I'm coming around to the idea of just having inline checks for "callee has not changed" at every inlining point in my #emacs jit. Slower, but a zillion times simpler than on-stack replacement.
Programs that send --help output to stderr are a pet peeve of mine. Latest offender: ninja.
Working on lldb for a couple of months has reinforced my belief in the need for good comments.
A time traveler tells me that USB-Q is the really good one and that I should wait.
In a few contexts now I've wanted to separate string storage management from the string APIs. I suppose lldb's string type proliferation is a way to work around this.
In lldb I think the string types you can use are std::string, ConstString, StringRef, SmallString, CachedHashString, CachedHashStringRef, and Twine. Though I think I may be missing one.
I'm looking at inlining for my #emacs JIT, but not seeing a good way to do this without implementing on-stack replacement, which seems hard, especially given that I'm using libjit rather than rolling my own back end.
When I hear Elephant Talk now, it just sounds like a list of messaging apps that I haven't heard of yet
I got my first (trivial) Rust expression to evaluate in lldb today.
My theory is that Satoshi is the aliens from The Three-Body Problem and that they are going to have us destroy ourselves via bitcoin mining.
@JordiGH @amphetamine
There are some decent gdb tutorials and GUIs. Most should be reasonably rust-savvy out of the box. Feel free to report problems to me, fixing the Rust debugging story is my current job.
