Skip to content

Latest commit

 

History

History
133 lines (122 loc) · 3.81 KB

TODO.md

File metadata and controls

133 lines (122 loc) · 3.81 KB
Raw

bugs in Wikifunctions that need to be filed

editing the label of Z10376 seems to take down stuff??? -- probably because the resursive test is very heavy and takes the orchestrator down

the following things throw errors: [ Z10015, five ] [ positive integer, five ] [ positive integer, zero, five ] [ positive integer, add( positive integer<"2">, positive integer<"2"> ) ] this works, though: [ positive integer, positive integer<"5"> ] [ positive integer ] [ string ] [ string, "a" ] [ string, "a", "b" ] [ string, "a", "b", concatenate("c", "d") ] [ string, "a", "b", concatenate("c", "d"), gello string ]

the following throws errors: length of list([Z1, "a", "b"]) length of list([Z6, "a", "b"]) this works though: length of list([string ]) length of list(string to list ("abc")) length of list([ Z1 ]) length of list([ ])

  • the following should have been a "cannot delabel reference 'one'" λ→ Z10058(zero, one) instead I get Error error type: "expected closing" error value: [ undefined K1: "ZSeparator" K2: "12" undefined K1: "ZSymbol" K2: "13" K3: "one" undefined K1: "ZCloseArg" K2: "17" ] 330 ms

minor fixes

λ→ Z13 [Object ] should be [Object]

λ→ typed pair(string, boolean)<"a", true> [object Object]<"a", true> should be Typed pair(String, Boolean)<"a", true>

next steps

  • add tests for literal notation

  • add documentation for literal notation

  • check if generics work, re parsing, both functions and types

  • full underscore implementation

  • have a test dataset in test that is used for tests instead of the real data in function-schemata. Use this to also test several languages

  • some way to show type declaration?

  • tests -- normalize -- canonicalize -- labelize -- answer -- format -- evaluate -- lambda -- interactive -- wellformed -- prettyprint

  • works a little: -- wellformed

previous next steps

-- parse / delabel / labelmap in load

  • how to manage labelmap state in different languages and endpoints
  • use search instead of labelmap for remote endpoints
  • how to deal with caches against different endpoints
  • what to do when more than one hit
  • when switching .wiki, also delete the cache and the labelmap

then:

  • prettyprint used to show the whole Persistentobject, not just the value, when asked for a ZID, should there be something like that?
  • lambda from other directories has issues, check config.json starting with ./
  • replace .reload with .clear
  • .language should also use language codes, not just ZIDs
  • eval, search, and load all make calls to the MW API with lots of repeat code
  • checking levels -- parses (JSON) -- parses mixed JSON into function call syntax and the other way around -- Wellformedness (simple wellformedness) -- Conforming (all inalienable truths) -- Linked (all ZIDs exist) -- checkable (the Type has a validator) -- valid (the validator returns no errrors)
  • deal with errors in many places

furthermore:

  • introduce recoverable errors when the line is not complete
  • allow JSON objects in syntax
  • labelmap from website
  • write a help message for --help
  • move repo to gerrit
  • rename master to main (will happen together with all other gerrit repos)
  • allow --config parameter
  • autocomplete (for default parser) -- needs labelmap or something to get all labels from .data source
  • result preview (for default parser)
  • check abstracttext capabilities and learn
  • defineCommand parser
  • defineCommand writer
  • defineCommand reload labelmap
  • defineCommand set evaluator
  • what about i18n of the commands?
  • check if FUTURE_README and README align
  • have several commands in one line, the last one counts, use of _ might make that interesting and readable
  • tests -- tests for aliases -- tests for delabeling other languages but english -- more tests for tokenizing and parsing
  • dont show timer on command line usage
  • extra empty line on normal evaluation