3 ms·
Checked C
- bluejekyll 10y agoI'm torn. I suppose it's easier to upgrade existing code bases to this, but it won't fix code locations that are improperly upgraded to this language. Which in my opinion means that The C code will still have lurking errors. Also: "either (1) the operation must convert that value to an in-range integer value or (2) the expression shall produce a runtime error" Is more defined, but does not actually help you much, because this will depend on platforms, and a overflow to 0 might produce similar issues when indexing as other errors today in creating bugs in encryption algorithms. If we all agree that C is not a safe language, and that we'd all prefer writing in a safe language, then why not write in a safe language, e.g. Rust. An aside: I've recently done a project where I explored replacing C library functions with Rust implementations, linking in the Rust library directly to the existing C library. I was very pleasantly surprised at the ease of doing this! This means that we really can migrate away from C in an iterative fashion, just as Checked C would allow us to do.
- Animats 10y agoDidn't we just cover this on HN? It's sort of halfway between C and C++, with constructs like array_ptr<T>
- rwmj 10y agoSince my paper on Bounds Checking C (Jones & Kelly, 1997) was mentioned in their bibliography, I've put my original 1995 BSc report that started the whole thing here: http://oirase.annexia.org/bounds/ http://oirase.annexia.org/bounds/ Original was a MS Word 6 document (!)