokie <[email protected]>
Signed-off-by: tabokie <[email protected]>
…901-parallel-write Signed-off-by: tabokie <[email protected]>
Signed-off-by: tabokie <[email protected]>
|
/cc @LykxSassinator |
Codecov ReportBase: 97.65% // Head: 97.66% // Increases project coverage by
Additional details and impacted files@@ Coverage Diff @@
## master #269 +/- ##
==========================================
+ Coverage 97.65% 97.66% +0.01%
==========================================
Files 30 30
Lines 10645 10633 -12
==========================================
- Hits 10395 10385 -10
+ Misses 250 248 -2
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 at Codecov. |
Signed-off-by: tabokie <[email protected]>
Signed-off-by: tabokie <[email protected]>
| @@ -401,15 +413,11 @@ impl<F: FileSystem> SinglePipe<F> { | |||
| Ok(handle) | |||
| } | |||
|
|
|||
| fn maybe_sync(&self, force: bool) -> Result<()> { | |||
| fn sync(&self) -> Result<()> { | |||
There was a problem hiding this comment.
In current implementation of sync, it still gets the mutex for avoiding concurrent racing at the beginning.
According to #258 , will it be refactored to support concurrent accessing in later work ? i.e. replace mutex with Rwlock ? or add an extra F::Handle in SinglePipe, making it be called concurrently ?
Signed-off-by: tabokie <[email protected]>
Signed-off-by: tabokie <[email protected]>
|
Ping @Connor1996 , this one is blocking #261 |
This is a preparing work for #258.
Changes:
fsyncto file handle instead of writer, so that we can call it concurrently in the future. In doing so, we have to remove the sync offset tracking and deprecatebytes_per_syncsupportprepare_writeinto pipe_log. Introduce a traitReactiveBytesfor this purpose.