Catch calls to initialize an oversize heap

[satur9nine]

Suggested fix for #77

[rhempel]

Hey thanks for submitting this PR, I am in the middle of a business trip, but should be back to integrate this next week.

I'm also working on improving the docs a bit so that it fits into my adaptabuild system so it's a good time to go over this module.

Do you mind if I ask where you are using this library? I gave a presentation on a separate topic a while back and out of about 30 attendees, two were using the library :-)

[satur9nine]

Hello, I just discovered this library a few weeks ago so I haven't adopted it for any project yet. Jacob

…

On Sun, Nov 30, 2025, 08:18 Ralph Hempel ***@***.***> wrote: *rhempel* left a comment (rhempel/umm_malloc#78) <#78 (comment)> Hey thanks for submitting this PR, I am in the middle of a business trip, but should be back to integrate this next week. I'm also working on improving the docs a bit so that it fits into my adaptabuild system so it's a good time to go over this module. Do you mind if I ask where you are using this library? I gave a presentation on a separate topic a while back and out of about 30 attendees, two were using the library :-) — Reply to this email directly, view it on GitHub <#78 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAYIW3424OBVJGQFPZ273TD37MKELAVCNFSM6AAAAACNSC5QTOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTKOJSG43TEMBZGU> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[rhempel]

I have implemented the check for oversized heap, but we need to still decide how to handle it when it happens.

My personal vote is either calling a user-defined failure function (preferred) or hanging (less preferred) or segfaulting (last resort)

[satur9nine]

I have implemented the check for oversized heap, but we need to still decide how to handle it when it happens.

My personal vote is either calling a user-defined failure function (preferred) or hanging (less preferred) or segfaulting (last resort)

You mentioned an idea in #77

We could consider having the init_xxx() APIs return true/false which would break existing code, or we can add an API for check_init_xxx() tjhat returns true/false .

I think new check_init_ functions that return int would be ideal (int that returns 0 on success is a bit more C standard).