rhi/quox
rhi/quox
1
0
Fork 0
Code Issues 15 Pull requests 1 Releases Wiki Activity

WIP: quantity polymorphism #44

Draft
rhi wants to merge 9 commits from qpoly into 🐉
pull from: qpoly
merge into: rhi:🐉
Conversation 0 Commits 9 Files changed 34 +2314 -1369
rhi commented 2024-05-06 15:07:14 -04:00
Owner
Copy link

new syntax

the existing π.(x : A) → … etc syntax only really works when π is a single token. and even then it took me a while to get used to it.

function types

  • (x : A | π) → B
    • parse
    • print
  • (A | π) → B for nondependent. parens needed to disambiguate nested function types
    • parse
    • print
  • todo less annoying short syntax for 0/1/ω
    • come up with
    • parse
    • print

boxes

  • [A | π] for consistency with new function syntax
    • parse
    • print

case/let

  • case‹π› or case.{π}, same for let; same as qty app
    • parse
    • print
  • case0 etc still exist

qty polymorphic definitions:

  • def app‹π, ρ› : type = rhs
    • parse
    • print
  • def app.{pi, rh} for non-unicode version
    • parse
    • print
  • todo is bounded quantification needed? probably :(
  • f‹π, ρ› or f.{pi, rh} for application
    • parse
    • print
## new syntax the existing `π.(x : A) → …` etc syntax only really works when π is a single token. and even then it took me a while to get used to it. ### function types - `(x : A | π) → B` - [ ] parse - [ ] print - `(A | π) → B` for nondependent. parens needed to disambiguate nested function types - [ ] parse - [ ] print - **todo** less annoying short syntax for 0/1/ω - [ ] come up with - [ ] parse - [ ] print ### boxes - `[A | π]` for consistency with new function syntax - [ ] parse - [ ] print ### case/let - `case‹π›` or `case.{π}`, same for `let`; same as qty app - [ ] parse - [ ] print - `case0` etc still exist ### qty polymorphic definitions: - `def app‹π, ρ› : type = rhs` - [ ] parse - [ ] print - `def app.{pi, rh}` for non-unicode version - [ ] parse - [ ] print - **todo** is bounded quantification needed? probably :( - `f‹π, ρ›` or `f.{pi, rh}` for application - [ ] parse - [ ] print
rhi force-pushed qpoly from 46889fefc8 to c5f2b2a2ce 2024-05-23 16:42:47 -04:00 Compare
rhi force-pushed qpoly from c5f2b2a2ce to ef273e819e 2024-05-27 15:28:43 -04:00 Compare
rhi force-pushed qpoly from ef273e819e to 4c008577b4 2024-05-27 15:32:54 -04:00 Compare
This pull request has changes conflicting with the target branch.
  • lib/Quox/OPE.idr
  • lib/Quox/Syntax/Term/Subst.idr
  • lib/Quox/Syntax/Term/Tighten.idr
  • lib/Quox/Whnf/Coercion.idr
  • lib/Quox/Whnf/Main.idr
  • lib/Quox/Whnf/TypeCase.idr
View command line instructions

Checkout

From your project repository, check out a new branch and test the changes.
git fetch -u origin qpoly:qpoly
git checkout qpoly

Merge

Merge the changes and update on Forgejo.
git checkout 🐉
git merge --no-ff qpoly
git checkout 🐉
git merge --ff-only qpoly
git checkout qpoly
git rebase 🐉
git checkout 🐉
git merge --no-ff qpoly
git checkout 🐉
git merge --squash qpoly
git checkout 🐉
git merge --ff-only qpoly
git checkout 🐉
git merge qpoly
git push origin 🐉
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels   
bug

something is arriving at the wrong answer

  
duplicate

this issue or pull request already exists

  
help wanted

   
info needed

   
invalid

   
language

new language feature

  
performance

source of peformance problems but not correctness

  
refactor

   
ux

ucks

  
wontfix
No labels
bug
duplicate
help wanted
info needed
invalid
language
performance
refactor
ux
wontfix
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference: rhi/quox#44
No description provided.
Delete branch "qpoly"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?