parser,checker: use keywords matcher trie for imported symbol presence checks, instead of x in p.imported_symbol (#25201)

This commit is contained in:
Delyan Angelov 2025-08-30 22:56:30 +03:00 committed by GitHub
parent 2ac3478296
commit cb2756e39c
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
9 changed files with 41 additions and 16 deletions

View file

@ -1022,6 +1022,7 @@ pub mut:
implied_imports []string // imports that the user's code uses but omitted to import explicitly, used by `vfmt` implied_imports []string // imports that the user's code uses but omitted to import explicitly, used by `vfmt`
embedded_files []EmbeddedFile // list of files to embed in the binary embedded_files []EmbeddedFile // list of files to embed in the binary
imported_symbols map[string]string // used for `import {symbol}`, it maps symbol => module.symbol imported_symbols map[string]string // used for `import {symbol}`, it maps symbol => module.symbol
imported_symbols_trie token.KeywordsMatcherTrie // constructed from imported_symbols, to accelerate presense checks
imported_symbols_used map[string]bool imported_symbols_used map[string]bool
errors []errors.Error // all the checker errors in the file errors []errors.Error // all the checker errors in the file
warnings []errors.Warning // all the checker warnings in the file warnings []errors.Warning // all the checker warnings in the file

View file

@ -4261,7 +4261,7 @@ fn (mut c Checker) ident(mut node ast.Ident) ast.Type {
} }
mut name := node.name mut name := node.name
// check for imported symbol // check for imported symbol
if name in c.file.imported_symbols { if c.file.imported_symbols_trie.matches(name) {
name = c.file.imported_symbols[name] name = c.file.imported_symbols[name]
} }
// prepend mod to look for fn call or const // prepend mod to look for fn call or const

View file

@ -61,7 +61,7 @@ fn (mut p Parser) enum_decl() ast.EnumDecl {
end_pos) end_pos)
return ast.EnumDecl{} return ast.EnumDecl{}
} }
if enum_name in p.imported_symbols { if p.is_imported_symbol(enum_name) {
p.error_with_pos('cannot register enum `${enum_name}`, this type was already imported', p.error_with_pos('cannot register enum `${enum_name}`, this type was already imported',
end_pos) end_pos)
return ast.EnumDecl{} return ast.EnumDecl{}

View file

@ -89,7 +89,7 @@ fn (mut p Parser) call_expr(language ast.Language, mod string) ast.CallExpr {
} }
or_kind = if is_not { .propagate_result } else { .propagate_option } or_kind = if is_not { .propagate_result } else { .propagate_option }
} }
if fn_name in p.imported_symbols { if p.is_imported_symbol(fn_name) {
check := !p.imported_symbols_used[fn_name] check := !p.imported_symbols_used[fn_name]
fn_name = p.imported_symbols[fn_name] fn_name = p.imported_symbols[fn_name]
if check { if check {
@ -152,7 +152,7 @@ fn (mut p Parser) call_args() []ast.CallArg {
} else { } else {
expr = p.expr(0) expr = p.expr(0)
if mut expr is ast.Ident { if mut expr is ast.Ident {
if expr.name in p.imported_symbols && !p.imported_symbols_used[expr.name] { if p.is_imported_symbol(expr.name) && !p.imported_symbols_used[expr.name] {
// func call arg is another function call // func call arg is another function call
// import term { bright_cyan, colorize } ... colorize(bright_cyan, 'hello') // import term { bright_cyan, colorize } ... colorize(bright_cyan, 'hello')
p.register_used_import_for_symbol_name(p.imported_symbols[expr.name]) p.register_used_import_for_symbol_name(p.imported_symbols[expr.name])
@ -393,7 +393,7 @@ fn (mut p Parser) fn_decl() ast.FnDecl {
} }
} }
if !p.pref.is_fmt { if !p.pref.is_fmt {
if name in p.imported_symbols { if p.is_imported_symbol(name) {
p.error_with_pos('cannot redefine imported function `${name}`', name_pos) p.error_with_pos('cannot redefine imported function `${name}`', name_pos)
return ast.FnDecl{ return ast.FnDecl{
scope: unsafe { nil } scope: unsafe { nil }

View file

@ -36,7 +36,7 @@ fn (mut p Parser) register_used_import(alias string) {
fn (mut p Parser) register_used_import_for_symbol_name(sym_name string) { fn (mut p Parser) register_used_import_for_symbol_name(sym_name string) {
short_import_name := sym_name.all_before_last('.').all_after_last('.') short_import_name := sym_name.all_before_last('.').all_after_last('.')
short_symbol_name := sym_name.all_after_last('.') short_symbol_name := sym_name.all_after_last('.')
if short_symbol_name in p.imported_symbols { if p.is_imported_symbol(short_symbol_name) {
p.imported_symbols_used[short_symbol_name] = true p.imported_symbols_used[short_symbol_name] = true
} }
for alias, mod in p.imports { for alias, mod in p.imports {
@ -322,12 +322,13 @@ fn (mut p Parser) import_syms(mut parent ast.Import) {
for p.tok.kind == .name { for p.tok.kind == .name {
pos := p.tok.pos() pos := p.tok.pos()
alias := p.check_name() alias := p.check_name()
if alias in p.imported_symbols { if p.is_imported_symbol(alias) {
p.error_with_pos('cannot register symbol `${alias}`, it was already imported', p.error_with_pos('cannot register symbol `${alias}`, it was already imported',
pos) pos)
return return
} }
p.imported_symbols[alias] = parent.mod + '.' + alias p.imported_symbols[alias] = parent.mod + '.' + alias
p.rebuild_imported_symbols_matcher(alias)
// so we can work with this in fmt+checker // so we can work with this in fmt+checker
parent.syms << ast.ImportSymbol{ parent.syms << ast.ImportSymbol{
pos: pos pos: pos
@ -347,3 +348,12 @@ fn (mut p Parser) import_syms(mut parent ast.Import) {
} }
p.next() p.next()
} }
fn (mut p Parser) rebuild_imported_symbols_matcher(name string) {
p.imported_symbols_trie = token.new_keywords_matcher_from_array_trie(p.imported_symbols.keys())
}
@[inline]
fn (mut p Parser) is_imported_symbol(name string) bool {
return p.imported_symbols_trie.matches(name)
}

View file

@ -636,7 +636,7 @@ fn (mut p Parser) parse_any_type(language ast.Language, is_ptr bool, check_dot b
} else if p.expr_mod != '' && !p.inside_generic_params { } else if p.expr_mod != '' && !p.inside_generic_params {
// p.expr_mod is from the struct and not from the generic parameter // p.expr_mod is from the struct and not from the generic parameter
name = p.expr_mod + '.' + name name = p.expr_mod + '.' + name
} else if name in p.imported_symbols { } else if p.is_imported_symbol(name) {
check := !p.imported_symbols_used[name] check := !p.imported_symbols_used[name]
name = p.imported_symbols[name] name = p.imported_symbols[name]
if check { if check {

View file

@ -86,6 +86,7 @@ mut:
implied_imports []string // imports that the user's code uses but omitted to import explicitly, used by `vfmt` implied_imports []string // imports that the user's code uses but omitted to import explicitly, used by `vfmt`
imported_symbols map[string]string imported_symbols map[string]string
imported_symbols_used map[string]bool imported_symbols_used map[string]bool
imported_symbols_trie token.KeywordsMatcherTrie
is_amp bool // for generating the right code for `&Foo{}` is_amp bool // for generating the right code for `&Foo{}`
returns bool returns bool
is_stmt_ident bool // true while the beginning of a statement is an ident/selector is_stmt_ident bool // true while the beginning of a statement is an ident/selector
@ -343,6 +344,7 @@ pub fn (mut p Parser) parse() &ast.File {
mod: module_decl mod: module_decl
imports: p.ast_imports imports: p.ast_imports
imported_symbols: p.imported_symbols imported_symbols: p.imported_symbols
imported_symbols_trie: token.new_keywords_matcher_from_array_trie(p.imported_symbols.keys())
imported_symbols_used: p.imported_symbols_used imported_symbols_used: p.imported_symbols_used
auto_imports: p.auto_imports auto_imports: p.auto_imports
used_imports: p.used_imports used_imports: p.used_imports
@ -592,8 +594,7 @@ fn (mut p Parser) check_name() string {
name := p.tok.lit name := p.tok.lit
if p.tok.kind != .name && p.peek_tok.kind == .dot && name in p.imports { if p.tok.kind != .name && p.peek_tok.kind == .dot && name in p.imports {
p.register_used_import(name) p.register_used_import(name)
} else if p.tok.kind == .name && p.peek_tok.kind == .dot && name in p.imported_symbols } else if p.tok.kind == .name && p.is_imported_symbol(name) && !p.imported_symbols_used[name] {
&& !p.imported_symbols_used[name] {
// symbols like Enum.field_name // symbols like Enum.field_name
p.register_used_import_for_symbol_name(p.imported_symbols[name]) p.register_used_import_for_symbol_name(p.imported_symbols[name])
} }
@ -2698,7 +2699,7 @@ fn (mut p Parser) type_decl() ast.TypeDecl {
return ast.FnTypeDecl{} return ast.FnTypeDecl{}
} }
} }
if name in p.imported_symbols { if p.is_imported_symbol(name) {
p.error_with_pos('cannot register alias `${name}`, this type was already imported', p.error_with_pos('cannot register alias `${name}`, this type was already imported',
end_pos) end_pos)
return ast.AliasTypeDecl{} return ast.AliasTypeDecl{}

View file

@ -73,7 +73,7 @@ fn (mut p Parser) struct_decl(is_anon bool) ast.StructDecl {
p.error_with_pos('struct names must have more than one character', name_pos) p.error_with_pos('struct names must have more than one character', name_pos)
return ast.StructDecl{} return ast.StructDecl{}
} }
if name in p.imported_symbols { if p.is_imported_symbol(name) {
p.error_with_pos('cannot register struct `${name}`, this type was already imported', p.error_with_pos('cannot register struct `${name}`, this type was already imported',
name_pos) name_pos)
return ast.StructDecl{} return ast.StructDecl{}
@ -651,7 +651,7 @@ fn (mut p Parser) interface_decl() ast.InterfaceDecl {
mut pre_comments := p.eat_comments() mut pre_comments := p.eat_comments()
p.check(.lcbr) p.check(.lcbr)
pre_comments << p.eat_comments() pre_comments << p.eat_comments()
if modless_name in p.imported_symbols { if p.is_imported_symbol(modless_name) {
p.error_with_pos('cannot register interface `${interface_name}`, this type was already imported', p.error_with_pos('cannot register interface `${interface_name}`, this type was already imported',
name_pos) name_pos)
return ast.InterfaceDecl{} return ast.InterfaceDecl{}

View file

@ -11,6 +11,11 @@ pub mut:
max_len int max_len int
} }
// str returns a short representation of matcher
pub fn (km &KeywordsMatcherTrie) str() string {
return 'KeywordsMatcherTrie{ /* nodes.len: ${km.nodes.len} */ min_len: ${km.min_len}, max_len: ${km.max_len} }'
}
// TrieNode is a single node from a trie, used by KeywordsMatcherTrie // TrieNode is a single node from a trie, used by KeywordsMatcherTrie
pub struct TrieNode { pub struct TrieNode {
pub mut: pub mut:
@ -18,6 +23,14 @@ pub mut:
value int = -1 // when != -1, it is a leaf node representing a match value int = -1 // when != -1, it is a leaf node representing a match
} }
// str returns a string representation of the node content
pub fn (node &TrieNode) str() string {
if isnil(node) {
return '&TrieNode(nil)'
}
return '&TrieNode{value: ${node.value}}'
}
// find tries to find the given `word` in the set of all previously added words // find tries to find the given `word` in the set of all previously added words
// to the KeywordsMatcherTrie instance. It returns -1 if the word was NOT found // to the KeywordsMatcherTrie instance. It returns -1 if the word was NOT found
// there at all. If the word was found, find will return the `value` (value => 0), // there at all. If the word was found, find will return the `value` (value => 0),