Immutable Paths, of minimum length 2 #906
Closed
Conversation
This file contains 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
This changes PATH! to no longer be an ANY-SERIES! type. This means it cannot be operated on with series operations like APPEND, and it can't be traversed to give new positions into paths with things like NEXT or SKIP. However, other operations like LENGTH OF, PICK, and FOR-EACH continue to work. Because there's no way to change them, ANY-PATH! is also immutable. This means that a path can be checked at the time of its creation for properties which cannot be violated by future mutations. The new rules are: * Paths must contain at least 2 elements. The minimum path is `/`, which is the path form of [_ _]...two blanks. * Paths may not directly contain other paths. This prevents various ambiguity problems like [a/b c] and [a b/c] both being possible interpretations of a/b/c. Indirect inclusion of other paths is legal, such as `a/(b/c)/d` It is believed that this change eliminates harmful degrees of freedom, without really sacrificing any behaviors people found that interesting. Additionally, being able to promise paths are immutable open doors to being able to create optimized representations of things like `/` or `/a`, which could be packed into a single cell with no array node. If one wishes to do surgery on path content, the way to do that is to convert it to a BLOCK! or GROUP! and use series operations on that, and then change it back to a PATH!. (Note: Any series embedded in a path, e.g. the GROUP! in `a/(b c d)/e`, is not subject to the immutability rule...and is subject to whatever mutability state it had when it was inserted.)
This commit does a rethinking of the definition of a REFINMENT! such that /word is the parallel to what would be [_ word] as a block. This opens up the possibilities for greater expressions of paths in the pattern `/foo/bar/...` Similar to how the QUOTED! change generalized what were previously dedicated types, this adds REFINEMENT! as a parse rule and some compatibility checks. Unlike LIT-WORD? and LIT-PATH? however, the matching of the pattern for REFINEMENT? and the term may be worth keeping to idiomatically identify this path subclass. As a first take, this is more inefficient than a refinement, as it actually does allocate 2-element arrays for each usage (which can't fit in 2 cell spaces, as it must account for an end as well, so it requires a full 4 cell-sized array allocation, in addition to a series node). This makes a refinement cost roughly the overhead of 6 additional value cells to the one cell it occupied previously. Next step will be to adapt the "in-situ" escaping technique (which was made to put higher quoting levels into cells directly) such that it can represent this form of 2-element path at nearly no added cost, while still appearing to be a PATH!.
hostilefork
force-pushed
the
master
branch
4 times, most recently
from
c8d2c31
to
f9d4985
Compare
|
Merged with fixes: |
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.
This changes PATH! to no longer be an ANY-SERIES! type. This means it
cannot be operated on with series operations like APPEND, and it can't
be traversed to give new positions into paths with things like NEXT
or SKIP. However, other operations like LENGTH OF, PICK, and FOR-EACH
continue to work.
Besides no series modifiers, ANY-PATH! can't be mutated with POKE-ing.
This means that a path can be checked at the time of its creation for
properties which cannot be violated by future mutations.
The new rules are:
Paths must contain at least 2 elements. The minimum path is
/,which is the path form of [_ _]...two blanks.
Paths may not directly contain other paths. This prevents various
ambiguity problems like [a/b c] and [a b/c] both being possible
interpretations of a/b/c. Indirect inclusion of other paths is
legal, such as
a/(b/c)/dIt is believed that this change eliminates harmful degrees of freedom,
without really sacrificing any behaviors people found that interesting.
Additionally, being able to promise paths are immutable open doors to
being able to create optimized representations of things like
/or/a, which could be packed into a single cell with no array node.If one wishes to do surgery on path content, the way to do that is to
convert it to a BLOCK! or GROUP! and use series operations on that,
and then change it back to a PATH!.
(Note: Any series embedded in a path, e.g. the GROUP! in
a/(b c d)/e,is not subject to the immutability rule...and is subject to whatever
mutability state it had when it was inserted.)