Stable ordering when generating StructMeta code #1
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Currently, when user specifies `derive(StructMeta)`, this crate generates a branch of code for each of the attributes. These attributes however have so far been stored inside a `HashMap` which means the ordering of the branches is more or less random. This has no real impact on the runtime: the order of the branches doesn't change the semantics of the code. It does matter however for compile time: `rustc` will calculate a different crate hash (effectively a sort of ABI signature) if code is re-ordered. This means that any time we recompile code with `derive(StructMeta)` somewhere, we're breaking the ABI even if there were _no_ changes at all to any code. Basically, we're throwing a dice at compile time as to what order the code will be laid out in. This is very bad for reproducability and plays very poorly with any sort of binary caching. I initially thought that the reproducability issue was coming from `rustc` itself but the more I was debugging, I made my way back to this crate. You can see the original rustc issue for context [here](rust-lang/rust#89904). I think this change should be basically invisible to anyone that's been using the crate so far.
|
I have verified that with this patch if I now, in If at all possible, a release to crates.io would be amazing, along with update to |
This was referenced Oct 18, 2021
|
Thanks for the excellent pull request. I have now released structmeta 0.1.4 and parse-display 0.5.3 which contain this fix. |
Thank you for being so quick! We'll switch over to it ASAP, we've been using my branch temporarily. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Currently, when user specifies
derive(StructMeta), this crategenerates a branch of code for each of the attributes.
These attributes however have so far been stored inside a
HashMapwhich means the ordering of the branches is more or less random. This
has no real impact on the runtime: the order of the branches doesn't
change the semantics of the code. It does matter however for compile
time:
rustcwill calculate a different crate hash (effectively a sortof ABI signature) if code is re-ordered.
This means that any time we recompile code with
derive(StructMeta)somewhere, we're breaking the ABI even if there were no changes at all
to any code. Basically, we're throwing a die at compile time as to what
order the code will be laid out in.
This is very bad for reproducability and plays very poorly with any sort
of binary caching.
I initially thought that the reproducability issue was coming from
rustcitself but the more I was debugging, I made my way back to thiscrate.
You can see the original rustc issue for context
here.
I think this change should be basically invisible to anyone that's been
using the crate so far.