| (* Doubly linked list node *) | |
| type 'a dlln = { mutable data : 'a ; | |
| mutable prev : 'a dlln ; | |
| mutable next : 'a dlln } | |
| (* Look ma, a cycle can be initialized without any NULL pointers! *) | |
| let rec singleton : int dlln = | |
| { data = 0 ; | |
| prev = singleton ; | |
| next = singleton } |
| module Inc = Incremental_lib.Incremental.Make () | |
| open Inc | |
| let make init = | |
| let num = 4 in | |
| let den_v = Var.create init in | |
| let den = Var.watch den_v in | |
| (* computes "if den = 0 then 0 else num / den" *) | |
| let div = bind den (fun den -> return (num / den)) in |
| Will, Dakota, Allan | |
| 20161026 -- Wed | |
| Q: Why clean "aggressively"? | |
| ----------------------------- | |
| Informal definition: "Aggressive cleaning", which Adapton (PLDI 2014) | |
| implements, consists of several aspects of the design of the DCG, | |
| which work together to help avoid needless re-evaluation: |
| (* The 'Name Term' sublanguage of Typed Adapton. *) | |
| (* See: https://arxiv.org/pdf/1610.00097v2.pdf | |
| Figure 6 (top) | |
| and Figure 8 (top) | |
| *) | |
| Inductive sort : Set := | |
| | Arr : sort -> sort -> sort | |
| | Nm : sort | |
| | Prod : sort -> sort -> sort |
| #[macro_use] | |
| extern crate adapton; | |
| use adapton::macros::*; | |
| use adapton::engine::*; | |
| fn main() { | |
| manage::init_dcg(); |
| error[E0275]: overflow evaluating the requirement `rustc::ty::ExistentialPredicate<'_>: std::marker::Sync` | |
| --> src/librustc_driver/driver.rs:968:13 | |
| | | |
| 968 | thread::spawn(move||profile_queries_thread(rx)); | |
| | ^^^^^^^^^^^^^ | |
| | | |
| = help: consider adding a `#![recursion_limit="128"]` attribute to your crate | |
| = note: required because it appears within the type `[rustc::ty::ExistentialPredicate<'_>]` | |
| = note: required because it appears within the type `rustc::ty::Slice<rustc::ty::ExistentialPredicate<'_>>` | |
| = note: required because it appears within the type `&rustc::ty::Slice<rustc::ty::ExistentialPredicate<'_>>` |
Instead of using call-by-value (CBV), I often use Call-by-push-value (CBPV) in my core calculi, and many people ask me "why?".
Most reasons relate to the "fine-grained" typed structure that it introduces, and the ability of this structure to describe "low-level" concerns, like the difference between partial and full function application (of curried,
| 4dfdd3d3fc HEAD@{0}: reset: moving to HEAD~10 | |
| d105172a10 HEAD@{1}: rebase -i (finish): returning to refs/heads/master | |
| d105172a10 HEAD@{2}: rebase -i (start): checkout refs/remotes/origin/master | |
| d105172a10 HEAD@{3}: rebase: aborting | |
| 580ab1ef6f HEAD@{4}: rebase -i (pick): Update release notes for 1.19.0 | |
| 4545f18562 HEAD@{5}: rebase -i (pick): Allow remote testing remotely when `TEST_DEVICE_ADDR` is set | |
| 814f1ea446 HEAD@{6}: rebase -i (pick): Add empty MIR pass for non-lexical lifetimes | |
| d573b9605a HEAD@{7}: rebase -i (pick): Fix checking for missing stability annotations | |
| 773ace032f HEAD@{8}: rebase -i (pick): explanatory error on `--print target-spec-json` without unstable options | |
| 06f7c0df4c HEAD@{9}: rebase -i (pick): reorder span labels |
| 825166bc46 (HEAD -> master) HEAD@{0}: rebase finished: returning to refs/heads/master | |
| 825166bc46 (HEAD -> master) HEAD@{1}: rebase: -Z profile-queries: compute 'self' duration; order final summary by 'self' duration, not 'total' duration | |
| 763e8ba4b0 HEAD@{2}: rebase: -Z profile-query-and-key, separate from -Z profile-query; query key is string option | |
| 40e26f2776 HEAD@{3}: rebase: -Z profile-queries includes dep_graph.with_task uses in output | |
| 06be62b0bb HEAD@{4}: rebase: count timed passes in totals; minor CSS stuff | |
| adf3845811 HEAD@{5}: rebase: -Z profile-queries: remove panic when channel is unset | |
| 8591d0dc3f HEAD@{6}: rebase: profiling with -Z profile-queries recognizes -Z time-passes | |
| 0148826367 HEAD@{7}: rebase: tidy: trailing ws | |
| f723e016d2 HEAD@{8}: rebase: comments from @michaelwoerister | |
| 9615098a94 HEAD@{9}: rebase: clean up and fixes, thanks to comments from @kennytm; moved dump logic to profile mod |
Archive:
⇨ Fall 2017 Calendar
People