
Boosted by jsonstein@masto.deoan.org ("Jeff Sonstein"):
cliffle@hachyderm.io wrote:
This evening I'm on an odd historical sojourn into the history of the tar format (thanks, @th, for inadvertently nerd-sniping me).
The problem is, this involves reading old C code, because SeLf DoCuMeNtInG or whatever. And I can't not see ghosts in C code anymore.
Anyway, guess how many things enshrined in later standards (like POSIX) are a product of what appear to be coding mistakes, including buffer overruns and use of uninitialized stack memory, in the V7 Unix tar implementation?
As an industry, we've _never_ been able to hold this tool (C) correctly.