social.noyu.me

here’s the least frustrating of those bugs: https://sourceware.org/bugzilla/show_bug.cgi?id=30385

here’s the much more frustrating one: https://sourceware.org/bugzilla/show_bug.cgi?id=30386

in fairness I might be the only person who’s ever needed this 🤔

I’ll build gdb from source so I can begin the process of fixing these bugs myself, I thought. It’ll be easy, I thought.

who’s surprised that the makefile doesn’t work if you have spaces in the path (sigh)

checking for memcpy... yes

I got gdb to build… eventually. most of the problems came from Apple shipping outdated stuff with macOS. though I also had to hack the source code to work around a C++ overload resolution issue. 😮‍💨

@hikari "just making sure there's no CVEivors"

@hikari you're fixing bugs in GDB, I think you're past the realm of easy,

@hikari Good luck building it with a recent SDK, there’s a file called VERSION that messes with a C++ include in the standard library for <version> and breaks things

@hikari btw your second bug looks familiar I’m pretty sure I ran into it during a CTF once

@saagar haha. a friend told me the other bug was familiar to them. I guess few people bother to report things :)

@hikari Normally I do try to report things and sometimes send in patches but either I forgot to do it that time or I probably found another bug report for it

@saagar the C preprocessor and its consequences have been a disaster for the human race

@hikari I found a Windows bug in GNU ld a few years ago.

Short version (pun intended!): A "short" was used in a place where PE technically expects "unsigned short".

I didn't even bother to file a bug b/c I know any change is going to be too invasive.

@hikari that's the fun part of programming tbh!

we keep finding bugs in wabt/wasm2c because of working on wasm2kotlin :3

@hikari when you have to gdb gdb you know you're having a bad bad day