Re: JSON/YAML/TOML/etc. parsing performance

[Paul Eggert]

Philipp Stephani wrote:
> Should I at least add an eassert to document this?

I wouldn't. Many calls to memory allocators would have problems if they request 
more than PTRDIFF_MAX bytes, given the problems that C programs have when doing 
pointer arithmetic on large objects. It would be a waste of time to document 
this in every call by doing an eassert. Simply calling a memory allocator that 
is documented to not return such objects, should make it clear to the reader 
what is going on.

> I've attached a new patch (which currently segfaults on
> decode_coding_gap, but the call to that function doesn't seem to be
> required anyway).

Thanks, I plan to take a look at it after the decode_coding_gap issue is 
addressed.