There’s a superb CppCon talk by Matt Godbolt which covers largely the same premise and goes even deeper. An LLVM dev I look up to called it a must-watch: https://www.youtube.com/watch?v=dOfucXtyEsU
TIL you can have emoji section names, and even include ANSI escape sequences for funky objdump output. I’m now searching for an opportunity to use this knowledge.
As I understand it, you can’t have emoji section names or ASCII section names, section names are simply binary data. A tool may choose to interpret these as text in some character set and, by convention, the common names are byte sequences that are meaningful ASCII strings. Nothing relies on the names being valid ASCII, UFT-8, or any other character set.
There’s a superb CppCon talk by Matt Godbolt which covers largely the same premise and goes even deeper. An LLVM dev I look up to called it a must-watch: https://www.youtube.com/watch?v=dOfucXtyEsU
TIL you can have emoji section names, and even include ANSI escape sequences for funky objdump output. I’m now searching for an opportunity to use this knowledge.
As I understand it, you can’t have emoji section names or ASCII section names, section names are simply binary data. A tool may choose to interpret these as text in some character set and, by convention, the common names are byte sequences that are meaningful ASCII strings. Nothing relies on the names being valid ASCII, UFT-8, or any other character set.