More progress towards Alpha cleanup #849
Conversation
arthurpaulino
force-pushed
the
more-migrations
branch
2 times, most recently
from
9253717
to
6f96d94
Compare
arthurpaulino
force-pushed
the
more-migrations
branch
2 times, most recently
from
f937580
to
1906714
Compare
|
You may want to make a mention of #494 |
|
It's already set to close #494 |
arthurpaulino
force-pushed
the
more-migrations
branch
4 times, most recently
from
2a91c45
to
7277446
Compare
|
@huitseeker there's no API for |
|
@arthurpaulino my bad you're right. It's in #881 if you want it. |
arthurpaulino
force-pushed
the
more-migrations
branch
2 times, most recently
from
80d75ef
to
e5d6615
Compare
arthurpaulino
force-pushed
the
more-migrations
branch
from
e5d6615
to
53d7635
Compare
arthurpaulino
force-pushed
the
more-migrations
branch
from
53d7635
to
eb2b663
Compare
* Get rid of unecessary parsing tests * Finish up the migration of gadgets specific to Lurk data construction and deconstruction
arthurpaulino
force-pushed
the
more-migrations
branch
from
eb2b663
to
f10ab75
Compare
| self.0 | ||
| .get(&FWrap(f)) | ||
| .ok_or_else(|| anyhow!("Global allocation not found for {}", f.hex_digits())) | ||
| .ok_or_else(|| SynthesisError::AssignmentMissing) |
There was a problem hiding this comment.
We may want to consider creating a real error to identify the case in which failure is due to 'lack of knowledge' — since this is likely to be an entirely real use case in eventual Lurk applications. Since 'knowledge is power', trying to evaluate (or prove) expressions to which (it turns out) one doesn't have the ability to dereference a pointer is conceptually equivalent to a kind of 'access denied'. It's true that this should lead to failure (i.e. proofs or evaluations cannot be created), but there are likely cases where this should be handled specially.
For example, it may be that when an agent tries to access privileged information, the user is prompted to explicitly grant access before continuing. (This is just a suggestive example, and one can imagine many more cases in which this kind of non-deterministic ignorance is either ephemeral or should be handled in a specific, non-fatal way.)
* fix: make z_data::serde private * refactor: restrict visibility of CLI backend/config * make circuit::data pub(crate) * refactor: make gadgets from #849 explicitly unused * make interpreter pub(crate) * refactor: make top-level expr, cont, hash, package private * refactor: make ptr pub(crate) * refactor: make symbol crate-private * refactor: make syntax crate-private * refactor: make tag crate-private * refactor: make uint crate-private * refactor: make writer crate-private * refactor: limit visibility in public_params * refactor: Refactor variant naming in public parameters module - Renamed `Error::IOError` to `Error::IO` and `Error::CacheError` to `Error::Cache` across multiple files for consistency. * fix: repair doctest
src/parser/syntax.rsalready contains the cases we want to coverCloses #494