func‿mod1‿mod2 ← •args lf←@+10 charSet‿cgl←(∾ ≍○< ≠¨)⟨ func # Function mod1 # 1-modifier mod2 # 2-modifier "⋄,"∾lf # Separator "⇐←↩" # Gets "(){}⟨⟩" # Bracket "‿" # Ligature "·" # nOthing # Use last character in case of UTF-16 (like dzaima/BQN) ¯1⊏˘10‿∘⥊"𝕊𝕏𝕎𝔽𝔾𝕤𝕩𝕨𝕗𝕘" # Input (𝕣 pending; ℝ not allowed) ".¯π∞" # Numeric '0'+↕10 # Digit "_"∾˜⥊"aA"+⌜↕na←26 # Alphabetic "•"∾(¯1↓"𝕨")∾" "∾@+9 # Whitespace (or special name prefix in UTF-16) "#'""@" # Preprocessed characters ⟩ bF‿b1‿b2‿bS‿bG‿bB‿bL‿bO‿bX‿bN‿bD‿bA‿bW‿bP←≍¨˜⟜(0»+`)cgl M←1⊸⊑(0⊸≤∧>)-⟜⊑ # ∊ for an init,length pair 𝕩 as above sep←⊑bS bI←5≍⟜-⊸+bX Pl←∾⟜("s"/˜1<≠) # Pluralize _tmpl←{∾𝕗{𝕎𝕩}¨<𝕩} # Template # Convert characters to numbers, mostly the same as tokens CharCode←charSet{ ErrUnknownChars←0!˜⟨"Unknown character"⊸Pl,": ",⊢⟩_tmpl Chk ← ⊢⊣ErrUnknownChars∘(≠/⊣)⍟≢⟜(⊏⟜𝕗) g←⍋𝕗 ⋄ ⊢ Chk g⊏˜1-˜1⌈(g⊏𝕗)⍋⊢ } vd←1+vi←⊑bN # Start of identifier numbering (plus dot) charRole←4∾˜∾⥊¨˜⟜(≠↑cgl˙)⟨1,2,3,¯1,¯3,¯1‿0,¯2,0,5/⌽↕2⟩ # For first vd chars T←⌈`× ⋄ IT←↕∘≠⊸T ⋄ I1T←(1+↕∘≠)⊸T # Source to ⟨tokens, roles, number of identifiers, literals⟩ # Identifiers then literal tokens are numbered starting at vi Tokenize←{ # Resolve comments and strings c←𝕩='#'⋄s←/0‿0⊸«⊸∧sm←𝕩='''⋄d←/dm←𝕩='"' g←⍋q←∾⟨ s⋄¯1↓d⋄/c⟩ ⋄q↩g⊏q # Open indices e← g⊏∾⟨2+s⋄ 1↓d⋄-⟜»∘⊏⟜(0∾+`c)⊸//(𝕩=lf)∾1⟩ # Matching close indices Se←{(⊏˜𝕨)Se 1¨⌾((𝕩/𝕨)⊸⊏)𝕩}⍟(0=¯1⊑⊢) # Mark reachable openings St←(≠𝕩)↑·/⁼((≠↑∾⟜≠Se 1∾0¨)q⍋e)⊸/ # All indices → reached mask a←St q⋄b←St e⋄f←1≠`ab←a∨b # Open/close masks; filter {⟨⊑/𝕩,"Unclosed quote"⟩!0}⍟(∨´)(sm∨dm)∧b↩qe←dm∧«a∧↩dm # Quote Escape "" str←𝕩⊔˜1-˜(si←a>»qe)(⊣+`⊸×○(∾⟜1)<)≠`dm∧ab # Strings (indices /si) # Extract words: identifiers and numbers ie←/f⋄is←ie≠⊸↑/1»f⋄Fs←{is/˜↩𝕨⋄𝕨/𝕩} # Token start and end is-↩is(-×⊏⟜c)ie # Comment → ending newline only t←CharCode ie⊏𝕩 nd←(t=⊑bN)>«t M bD # Namespace dot w←»⊸us wk←na⌊∘÷˜(⊑bA)-˜w/t # Kind of word from first char t-↩na×l∧t≥na+⊑bA # Case-insensitive {⟨𝕩/is,"System dot with no name"⟩!0}⍟(∨´)sy>«l w≠↩»⊸∨sy # Start system word at dot wi←0n←l∧(+`w)⊏0∾¬wi # Identifier/Number masks num←is ReadNums○(((0∾us)<∨⟜«0∾n)/0⊸∾) t×l # Numbers ir←(us/˜«⊸us)×+`w>n # Identifier groups {⟨is⊏˜𝕩/𝕨,"Numbers can't start with underscores"⟩!0}⍟(∨´⊢)⟜(ws<(⊑bA)>⊏⟜t)/(¯1»⌈`)⊸¯1»⌈`)⊸/¨˜↩j←⊐¨k # IDs j into uniques k k↩𝕎⌾(1⊸⊑)k # System value lookup wf←¬l∨t M bW⋄ie/˜↩wf∨>⟜«l # Index management for... t↩(w∨wf)Fs(∾j++`vd»kk←≠¨k)⌾(ki⊸⊏)t # Add IDs; remove words/whitespace t-↩t(M×-⟜⊑)bS # Separators are equivalent p←≠`1¨sb←¯1↓1↓/1(∾≠∾˜)t=sep # Separator group boundaries (excludes leading and trailing) sk←sb/˜p>∨⟜«(p+(sb-p)⊏t)∊3‿5+⊑bB # Keep the first of each group that's not just inside a bracket t{ie/˜↩𝕨⋄𝕨Fs𝕩}˜↩1¨⌾(sk⊸⊏)t≠sep # Remove the rest r←ir⌾((t M vd≍+´2↑kk)⊸/)(vd⌊t)⊏charRole∾0 # Role t+↩5×t M⟨⊑bX,5⟩ # Case-insensitive special names t-↩vi(<+10×=)t # Shift . to bX and variables back one ⟨t,r,k,is,ie⟩ } # 𝕩 is a list of tokens that contains the numeric literals, each # preceded by 0. Return the numbers. ReadNums←{ _err_←{(0!˜/⟜𝔾≍○<𝔽)⍟(∨´)} EChars←⟨"Letter"⊸Pl," """,⊏⟜charSet,""" not allowed in numbers"⟩_tmpl e‿d‿n‿p‿i←=⟜𝕩¨((⊑bA)+-´"ea")∾+⟜↕´bN # Masks for e.¯π∞ EChars∘(/⟜𝕩)_err_𝕨 ¬e∨𝕩<⊑bA s←d∨c←e∨z←0=𝕩⋄m←¬n∨c "Negative sign in the middle of a number"_err_𝕨 n>»c "Portion of a number is empty"_err_𝕨 (1«s)∧n∨s "Ill-formed decimal or exponent use"_err_(s/𝕎) ¬(0⊸=∨»⊸<)s/𝕩 "π and ∞ must occur alone"_err_𝕨 (p∨i)>1(»∧(p∧«e)∨«)z∨n>»e f←(17≥¬(⊢-T)+`)⊸∧g←(«≤(d<𝕩≠⊑bD)>○I1T¬)⊸∧m # No leading 0s; max 17 digits l←(¯1∾π‿1∾↕10)⊏˜(¬d)/f×𝕩-1+⊑bN # Digit lookup, with ∞ as 1 to avoid ∞×0 v←(>⟜«0≤l)/0(0⊸≤××⟜10⊸+)`l # Numeric values—mantissas and exponents v×↩1‿¯1⊏˜(r←>⟜»m)/»n # Negate if ¯ vm←c/z # Mask of mantissas in l dp←vm/f(--»⊸-(<×⊢)⊏⟜(I1T«d)⊸-)○(/>⟜«)g # Decimal position t←10⋆|ee←dp-˜vm/«v׬vm # Power of 10 t÷˜⌾((0>ee)⊸/)t×⌾((0○≠◶⟨"opening"_mis_ 0, "closing"_mis_ 1˜⟩ 0 !˜ 𝕨 (Lcs<⌊○≠)◶⟨Msg,"Malformed bracket nesting"⟩ 𝕩 } _err_←{(0!˜(∧∘⍉(i≍e˙)⊏⎉1˜/⟜𝔾)≍○<𝔽)⍟(∨´⍟=)} # Bracket and ligature validation and handling # Open brackets have role ¯1 and closed ones have role 0 "Empty program" ! 0<≠𝕩 g←⍋pd←+`p←(¯1-2×r)×𝕩 M bB⋄gb←g⊏r=¯1 # Paren (actually any bracket type) depth and grade (g⊏p)(>⟜0⊸/ErrMismatchedBrackets⍟≢1-˜<⟜0⊸/)gx←g⊏𝕩 "Swapped open and closed brackets" ! 0≤(⊑g)⊑pd "Parentheses can't contain separators"_err_(gb/G) ((⊑bB)⊸=»⊸∧sep⊸=)gb/gx {"Empty statement or expression"_err_(𝕩/0∾G) (4+⊑bB)≠𝕩/0∾gx}1⊸«⊸∧1∾gb r-↩(𝕩=⊑bG)>ec←«0≤r+p # Role ¯4 for exports: ⊑bG is ⇐ "Invalid assignment or stranding use"_err_(↕∘≠) ((¯4⊸<∧≤⟜¯2)r)>ec∧»0≤r "Dot must be followed by a name"_err_(↕∘≠) (r=4)>«𝕩 M vi‿nv dl←«⊸∨dc←r=4 # Dot left sr←»⌾(((⍋⊏⟜dl)⊸⊏g)⊸⊏)sl←«⊸∨r=¯2⋄ns←¬sl∨sr # Strand right and left; not stranded cp←𝕩=1+⊑bB # Closed paren nr←(IT¬cp)⊏𝕩=⊑bO # Nothingness role: 1 for · (handle 𝕨 later?) "Can't strand Nothing (·)"_err_(↕∘≠) ns0⋄c←/br<0 # Block Begin (mask) and Close (index), in matching order # Block properties PN←1(∾/∾˜)(∨/⊣)⋄H←b¬∘PN= # Which blocks have a property sn←(0‿3‿4‿5+⊑bI)⍋𝕩⋄sp←/𝕩 M bI # Special name index fx←1 H sn⋄fr←(fx∨0⊸<)⊸+ft←(2⊸H⌈2×3⊸H)sn # Block immediacy ¬fx and type ft "Special name outside of any block"_err_(/{(0=fi)∧𝕩 M bI}∘𝕩) 0<⊑fr fsc←(ft⊏0‿2‿3)+3×fx # Special name count # Propagate roles through parentheses # ir is the role of the expression ending at each position (truncated to the right) r↩sl-˜ns×(1↓fr)⌾((c⊏rev)⊸⊏)r # Add block roles; make strand elements ¯1 pt←cp∧ns # Pass-through parentheses: not in strands pp←pt∧»es←1⊸»⌾(g⊸⊏)r<0 # Parens enclosing one object (maybe with assignment) don't change roles ir←((»⌾(g⊸⊏)(1+es)×3=⊢)⌈⊢-es<2≤⊢)r+pp×(IT¬pp)⊏r # Propagate modifier roles ir⌈↩(IT¬pt∧ir=0)((⊏-⊢)⟜(+`¬pp)(⊢⌊1⌈+)⊏)ir # ...and function roles r+↩pt×»ir # Roles at pt were 0; set them now ir(×⟜¬-⊢)↩nr # Include nothingness (doesn't handle 𝕎) r-↩(r=¯4)∧1»r=¯1 # Lone ⇐ to role ¯5 "Dot must be preceded by a subject"_err_(⍋∘rev) dc>»r=0 r(×⟜¬-⊢)↩dl⋄dc⊏˜↩rev # Namespace and dot to ¯1 # Prep for lexical resolution before reordering 𝕩 xv←𝕩-vi # Parsing part 1 p-↩br⋄g↩⍋+`p⋄g⊏˜↩⍋g⊏«⊸∨dc⋄gr←g⊏rev # Order by non-brace bracket depth, then dots sll←1+2÷˜0(<-○/>)gr⊏sr-sl⋄l←/g⊏𝕩=5+⊑bB # Strand length; list starts bp←⟨/b,c⟩∾¨0(<≍○()g⊏p # Bracket pairs {i↩(𝕨⊏i)⌾(𝕩⊸⊏)i⋄e↩(𝕩⊏e)⌾(𝕨⊸⊏)e}´bp # Highlight all contents of a pair when error reporting g⊏˜↩gs←⍋gr⊏sl⋄gr↩g⊏rev⋄gi←⍋g # Send strand prefixes *‿ to the end 𝕩⊏˜↩g⋄r⊏˜↩gr⋄ir⊏˜↩gr⋄l↩(l⊏⍋gs)∾/gr⊏sr>sl a←(¯5⊸<∧≤⟜¯3)r⋄ps←a(«∧m2≤»)noaa af>↩alm←ai⊏aa⋄al←alm/ai # aliases al # Lexical resolution (independent of parsing part 2 below) di←/dm←»dc # Dots aren't scoped id←/idm←(gi⊏«aa∧a))0∨|)ps(⊢-T)+`¬ro # Train argument (first-level) fa←/(fe←ta∨ro∨«⊸∨psfe∨ff dy←fa⊏«no<(tr∧r≥0)∨roblk, oi, tx⟩ }