Split the lighttpd.rs in #727 into a currently working and full version
#728
Assignees
Comments
kkysen
added a commit
that referenced
this issue
Nov 19, 2022
Provide an initial stubbed `lighttpd` connection path. This will continue to be split out into multiple versions as described in #728. It'll also be moved to a separate crate (not the existing `analysis/test`), tested for no UB with `miri`, and a script added to run `c2rust-analyze` on it/them).
|
Could you help me understand why the third option is useful? Are you envisioning a workflow where we take the most recent currently-working file and add features from the version containing all the code to play with locally, without accidentally committing changes to the former? |
|
We can see as we go, but I imagined a 3rd version that could be anything in-between could be useful. If ends up not being, we can remove it. I do think we need it at least at the very beginning, though, as I want to split out the min version from it, and split out the max version from it, but without changing itself at all. So that'll leave us with 3 versions, and if we realize the 3rd middle one isn't useful at all, we can get rid of it. |
This was referenced Nov 20, 2022
aneksteind
added a commit
that referenced
this issue
Nov 23, 2022
…ts (#729) This moves `analysis/test` to `analysis/tests/misc` to prepare for more tests. This is in preparation for #728, where I'll add separate `lighttpd` tests. In order to stay as close as possible to the original transpiled `lighttpd` code, we'll need to put duplicate `#[no_mangle]` functions in separate crates, which will be created as directories in `analysis/tests`. I think putting them in separate crates is a good idea regardless of the `#[no_mangle]` issue, as when we later want to test everything with dynamic analysis (`c2rust-instrument`), too, it operates on crates, and eventually, I think we'd like for `c2rust-analyze` to do so, too.
aneksteind
added a commit
that referenced
this issue
Nov 23, 2022
…tests/lighttpd` (#730) This moves `analysis/tests/misc/lighttpd.rs` into its own crate: `analysis/tests/lighttpd`. This is in preparation for #728. This PR simply creates the `lighttpd` test crate and moves the existing `lighttpd.rs` file there as is. It also duplicates its `main.rs` from `analysis/tests/misc/main.rs` as much as possible. Any other changes I might add I'll save for another PR, as I want the diff in this one to simply be a move of `lighttpd.rs`. We can also now easily check that the `lighttpd` test crate is UB-free according to `cargo miri run` (it is).
This was referenced Nov 28, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Split the
lighttpd.rsin #727 into a currently working and full version.That is, we should have 3 versions:
lighttpdinto a currently-workinglighttpd-minimal#752: the subset oflighttpd.rsthat currently works withc2rust-analyze. This will be updated to include parts that start work as fixes are made toc2rust-analyze. We should also runc2rust-analyzeon this in CI (Testc2rust-analyzeonlighttpd-minimal#754).lighttpdinto a fulllighttpd-maximal#753: the maximallighttpd.rsthat is taken directly fromlighttpdand contains all of code (for the function(s) we're looking at for now), nothing stripped down and pared yet.lighttpd.rs, which we can use for experimentation.Remaining parts:
lighttpdinto a currently-workinglighttpd-minimal#752lighttpdinto a fulllighttpd-maximal#753c2rust-analyzeonlighttpd-minimal#754The text was updated successfully, but these errors were encountered: