Handle null raw_choice - fixes #235 #246
Conversation
ccmain/control.cpp
Outdated
| @@ -795,6 +795,8 @@ static int SelectBestWords(double rating_ratio, | |||
| bool debug, | |||
| PointerVector<WERD_RES>* new_words, | |||
| PointerVector<WERD_RES>* best_words) { | |||
|
|
|||
|
|
|||
There was a problem hiding this comment.
What is the reason for empty lines?
|
I've cleaned up the cosmetic issues. Any insight into what the root cause is? I'm not sure that this is more than a bandaid. |
|
I am not sure if it is an good idea to return valid output (0.0) when tesseract did not produced certainty... |
|
@theraysmith Will this PR help with the word getting dropped - #681 |
|
It's not related at all to #681. |
|
Can somebody (@Shreeshrii ?) check if this is relevant for 4.0 (if it helps)? |
|
I will request @stweil to review. Thanks. |
|
Link to issue: #235. |
|
I added the issue to the planning list and will have a look next week. |
|
@tfmorris : do you have test case for this PR? |
|
@zdenop Sorry for the delayed reply. I do not have a unit test for the PR. It's 2 1/2 year old, but from looking at the historical context, I'm guessing that I was able to reproduce the failure in #235 and manually tested using that test case. Since the test case in #235 was tested and no longer crashes, I don't think this PR is needed. |
|
The segfault is gone, but the other part of the issue still exist with the 'tessdada' repo (but not with best / fast) |
|
In any case, patches that don't solve the root of the issue are problematic IMO. |
I'm not sure this is the best fix for the issue, but it does provide a workaround while someone works out what the root cause is and where it should really be fixed.