Pierre-Clément Tosi

Hi, I'm the author of daniel5151/gdbstub#109, which was written for `crosvm` (see the [CLs on Gerrit](https://chromium-review.googlesource.com/c/crosvm/crosvm/+/3785468)). @michael2012z, given that you've implemented a very similar `Arch`, it would be very interesting...

> One (currently unimplemented) feature that could be particularly useful for AArch64 is , which I describe in https://github.com/daniel5151/gdbstub/issues/12#issuecomment-991195719. That's a great idea: ship all of the XML in the...

> how this would affect `type Registers`? From my basic understanding of the GDB protocol (you surely know more than I do), I had assumed that `Registers` (`{read,write}_registers`) represented the...

> rather, a target may send back however many registers it wants, and GDB will map said data onto registers based on their number value (as determined in the target.xml...

> Instead, I'd just swap out the unreachable for a `None`, which while not perfect, seems reasonable. Sounds good, I dropped 3796559ac7d99db1c89888a1847903b3ec6332f4 > Though fwiw, I would just inline the...

Thanks, @michael2012z! > `info registers` printed the expected long list of registers It's not pretty but at least, it works; I'll try to improve that: https://github.com/daniel5151/gdbstub/issues/12#issuecomment-1217836739

I plan on working on this by implementing https://github.com/daniel5151/gdbstub/pull/109#issuecomment-1198355769 _i.e._ making the `Arch` provide a standard XML ``-ing all its (`Arch`-defined + GDB-standard) features and make the `gdbstub` backend forward...