Kyle's comments on Niko's draft #15
Conversation
| explicit intention to permit Rust panics to propogate through native frames | ||
| * Announce "project group" with the purpose of designing subsequent RFCs to | ||
| flesh out the details of the "C unwind" ABI; invite interested Rustaceans to | ||
| join |
There was a problem hiding this comment.
These bullet points were duplicated below, so I made a dramatically abbreviated version for the summary section.
| details some of those use cases and their specific requirements. | ||
|
|
||
| ### Lucet: unwind past empty native frames | ||
| ### Lucet: unwind past foreign frames |
There was a problem hiding this comment.
Discussion on changing "native" to "foreign" here: https://rust-lang.zulipchat.com/#narrow/stream/210922-wg-ffi-unwind/topic/Native.20frame.20definition/near/178713063
| Behavior] -- however, we would like to support longjmp for Rust frames that do | ||
| not contain destructors. At minimum, this group should take the behavior of | ||
| longjmp into account; if it is easy to do, we may also introduce mechanisms or | ||
| RFCs that help to specify the use of longjmp in particular cases. |
There was a problem hiding this comment.
This is mostly a change in where I put my newlines. Sorry, I typically use 80-column lines and aggressively use Vim's gq to realign! I can reformat to minimize the diff if you'd like.
| when unwinding over C code compiled with Clang using the `-fexceptions` | ||
| compiler option. | ||
|
|
||
| > XXX I don't believe we've ever seen improper destructor behavior when |
There was a problem hiding this comment.
Throughout, I've used > XXX to indicate a comment from me. You had also used XXX to mark something like that, so I changed it to > Niko XXX.
BatmanAoD
force-pushed
the
kjs-comments-niko-rfc0
branch
from
e0e480b
to
a646a94
Compare
|
Superseded by #28 |
This is a work-in-progress, but I wanted to publish it sooner rather than later in case Niko wants to take a look.