Conversation
0841683 to
ed49a2a
Compare
ed49a2a to
b9abcd4
Compare
|
@tpunt, good catch! So you think there's really no chance to fix the pre-condition? As adding one more condition to the loop which can have many iterations doesn't look nice. I haven't look deep, but there should be a way to have a formula like for highest/lowest value to double check, just for the case. Thanks. |
|
I've refactored the patch so that the range of values are calculated from the array size calculation (rather than checking the intermediary Thanks, |
|
Dups with #1677 ? |
|
Ok thx for details |
There was a problem hiding this comment.
Ok, so __calc_size dosen't need to go out of scope anymore.
There was a problem hiding this comment.
Correct. I'll rectify that now (as well as in PR #1677).
|
@tpunt can we merge them please? It is the same function, so it'll be easier for having an overview. Thanks. |
|
I mean merge two PRs into one :) Thanks. |
|
Ok, I'll merge them now and submit a new PR. |
|
Superseded by PR #1695. |
The segfaults are caused by precision loss of large longs being converted to doubles when the
stepparameter is a double. The for loops continue infinitely since thestepbeing added/subtracted upon each iteration is too small to be represented accurately as a double.The segfaults can be reproduced with:
This fixes Bug #71197