Commit bf2e4a4
authored
Defer error chain next_node creation (#191)
Try and optimize arena allloc by deferring it until it is guaranteed to
be needed (i.e. since many nodes get pruned by beam, no_revisit_dets,
etc. etc. in between)
For context, I am hunting down something that looks like a memory leak
right now in Tesseract. This popped up as a possible factor but I have
not yet measured the impact on memory (or runtime for that matter). But
wanted to get advice on it.1 parent 1e0cd87 commit bf2e4a4
1 file changed
Lines changed: 7 additions & 7 deletions
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
433 | 433 | | |
434 | 434 | | |
435 | 435 | | |
436 | | - | |
437 | | - | |
438 | | - | |
439 | | - | |
440 | | - | |
441 | | - | |
442 | | - | |
443 | 436 | | |
444 | 437 | | |
445 | 438 | | |
| |||
483 | 476 | | |
484 | 477 | | |
485 | 478 | | |
| 479 | + | |
| 480 | + | |
| 481 | + | |
| 482 | + | |
| 483 | + | |
| 484 | + | |
| 485 | + | |
486 | 486 | | |
487 | 487 | | |
488 | 488 | | |
| |||
0 commit comments