BUG: fix buffer overflow in data array of ldexp and frexp #4839
Conversation
the number types does not match the size of the blank array anymore, use a larger array and add asserts to ensure the size matches.
|
found with gcc's address sanitizer (-fsanitize=address) |
There was a problem hiding this comment.
static data is zero initialized
|
How about generating these ifucking like is done with the rest of them? |
|
I guess there is probably a reason for why they are not generated but I lack the time to figure it out |
BUG: fix buffer overflow in data array of ldexp and frexp
|
@ewmoore It would be nice to generate them, but the current code generation machinery doesn't allow for the needed specifation. |
|
FWIW there is a commit or two over in my stalled reworking of the complex npymath functions (#3010)that adds generating them. Also sorry about the typo'd curse above, that was unintentional and not directed at anyone. I'm on my phone. |
|
@ewmoore A PR improving the generator and moving those declarations elsewhere would be welcome. |
|
@ewmoore That typo (ahem) almost motivated me to buy a Mac. |
the number types does not match the size of the blank array anymore, use
a larger array and add asserts to ensure the size matches.