9 May
2025
9 May
'25
7:21 a.m.
On Thu, 2025-05-08 at 11:31 -0700, Joe Damato wrote:
Nit: it looks like in the original we didn't care about malloc potentially failing. Do we care about checking for that now with this cleanup?
Thank you for the review, Joe.
I looked a bit into adding error messages and I think it wouldn't make sense just for these allocations, as there are others which do not check for malloc/calloc error (e.g. all generated ynl libs).
Furthermore, I am under the impression that on the kind of systems ncdevmem would run, malloc/calloc would almost never fail because of overcommit.
Finally, even if they did fail, the user program would just segfault (not very user friendly, but good enough).
Cosmin.