You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am usually working in an environment where custom allocators are used for various reasons and usage of plain malloc/free is frowned upon. Thankfully many libraries allow their users to supply custom (de)allocation routines.
While support for this can be as simple as some defines that default to malloc/free, I always found that in practice it is way more convenient to be able to pass some user data (often just a void*) to the (de)allocation routines (similar to what is suggested in #356). See for example PCRE2 that does this upon context creation (e.g. pcre2_general_context_create).
If you feel that this would fit MIR and is worth pursuing, I could try to put together a pull request for this :-)
The text was updated successfully, but these errors were encountered:
Sorry for long delay with the answer, I overlooked your message. I think custom allocation would be useful for some applications. So it is worth to pursue this. I only have one important requirement is to save somehow the old library interface for compatibility.
Thanks for your feedback :-)
I am currently a bit busy with some other stuff, so it might be a while until I have something to show.
Just to clarify: by library interface you mean the functions/types in mir.h and mir-gen.h and the like, but not e.g. mir-htab.h and mir-varr.h, correct?
I am usually working in an environment where custom allocators are used for various reasons and usage of plain
malloc
/free
is frowned upon. Thankfully many libraries allow their users to supply custom (de)allocation routines.While support for this can be as simple as some defines that default to
malloc
/free
, I always found that in practice it is way more convenient to be able to pass some user data (often just avoid*
) to the (de)allocation routines (similar to what is suggested in #356). See for example PCRE2 that does this upon context creation (e.g.pcre2_general_context_create
).If you feel that this would fit MIR and is worth pursuing, I could try to put together a pull request for this :-)
The text was updated successfully, but these errors were encountered: