waffenklang

Make a german version (with germany related tax, stock and inflation data) and I "shut up and take my money."

No, Op seem to have mistaken B and M.

And thats fine for you! But as a keyboard maker System76 could have make it optional as a lot of other keyboard maker are able to do.

No ISO, no money.

Mh, thanks for the repost. I'm curious if anything has actually improved since 2014? Whats the state of the art on the graphene front?

Wow.. I'm a bit ashamed as I searched for algorithms on that topic but was obviously unable to get the right keywords to find that one. Thanks for pointing it out. I will look into that one more deeply, but on the first glance its basically the same algorithm.

you made my day.

didnt read the whole comments; so sorry, if duplicate.

After a short review of the add function for handling of complex structs i found a possible bug as they miss to perform a deep copy of objects containing references to other objects. they would only copy the value of the pointers of a struct like that: struct { obj* ptr; } and not the content of ptr.

I agree in the point that its the task of the fv-system to take care of this, and so its a flaw in it, if its necessary.

imho,in this case it is a flaw, as the 'public' api function base64_encode wants the user to input pointer to 2 ptr to buffer with only for 1 given the length. and the code makes implicit assumption of the user input without checking it (encoded is manipulated).

Good example for the weaknesses of the 'formally verified' sale strategy. It may be formally correct, but the code is full of weaknesses, like no NULL checks, no out of bounds checks, one endless loop based on pointer arithmetic..

good and fast c code looks not like this.