From ab885f07b844fd60adb9d49ed7884f3cd2d939a7 Mon Sep 17 00:00:00 2001 From: Matthew Holt Date: Fri, 9 Aug 2019 12:05:47 -0600 Subject: Implement config adapters and beginning of Caddyfile adapter Along with several other changes, such as renaming caddyhttp.ServerRoute to caddyhttp.Route, exporting some types that were not exported before, and tweaking the caddytls TLS values to be more consistent. Notably, we also now disable automatic cert management for names which already have a cert (manually) loaded into the cache. These names no longer need to be specified in the "skip_certificates" field of the automatic HTTPS config, because they will be skipped automatically. --- caddyconfig/caddyfile/adapter.go | 93 +++ caddyconfig/caddyfile/dispenser.go | 333 +++++++++++ caddyconfig/caddyfile/dispenser_test.go | 316 +++++++++++ caddyconfig/caddyfile/lexer.go | 150 +++++ caddyconfig/caddyfile/lexer_test.go | 196 +++++++ caddyconfig/caddyfile/parse.go | 492 ++++++++++++++++ caddyconfig/caddyfile/parse_test.go | 718 ++++++++++++++++++++++++ caddyconfig/caddyfile/testdata/import_glob0.txt | 6 + caddyconfig/caddyfile/testdata/import_glob1.txt | 4 + caddyconfig/caddyfile/testdata/import_glob2.txt | 3 + caddyconfig/caddyfile/testdata/import_test1.txt | 2 + caddyconfig/caddyfile/testdata/import_test2.txt | 4 + 12 files changed, 2317 insertions(+) create mode 100644 caddyconfig/caddyfile/adapter.go create mode 100755 caddyconfig/caddyfile/dispenser.go create mode 100755 caddyconfig/caddyfile/dispenser_test.go create mode 100755 caddyconfig/caddyfile/lexer.go create mode 100755 caddyconfig/caddyfile/lexer_test.go create mode 100755 caddyconfig/caddyfile/parse.go create mode 100755 caddyconfig/caddyfile/parse_test.go create mode 100755 caddyconfig/caddyfile/testdata/import_glob0.txt create mode 100755 caddyconfig/caddyfile/testdata/import_glob1.txt create mode 100755 caddyconfig/caddyfile/testdata/import_glob2.txt create mode 100755 caddyconfig/caddyfile/testdata/import_test1.txt create mode 100755 caddyconfig/caddyfile/testdata/import_test2.txt (limited to 'caddyconfig/caddyfile') diff --git a/caddyconfig/caddyfile/adapter.go b/caddyconfig/caddyfile/adapter.go new file mode 100644 index 0000000..ab4905a --- /dev/null +++ b/caddyconfig/caddyfile/adapter.go @@ -0,0 +1,93 @@ +// Copyright 2015 Matthew Holt and The Caddy Authors +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +package caddyfile + +import ( + "bytes" + "encoding/json" + "fmt" + + "github.com/caddyserver/caddy/v2" + "github.com/caddyserver/caddy/v2/caddyconfig" +) + +// Adapter adapts Caddyfile to Caddy JSON. +type Adapter struct { + ServerType ServerType +} + +// Adapt converts the Caddyfile config in body to Caddy JSON. +func (a Adapter) Adapt(body []byte, options map[string]string) ([]byte, []caddyconfig.Warning, error) { + if a.ServerType == nil { + return nil, nil, fmt.Errorf("no server type") + } + if options == nil { + options = make(map[string]string) + } + + directives := a.ServerType.ValidDirectives() + + filename := options["filename"] + if filename == "" { + filename = "Caddyfile" + } + + serverBlocks, err := Parse(filename, bytes.NewReader(body), directives) + if err != nil { + return nil, nil, err + } + + cfg, warnings, err := a.ServerType.Setup(serverBlocks, options) + if err != nil { + return nil, warnings, err + } + + marshalFunc := json.Marshal + if options["pretty"] == "true" { + marshalFunc = caddyconfig.JSONIndent + } + result, err := marshalFunc(cfg) + + return result, warnings, err +} + +// Unmarshaler is a type that can unmarshal +// Caddyfile tokens to set itself up for a +// JSON encoding. The goal of an unmarshaler +// is not to set itself up for actual use, +// but to set itself up for being marshaled +// into JSON. Caddyfile-unmarshaled values +// will not be used directly; they will be +// encoded as JSON and then used from that. +type Unmarshaler interface { + UnmarshalCaddyfile(d *Dispenser) error +} + +// ServerType is a type that can evaluate a Caddyfile and set up a caddy config. +type ServerType interface { + // ValidDirectives returns a list of the + // server type's recognized directives. + ValidDirectives() []string + + // Setup takes the server blocks which + // contain tokens, as well as options + // (e.g. CLI flags) and creates a Caddy + // config, along with any warnings or + // an error. + Setup([]ServerBlock, map[string]string) (*caddy.Config, []caddyconfig.Warning, error) +} + +// Interface guard +var _ caddyconfig.Adapter = (*Adapter)(nil) diff --git a/caddyconfig/caddyfile/dispenser.go b/caddyconfig/caddyfile/dispenser.go new file mode 100755 index 0000000..1cf5d04 --- /dev/null +++ b/caddyconfig/caddyfile/dispenser.go @@ -0,0 +1,333 @@ +// Copyright 2015 Matthew Holt and The Caddy Authors +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +package caddyfile + +import ( + "errors" + "fmt" + "strings" +) + +// Dispenser is a type that dispenses tokens, similarly to a lexer, +// except that it can do so with some notion of structure. An empty +// Dispenser is invalid; call NewDispenser to make a proper instance. +type Dispenser struct { + filename string + tokens []Token + cursor int + nesting int +} + +// NewDispenser returns a Dispenser filled with the given tokens. +func NewDispenser(filename string, tokens []Token) *Dispenser { + return &Dispenser{ + filename: filename, + tokens: tokens, + cursor: -1, + } +} + +// Next loads the next token. Returns true if a token +// was loaded; false otherwise. If false, all tokens +// have been consumed. +func (d *Dispenser) Next() bool { + if d.cursor < len(d.tokens)-1 { + d.cursor++ + return true + } + return false +} + +// Prev moves to the previous token. It does the inverse +// of Next(). Generally, this should only be used in +// special cases such as deleting a token from the slice +// that d is iterating. In that case, without using Prev(), +// the dispenser would be pointing at the wrong token since +// deleting a token implicitly advances the cursor. +func (d *Dispenser) Prev() bool { + if d.cursor > 0 { + d.cursor-- + return true + } + return false +} + +// NextArg loads the next token if it is on the same +// line and if it is not a block opening (open curly +// brace). Returns true if an argument token was +// loaded; false otherwise. If false, all tokens on +// the line have been consumed except for potentially +// a block opening. It handles imported tokens +// correctly. +func (d *Dispenser) NextArg() bool { + if !d.nextOnSameLine() { + return false + } + if d.Val() == "{" { + // roll back; a block opening is not an argument + d.cursor-- + return false + } + return true +} + +// nextOnSameLine advances the cursor if the next +// token is on the same line of the same file. +func (d *Dispenser) nextOnSameLine() bool { + if d.cursor < 0 { + d.cursor++ + return true + } + if d.cursor >= len(d.tokens) { + return false + } + if d.cursor < len(d.tokens)-1 && + d.tokens[d.cursor].File == d.tokens[d.cursor+1].File && + d.tokens[d.cursor].Line+d.numLineBreaks(d.cursor) == d.tokens[d.cursor+1].Line { + d.cursor++ + return true + } + return false +} + +// NextLine loads the next token only if it is not on the same +// line as the current token, and returns true if a token was +// loaded; false otherwise. If false, there is not another token +// or it is on the same line. It handles imported tokens correctly. +func (d *Dispenser) NextLine() bool { + if d.cursor < 0 { + d.cursor++ + return true + } + if d.cursor >= len(d.tokens) { + return false + } + if d.cursor < len(d.tokens)-1 && + (d.tokens[d.cursor].File != d.tokens[d.cursor+1].File || + d.tokens[d.cursor].Line+d.numLineBreaks(d.cursor) < d.tokens[d.cursor+1].Line) { + d.cursor++ + return true + } + return false +} + +// NextBlock can be used as the condition of a for loop +// to load the next token as long as it opens a block or +// is already in a block. It returns true if a token was +// loaded, or false when the block's closing curly brace +// was loaded and thus the block ended. Nested blocks are +// not supported. +func (d *Dispenser) NextBlock() bool { + if d.nesting > 0 { + d.Next() + if d.Val() == "}" { + d.nesting-- + return false + } + return true + } + if !d.nextOnSameLine() { // block must open on same line + return false + } + if d.Val() != "{" { + d.cursor-- // roll back if not opening brace + return false + } + d.Next() + if d.Val() == "}" { + // open and then closed right away + return false + } + d.nesting++ + return true +} + +// Nested returns true if the token is currently nested +// inside a block (i.e. an open curly brace was consumed). +func (d *Dispenser) Nested() bool { + return d.nesting > 0 +} + +// Val gets the text of the current token. If there is no token +// loaded, it returns empty string. +func (d *Dispenser) Val() string { + if d.cursor < 0 || d.cursor >= len(d.tokens) { + return "" + } + return d.tokens[d.cursor].Text +} + +// Line gets the line number of the current token. If there is no token +// loaded, it returns 0. +func (d *Dispenser) Line() int { + if d.cursor < 0 || d.cursor >= len(d.tokens) { + return 0 + } + return d.tokens[d.cursor].Line +} + +// File gets the filename of the current token. If there is no token loaded, +// it returns the filename originally given when parsing started. +func (d *Dispenser) File() string { + if d.cursor < 0 || d.cursor >= len(d.tokens) { + return d.filename + } + if tokenFilename := d.tokens[d.cursor].File; tokenFilename != "" { + return tokenFilename + } + return d.filename +} + +// Args is a convenience function that loads the next arguments +// (tokens on the same line) into an arbitrary number of strings +// pointed to in targets. If there are fewer tokens available +// than string pointers, the remaining strings will not be changed +// and false will be returned. If there were enough tokens available +// to fill the arguments, then true will be returned. +func (d *Dispenser) Args(targets ...*string) bool { + for i := 0; i < len(targets); i++ { + if !d.NextArg() { + return false + } + *targets[i] = d.Val() + } + return true +} + +// RemainingArgs loads any more arguments (tokens on the same line) +// into a slice and returns them. Open curly brace tokens also indicate +// the end of arguments, and the curly brace is not included in +// the return value nor is it loaded. +func (d *Dispenser) RemainingArgs() []string { + var args []string + for d.NextArg() { + args = append(args, d.Val()) + } + return args +} + +// NewFromNextTokens returns a new dispenser with a copy of +// the tokens from the current token until the end of the +// "directive" whether that be to the end of the line or +// the end of a block that starts at the end of the line. +func (d *Dispenser) NewFromNextTokens() *Dispenser { + var tkns []Token + tkns = append(tkns, d.Token()) + for d.NextArg() { + tkns = append(tkns, d.Token()) + } + if d.Next() && d.Val() == "{" { + tkns = append(tkns, d.Token()) + for d.NextBlock() { + for d.Nested() { + tkns = append(tkns, d.Token()) + d.NextBlock() + } + } + tkns = append(tkns, d.Token()) + } else { + d.cursor-- + } + return NewDispenser(d.filename, tkns) +} + +// Token returns the current token. +func (d *Dispenser) Token() Token { + if d.cursor < 0 || d.cursor >= len(d.tokens) { + return Token{} + } + return d.tokens[d.cursor] +} + +// Cursor returns the current cursor (token index). +func (d *Dispenser) Cursor() int { + return d.cursor +} + +// ArgErr returns an argument error, meaning that another +// argument was expected but not found. In other words, +// a line break or open curly brace was encountered instead of +// an argument. +func (d *Dispenser) ArgErr() error { + if d.Val() == "{" { + return d.Err("Unexpected token '{', expecting argument") + } + return d.Errf("Wrong argument count or unexpected line ending after '%s'", d.Val()) +} + +// SyntaxErr creates a generic syntax error which explains what was +// found and what was expected. +func (d *Dispenser) SyntaxErr(expected string) error { + msg := fmt.Sprintf("%s:%d - Syntax error: Unexpected token '%s', expecting '%s'", d.File(), d.Line(), d.Val(), expected) + return errors.New(msg) +} + +// EOFErr returns an error indicating that the dispenser reached +// the end of the input when searching for the next token. +func (d *Dispenser) EOFErr() error { + return d.Errf("Unexpected EOF") +} + +// Err generates a custom parse-time error with a message of msg. +func (d *Dispenser) Err(msg string) error { + msg = fmt.Sprintf("%s:%d - Error during parsing: %s", d.File(), d.Line(), msg) + return errors.New(msg) +} + +// Errf is like Err, but for formatted error messages +func (d *Dispenser) Errf(format string, args ...interface{}) error { + return d.Err(fmt.Sprintf(format, args...)) +} + +// Delete deletes the current token and returns the updated slice +// of tokens. The cursor is not advanced to the next token. +// Because deletion modifies the underlying slice, this method +// should only be called if you have access to the original slice +// of tokens and/or are using the slice of tokens outside this +// Dispenser instance. If you do not re-assign the slice with the +// return value of this method, inconsistencies in the token +// array will become apparent (or worse, hide from you like they +// did me for 3 and a half freaking hours late one night). +func (d *Dispenser) Delete() []Token { + if d.cursor >= 0 && d.cursor < len(d.tokens)-1 { + d.tokens = append(d.tokens[:d.cursor], d.tokens[d.cursor+1:]...) + d.cursor-- + } + return d.tokens +} + +// numLineBreaks counts how many line breaks are in the token +// value given by the token index tknIdx. It returns 0 if the +// token does not exist or there are no line breaks. +func (d *Dispenser) numLineBreaks(tknIdx int) int { + if tknIdx < 0 || tknIdx >= len(d.tokens) { + return 0 + } + return strings.Count(d.tokens[tknIdx].Text, "\n") +} + +// isNewLine determines whether the current token is on a different +// line (higher line number) than the previous token. It handles imported +// tokens correctly. If there isn't a previous token, it returns true. +func (d *Dispenser) isNewLine() bool { + if d.cursor < 1 { + return true + } + if d.cursor > len(d.tokens)-1 { + return false + } + return d.tokens[d.cursor-1].File != d.tokens[d.cursor].File || + d.tokens[d.cursor-1].Line+d.numLineBreaks(d.cursor-1) < d.tokens[d.cursor].Line +} diff --git a/caddyconfig/caddyfile/dispenser_test.go b/caddyconfig/caddyfile/dispenser_test.go new file mode 100755 index 0000000..9860bed --- /dev/null +++ b/caddyconfig/caddyfile/dispenser_test.go @@ -0,0 +1,316 @@ +// Copyright 2015 Matthew Holt and The Caddy Authors +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +package caddyfile + +import ( + "io" + "log" + "reflect" + "strings" + "testing" +) + +func TestDispenser_Val_Next(t *testing.T) { + input := `host:port + dir1 arg1 + dir2 arg2 arg3 + dir3` + d := newTestDispenser(input) + + if val := d.Val(); val != "" { + t.Fatalf("Val(): Should return empty string when no token loaded; got '%s'", val) + } + + assertNext := func(shouldLoad bool, expectedCursor int, expectedVal string) { + if loaded := d.Next(); loaded != shouldLoad { + t.Errorf("Next(): Expected %v but got %v instead (val '%s')", shouldLoad, loaded, d.Val()) + } + if d.cursor != expectedCursor { + t.Errorf("Expected cursor to be %d, but was %d", expectedCursor, d.cursor) + } + if d.nesting != 0 { + t.Errorf("Nesting should be 0, was %d instead", d.nesting) + } + if val := d.Val(); val != expectedVal { + t.Errorf("Val(): Expected '%s' but got '%s'", expectedVal, val) + } + } + + assertNext(true, 0, "host:port") + assertNext(true, 1, "dir1") + assertNext(true, 2, "arg1") + assertNext(true, 3, "dir2") + assertNext(true, 4, "arg2") + assertNext(true, 5, "arg3") + assertNext(true, 6, "dir3") + // Note: This next test simply asserts existing behavior. + // If desired, we may wish to empty the token value after + // reading past the EOF. Open an issue if you want this change. + assertNext(false, 6, "dir3") +} + +func TestDispenser_NextArg(t *testing.T) { + input := `dir1 arg1 + dir2 arg2 arg3 + dir3` + d := newTestDispenser(input) + + assertNext := func(shouldLoad bool, expectedVal string, expectedCursor int) { + if d.Next() != shouldLoad { + t.Errorf("Next(): Should load token but got false instead (val: '%s')", d.Val()) + } + if d.cursor != expectedCursor { + t.Errorf("Next(): Expected cursor to be at %d, but it was %d", expectedCursor, d.cursor) + } + if val := d.Val(); val != expectedVal { + t.Errorf("Val(): Expected '%s' but got '%s'", expectedVal, val) + } + } + + assertNextArg := func(expectedVal string, loadAnother bool, expectedCursor int) { + if !d.NextArg() { + t.Error("NextArg(): Should load next argument but got false instead") + } + if d.cursor != expectedCursor { + t.Errorf("NextArg(): Expected cursor to be at %d, but it was %d", expectedCursor, d.cursor) + } + if val := d.Val(); val != expectedVal { + t.Errorf("Val(): Expected '%s' but got '%s'", expectedVal, val) + } + if !loadAnother { + if d.NextArg() { + t.Fatalf("NextArg(): Should NOT load another argument, but got true instead (val: '%s')", d.Val()) + } + if d.cursor != expectedCursor { + t.Errorf("NextArg(): Expected cursor to remain at %d, but it was %d", expectedCursor, d.cursor) + } + } + } + + assertNext(true, "dir1", 0) + assertNextArg("arg1", false, 1) + assertNext(true, "dir2", 2) + assertNextArg("arg2", true, 3) + assertNextArg("arg3", false, 4) + assertNext(true, "dir3", 5) + assertNext(false, "dir3", 5) +} + +func TestDispenser_NextLine(t *testing.T) { + input := `host:port + dir1 arg1 + dir2 arg2 arg3` + d := newTestDispenser(input) + + assertNextLine := func(shouldLoad bool, expectedVal string, expectedCursor int) { + if d.NextLine() != shouldLoad { + t.Errorf("NextLine(): Should load token but got false instead (val: '%s')", d.Val()) + } + if d.cursor != expectedCursor { + t.Errorf("NextLine(): Expected cursor to be %d, instead was %d", expectedCursor, d.cursor) + } + if val := d.Val(); val != expectedVal { + t.Errorf("Val(): Expected '%s' but got '%s'", expectedVal, val) + } + } + + assertNextLine(true, "host:port", 0) + assertNextLine(true, "dir1", 1) + assertNextLine(false, "dir1", 1) + d.Next() // arg1 + assertNextLine(true, "dir2", 3) + assertNextLine(false, "dir2", 3) + d.Next() // arg2 + assertNextLine(false, "arg2", 4) + d.Next() // arg3 + assertNextLine(false, "arg3", 5) +} + +func TestDispenser_NextBlock(t *testing.T) { + input := `foobar1 { + sub1 arg1 + sub2 + } + foobar2 { + }` + d := newTestDispenser(input) + + assertNextBlock := func(shouldLoad bool, expectedCursor, expectedNesting int) { + if loaded := d.NextBlock(); loaded != shouldLoad { + t.Errorf("NextBlock(): Should return %v but got %v", shouldLoad, loaded) + } + if d.cursor != expectedCursor { + t.Errorf("NextBlock(): Expected cursor to be %d, was %d", expectedCursor, d.cursor) + } + if d.nesting != expectedNesting { + t.Errorf("NextBlock(): Nesting should be %d, not %d", expectedNesting, d.nesting) + } + } + + assertNextBlock(false, -1, 0) + d.Next() // foobar1 + assertNextBlock(true, 2, 1) + assertNextBlock(true, 3, 1) + assertNextBlock(true, 4, 1) + assertNextBlock(false, 5, 0) + d.Next() // foobar2 + assertNextBlock(false, 8, 0) // empty block is as if it didn't exist +} + +func TestDispenser_Args(t *testing.T) { + var s1, s2, s3 string + input := `dir1 arg1 arg2 arg3 + dir2 arg4 arg5 + dir3 arg6 arg7 + dir4` + d := newTestDispenser(input) + + d.Next() // dir1 + + // As many strings as arguments + if all := d.Args(&s1, &s2, &s3); !all { + t.Error("Args(): Expected true, got false") + } + if s1 != "arg1" { + t.Errorf("Args(): Expected s1 to be 'arg1', got '%s'", s1) + } + if s2 != "arg2" { + t.Errorf("Args(): Expected s2 to be 'arg2', got '%s'", s2) + } + if s3 != "arg3" { + t.Errorf("Args(): Expected s3 to be 'arg3', got '%s'", s3) + } + + d.Next() // dir2 + + // More strings than arguments + if all := d.Args(&s1, &s2, &s3); all { + t.Error("Args(): Expected false, got true") + } + if s1 != "arg4" { + t.Errorf("Args(): Expected s1 to be 'arg4', got '%s'", s1) + } + if s2 != "arg5" { + t.Errorf("Args(): Expected s2 to be 'arg5', got '%s'", s2) + } + if s3 != "arg3" { + t.Errorf("Args(): Expected s3 to be unchanged ('arg3'), instead got '%s'", s3) + } + + // (quick cursor check just for kicks and giggles) + if d.cursor != 6 { + t.Errorf("Cursor should be 6, but is %d", d.cursor) + } + + d.Next() // dir3 + + // More arguments than strings + if all := d.Args(&s1); !all { + t.Error("Args(): Expected true, got false") + } + if s1 != "arg6" { + t.Errorf("Args(): Expected s1 to be 'arg6', got '%s'", s1) + } + + d.Next() // dir4 + + // No arguments or strings + if all := d.Args(); !all { + t.Error("Args(): Expected true, got false") + } + + // No arguments but at least one string + if all := d.Args(&s1); all { + t.Error("Args(): Expected false, got true") + } +} + +func TestDispenser_RemainingArgs(t *testing.T) { + input := `dir1 arg1 arg2 arg3 + dir2 arg4 arg5 + dir3 arg6 { arg7 + dir4` + d := newTestDispenser(input) + + d.Next() // dir1 + + args := d.RemainingArgs() + if expected := []string{"arg1", "arg2", "arg3"}; !reflect.DeepEqual(args, expected) { + t.Errorf("RemainingArgs(): Expected %v, got %v", expected, args) + } + + d.Next() // dir2 + + args = d.RemainingArgs() + if expected := []string{"arg4", "arg5"}; !reflect.DeepEqual(args, expected) { + t.Errorf("RemainingArgs(): Expected %v, got %v", expected, args) + } + + d.Next() // dir3 + + args = d.RemainingArgs() + if expected := []string{"arg6"}; !reflect.DeepEqual(args, expected) { + t.Errorf("RemainingArgs(): Expected %v, got %v", expected, args) + } + + d.Next() // { + d.Next() // arg7 + d.Next() // dir4 + + args = d.RemainingArgs() + if len(args) != 0 { + t.Errorf("RemainingArgs(): Expected %v, got %v", []string{}, args) + } +} + +func TestDispenser_ArgErr_Err(t *testing.T) { + input := `dir1 { + } + dir2 arg1 arg2` + d := newTestDispenser(input) + + d.cursor = 1 // { + + if err := d.ArgErr(); err == nil || !strings.Contains(err.Error(), "{") { + t.Errorf("ArgErr(): Expected an error message with { in it, but got '%v'", err) + } + + d.cursor = 5 // arg2 + + if err := d.ArgErr(); err == nil || !strings.Contains(err.Error(), "arg2") { + t.Errorf("ArgErr(): Expected an error message with 'arg2' in it; got '%v'", err) + } + + err := d.Err("foobar") + if err == nil { + t.Fatalf("Err(): Expected an error, got nil") + } + + if !strings.Contains(err.Error(), "Testfile:3") { + t.Errorf("Expected error message with filename:line in it; got '%v'", err) + } + + if !strings.Contains(err.Error(), "foobar") { + t.Errorf("Expected error message with custom message in it ('foobar'); got '%v'", err) + } +} + +func newTestDispenser(input string) *Dispenser { + tokens, err := allTokens(strings.NewReader(input)) + if err != nil && err != io.EOF { + log.Fatalf("getting all tokens from input: %v", err) + } + return NewDispenser("Testfile", tokens) +} diff --git a/caddyconfig/caddyfile/lexer.go b/caddyconfig/caddyfile/lexer.go new file mode 100755 index 0000000..efe648d --- /dev/null +++ b/caddyconfig/caddyfile/lexer.go @@ -0,0 +1,150 @@ +// Copyright 2015 Light Code Labs, LLC +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +package caddyfile + +import ( + "bufio" + "io" + "unicode" +) + +type ( + // lexer is a utility which can get values, token by + // token, from a Reader. A token is a word, and tokens + // are separated by whitespace. A word can be enclosed + // in quotes if it contains whitespace. + lexer struct { + reader *bufio.Reader + token Token + line int + } + + // Token represents a single parsable unit. + Token struct { + File string + Line int + Text string + } +) + +// load prepares the lexer to scan an input for tokens. +// It discards any leading byte order mark. +func (l *lexer) load(input io.Reader) error { + l.reader = bufio.NewReader(input) + l.line = 1 + + // discard byte order mark, if present + firstCh, _, err := l.reader.ReadRune() + if err != nil { + return err + } + if firstCh != 0xFEFF { + err := l.reader.UnreadRune() + if err != nil { + return err + } + } + + return nil +} + +// next loads the next token into the lexer. +// A token is delimited by whitespace, unless +// the token starts with a quotes character (") +// in which case the token goes until the closing +// quotes (the enclosing quotes are not included). +// Inside quoted strings, quotes may be escaped +// with a preceding \ character. No other chars +// may be escaped. The rest of the line is skipped +// if a "#" character is read in. Returns true if +// a token was loaded; false otherwise. +func (l *lexer) next() bool { + var val []rune + var comment, quoted, escaped bool + + makeToken := func() bool { + l.token.Text = string(val) + return true + } + + for { + ch, _, err := l.reader.ReadRune() + if err != nil { + if len(val) > 0 { + return makeToken() + } + if err == io.EOF { + return false + } + panic(err) + } + + if quoted { + if !escaped { + if ch == '\\' { + escaped = true + continue + } else if ch == '"' { + quoted = false + return makeToken() + } + } + if ch == '\n' { + l.line++ + } + if escaped { + // only escape quotes and newlines + if ch != '"' && ch != '\n' { + val = append(val, '\\') + } + } + val = append(val, ch) + escaped = false + continue + } + + if unicode.IsSpace(ch) { + if ch == '\r' { + continue + } + if ch == '\n' { + l.line++ + comment = false + } + if len(val) > 0 { + return makeToken() + } + continue + } + + if ch == '#' { + comment = true + } + + if comment { + continue + } + + if len(val) == 0 { + l.token = Token{Line: l.line} + if ch == '"' { + quoted = true + continue + } + } + + val = append(val, ch) + } +} diff --git a/caddyconfig/caddyfile/lexer_test.go b/caddyconfig/caddyfile/lexer_test.go new file mode 100755 index 0000000..f9a843c --- /dev/null +++ b/caddyconfig/caddyfile/lexer_test.go @@ -0,0 +1,196 @@ +// Copyright 2015 Light Code Labs, LLC +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +package caddyfile + +import ( + "log" + "strings" + "testing" +) + +type lexerTestCase struct { + input string + expected []Token +} + +func TestLexer(t *testing.T) { + testCases := []lexerTestCase{ + { + input: `host:123`, + expected: []Token{ + {Line: 1, Text: "host:123"}, + }, + }, + { + input: `host:123 + + directive`, + expected: []Token{ + {Line: 1, Text: "host:123"}, + {Line: 3, Text: "directive"}, + }, + }, + { + input: `host:123 { + directive + }`, + expected: []Token{ + {Line: 1, Text: "host:123"}, + {Line: 1, Text: "{"}, + {Line: 2, Text: "directive"}, + {Line: 3, Text: "}"}, + }, + }, + { + input: `host:123 { directive }`, + expected: []Token{ + {Line: 1, Text: "host:123"}, + {Line: 1, Text: "{"}, + {Line: 1, Text: "directive"}, + {Line: 1, Text: "}"}, + }, + }, + { + input: `host:123 { + #comment + directive + # comment + foobar # another comment + }`, + expected: []Token{ + {Line: 1, Text: "host:123"}, + {Line: 1, Text: "{"}, + {Line: 3, Text: "directive"}, + {Line: 5, Text: "foobar"}, + {Line: 6, Text: "}"}, + }, + }, + { + input: `a "quoted value" b + foobar`, + expected: []Token{ + {Line: 1, Text: "a"}, + {Line: 1, Text: "quoted value"}, + {Line: 1, Text: "b"}, + {Line: 2, Text: "foobar"}, + }, + }, + { + input: `A "quoted \"value\" inside" B`, + expected: []Token{ + {Line: 1, Text: "A"}, + {Line: 1, Text: `quoted "value" inside`}, + {Line: 1, Text: "B"}, + }, + }, + { + input: "A \"newline \\\ninside\" quotes", + expected: []Token{ + {Line: 1, Text: "A"}, + {Line: 1, Text: "newline \ninside"}, + {Line: 2, Text: "quotes"}, + }, + }, + { + input: `"don't\escape"`, + expected: []Token{ + {Line: 1, Text: `don't\escape`}, + }, + }, + { + input: `"don't\\escape"`, + expected: []Token{ + {Line: 1, Text: `don't\\escape`}, + }, + }, + { + input: `A "quoted value with line + break inside" { + foobar + }`, + expected: []Token{ + {Line: 1, Text: "A"}, + {Line: 1, Text: "quoted value with line\n\t\t\t\t\tbreak inside"}, + {Line: 2, Text: "{"}, + {Line: 3, Text: "foobar"}, + {Line: 4, Text: "}"}, + }, + }, + { + input: `"C:\php\php-cgi.exe"`, + expected: []Token{ + {Line: 1, Text: `C:\php\php-cgi.exe`}, + }, + }, + { + input: `empty "" string`, + expected: []Token{ + {Line: 1, Text: `empty`}, + {Line: 1, Text: ``}, + {Line: 1, Text: `string`}, + }, + }, + { + input: "skip those\r\nCR characters", + expected: []Token{ + {Line: 1, Text: "skip"}, + {Line: 1, Text: "those"}, + {Line: 2, Text: "CR"}, + {Line: 2, Text: "characters"}, + }, + }, + { + input: "\xEF\xBB\xBF:8080", // test with leading byte order mark + expected: []Token{ + {Line: 1, Text: ":8080"}, + }, + }, + } + + for i, testCase := range testCases { + actual := tokenize(testCase.input) + lexerCompare(t, i, testCase.expected, actual) + } +} + +func tokenize(input string) (tokens []Token) { + l := lexer{} + if err := l.load(strings.NewReader(input)); err != nil { + log.Printf("[ERROR] load failed: %v", err) + } + for l.next() { + tokens = append(tokens, l.token) + } + return +} + +func lexerCompare(t *testing.T, n int, expected, actual []Token) { + if len(expected) != len(actual) { + t.Errorf("Test case %d: expected %d token(s) but got %d", n, len(expected), len(actual)) + } + + for i := 0; i < len(actual) && i < len(expected); i++ { + if actual[i].Line != expected[i].Line { + t.Errorf("Test case %d token %d ('%s'): expected line %d but was line %d", + n, i, expected[i].Text, expected[i].Line, actual[i].Line) + break + } + if actual[i].Text != expected[i].Text { + t.Errorf("Test case %d token %d: expected text '%s' but was '%s'", + n, i, expected[i].Text, actual[i].Text) + break + } + } +} diff --git a/caddyconfig/caddyfile/parse.go b/caddyconfig/caddyfile/parse.go new file mode 100755 index 0000000..cc91e3d --- /dev/null +++ b/caddyconfig/caddyfile/parse.go @@ -0,0 +1,492 @@ +// Copyright 2015 Light Code Labs, LLC +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +package caddyfile + +import ( + "io" + "log" + "os" + "path/filepath" + "strings" +) + +// Parse parses the input just enough to group tokens, in +// order, by server block. No further parsing is performed. +// Server blocks are returned in the order in which they appear. +// Directives that do not appear in validDirectives will cause +// an error. If you do not want to check for valid directives, +// pass in nil instead. +func Parse(filename string, input io.Reader, validDirectives []string) ([]ServerBlock, error) { + tokens, err := allTokens(input) + if err != nil { + return nil, err + } + p := parser{Dispenser: NewDispenser(filename, tokens), validDirectives: validDirectives} + return p.parseAll() +} + +// allTokens lexes the entire input, but does not parse it. +// It returns all the tokens from the input, unstructured +// and in order. +func allTokens(input io.Reader) ([]Token, error) { + l := new(lexer) + err := l.load(input) + if err != nil { + return nil, err + } + var tokens []Token + for l.next() { + tokens = append(tokens, l.token) + } + return tokens, nil +} + +type parser struct { + *Dispenser + block ServerBlock // current server block being parsed + validDirectives []string // a directive must be valid or it's an error + eof bool // if we encounter a valid EOF in a hard place + definedSnippets map[string][]Token +} + +func (p *parser) parseAll() ([]ServerBlock, error) { + var blocks []ServerBlock + + for p.Next() { + err := p.parseOne() + if err != nil { + return blocks, err + } + if len(p.block.Keys) > 0 { + blocks = append(blocks, p.block) + } + } + + return blocks, nil +} + +func (p *parser) parseOne() error { + p.block = ServerBlock{Tokens: make(map[string][]Token)} + + return p.begin() +} + +func (p *parser) begin() error { + if len(p.tokens) == 0 { + return nil + } + + err := p.addresses() + + if err != nil { + return err + } + + if p.eof { + // this happens if the Caddyfile consists of only + // a line of addresses and nothing else + return nil + } + + if ok, name := p.isSnippet(); ok { + if p.definedSnippets == nil { + p.definedSnippets = map[string][]Token{} + } + if _, found := p.definedSnippets[name]; found { + return p.Errf("redeclaration of previously declared snippet %s", name) + } + // consume all tokens til matched close brace + tokens, err := p.snippetTokens() + if err != nil { + return err + } + p.definedSnippets[name] = tokens + // empty block keys so we don't save this block as a real server. + p.block.Keys = nil + return nil + } + + return p.blockContents() +} + +func (p *parser) addresses() error { + var expectingAnother bool + + for { + tkn := replaceEnvVars(p.Val()) + + // special case: import directive replaces tokens during parse-time + if tkn == "import" && p.isNewLine() { + err := p.doImport() + if err != nil { + return err + } + continue + } + + // Open brace definitely indicates end of addresses + if tkn == "{" { + if expectingAnother { + return p.Errf("Expected another address but had '%s' - check for extra comma", tkn) + } + break + } + + if tkn != "" { // empty token possible if user typed "" + // Trailing comma indicates another address will follow, which + // may possibly be on the next line + if tkn[len(tkn)-1] == ',' { + tkn = tkn[:len(tkn)-1] + expectingAnother = true + } else { + expectingAnother = false // but we may still see another one on this line + } + + p.block.Keys = append(p.block.Keys, tkn) + } + + // Advance token and possibly break out of loop or return error + hasNext := p.Next() + if expectingAnother && !hasNext { + return p.EOFErr() + } + if !hasNext { + p.eof = true + break // EOF + } + if !expectingAnother && p.isNewLine() { + break + } + } + + return nil +} + +func (p *parser) blockContents() error { + errOpenCurlyBrace := p.openCurlyBrace() + if errOpenCurlyBrace != nil { + // single-server configs don't need curly braces + p.cursor-- + } + + err := p.directives() + if err != nil { + return err + } + + // Only look for close curly brace if there was an opening + if errOpenCurlyBrace == nil { + err = p.closeCurlyBrace() + if err != nil { + return err + } + } + + return nil +} + +// directives parses through all the lines for directives +// and it expects the next token to be the first +// directive. It goes until EOF or closing curly brace +// which ends the server block. +func (p *parser) directives() error { + for p.Next() { + // end of server block + if p.Val() == "}" { + break + } + + // special case: import directive replaces tokens during parse-time + if p.Val() == "import" { + err := p.doImport() + if err != nil { + return err + } + p.cursor-- // cursor is advanced when we continue, so roll back one more + continue + } + + // normal case: parse a directive on this line + if err := p.directive(); err != nil { + return err + } + } + return nil +} + +// doImport swaps out the import directive and its argument +// (a total of 2 tokens) with the tokens in the specified file +// or globbing pattern. When the function returns, the cursor +// is on the token before where the import directive was. In +// other words, call Next() to access the first token that was +// imported. +func (p *parser) doImport() error { + // syntax checks + if !p.NextArg() { + return p.ArgErr() + } + importPattern := replaceEnvVars(p.Val()) + if importPattern == "" { + return p.Err("Import requires a non-empty filepath") + } + if p.NextArg() { + return p.Err("Import takes only one argument (glob pattern or file)") + } + // splice out the import directive and its argument (2 tokens total) + tokensBefore := p.tokens[:p.cursor-1] + tokensAfter := p.tokens[p.cursor+1:] + var importedTokens []Token + + // first check snippets. That is a simple, non-recursive replacement + if p.definedSnippets != nil && p.definedSnippets[importPattern] != nil { + importedTokens = p.definedSnippets[importPattern] + } else { + // make path relative to the file of the _token_ being processed rather + // than current working directory (issue #867) and then use glob to get + // list of matching filenames + absFile, err := filepath.Abs(p.Dispenser.File()) + if err != nil { + return p.Errf("Failed to get absolute path of file: %s: %v", p.Dispenser.filename, err) + } + + var matches []string + var globPattern string + if !filepath.IsAbs(importPattern) { + globPattern = filepath.Join(filepath.Dir(absFile), importPattern) + } else { + globPattern = importPattern + } + if strings.Count(globPattern, "*") > 1 || strings.Count(globPattern, "?") > 1 || + (strings.Contains(globPattern, "[") && strings.Contains(globPattern, "]")) { + // See issue #2096 - a pattern with many glob expansions can hang for too long + return p.Errf("Glob pattern may only contain one wildcard (*), but has others: %s", globPattern) + } + matches, err = filepath.Glob(globPattern) + + if err != nil { + return p.Errf("Failed to use import pattern %s: %v", importPattern, err) + } + if len(matches) == 0 { + if strings.ContainsAny(globPattern, "*?[]") { + log.Printf("[WARNING] No files matching import glob pattern: %s", importPattern) + } else { + return p.Errf("File to import not found: %s", importPattern) + } + } + + // collect all the imported tokens + + for _, importFile := range matches { + newTokens, err := p.doSingleImport(importFile) + if err != nil { + return err + } + importedTokens = append(importedTokens, newTokens...) + } + } + + // splice the imported tokens in the place of the import statement + // and rewind cursor so Next() will land on first imported token + p.tokens = append(tokensBefore, append(importedTokens, tokensAfter...)...) + p.cursor-- + + return nil +} + +// doSingleImport lexes the individual file at importFile and returns +// its tokens or an error, if any. +func (p *parser) doSingleImport(importFile string) ([]Token, error) { + file, err := os.Open(importFile) + if err != nil { + return nil, p.Errf("Could not import %s: %v", importFile, err) + } + defer file.Close() + + if info, err := file.Stat(); err != nil { + return nil, p.Errf("Could not import %s: %v", importFile, err) + } else if info.IsDir() { + return nil, p.Errf("Could not import %s: is a directory", importFile) + } + + importedTokens, err := allTokens(file) + if err != nil { + return nil, p.Errf("Could not read tokens while importing %s: %v", importFile, err) + } + + // Tack the file path onto these tokens so errors show the imported file's name + // (we use full, absolute path to avoid bugs: issue #1892) + filename, err := filepath.Abs(importFile) + if err != nil { + return nil, p.Errf("Failed to get absolute path of file: %s: %v", p.Dispenser.filename, err) + } + for i := 0; i < len(importedTokens); i++ { + importedTokens[i].File = filename + } + + return importedTokens, nil +} + +// directive collects tokens until the directive's scope +// closes (either end of line or end of curly brace block). +// It expects the currently-loaded token to be a directive +// (or } that ends a server block). The collected tokens +// are loaded into the current server block for later use +// by directive setup functions. +func (p *parser) directive() error { + dir := replaceEnvVars(p.Val()) + nesting := 0 + + if !p.validDirective(dir) { + return p.Errf("Unknown directive '%s'", dir) + } + + // The directive itself is appended as a relevant token + p.block.Tokens[dir] = append(p.block.Tokens[dir], p.tokens[p.cursor]) + + for p.Next() { + if p.Val() == "{" { + nesting++ + } else if p.isNewLine() && nesting == 0 { + p.cursor-- // read too far + break + } else if p.Val() == "}" && nesting > 0 { + nesting-- + } else if p.Val() == "}" && nesting == 0 { + return p.Err("Unexpected '}' because no matching opening brace") + } else if p.Val() == "import" && p.isNewLine() { + if err := p.doImport(); err != nil { + return err + } + p.cursor-- // cursor is advanced when we continue, so roll back one more + continue + } + p.tokens[p.cursor].Text = replaceEnvVars(p.tokens[p.cursor].Text) + p.block.Tokens[dir] = append(p.block.Tokens[dir], p.tokens[p.cursor]) + } + + if nesting > 0 { + return p.EOFErr() + } + return nil +} + +// openCurlyBrace expects the current token to be an +// opening curly brace. This acts like an assertion +// because it returns an error if the token is not +// a opening curly brace. It does NOT advance the token. +func (p *parser) openCurlyBrace() error { + if p.Val() != "{" { + return p.SyntaxErr("{") + } + return nil +} + +// closeCurlyBrace expects the current token to be +// a closing curly brace. This acts like an assertion +// because it returns an error if the token is not +// a closing curly brace. It does NOT advance the token. +func (p *parser) closeCurlyBrace() error { + if p.Val() != "}" { + return p.SyntaxErr("}") + } + return nil +} + +// validDirective returns true if dir is in p.validDirectives. +func (p *parser) validDirective(dir string) bool { + if p.validDirectives == nil { + return true + } + for _, d := range p.validDirectives { + if d == dir { + return true + } + } + return false +} + +// replaceEnvVars replaces environment variables that appear in the token +// and understands both the $UNIX and %WINDOWS% syntaxes. +func replaceEnvVars(s string) string { + s = replaceEnvReferences(s, "{%", "%}") + s = replaceEnvReferences(s, "{$", "}") + return s +} + +// replaceEnvReferences performs the actual replacement of env variables +// in s, given the placeholder start and placeholder end strings. +func replaceEnvReferences(s, refStart, refEnd string) string { + index := strings.Index(s, refStart) + for index != -1 { + endIndex := strings.Index(s[index:], refEnd) + if endIndex == -1 { + break + } + + endIndex += index + if endIndex > index+len(refStart) { + ref := s[index : endIndex+len(refEnd)] + s = strings.Replace(s, ref, os.Getenv(ref[len(refStart):len(ref)-len(refEnd)]), -1) + } else { + return s + } + index = strings.Index(s, refStart) + } + return s +} + +// ServerBlock associates any number of keys (usually addresses +// of some sort) with tokens (grouped by directive name). +type ServerBlock struct { + Keys []string + Tokens map[string][]Token +} + +func (p *parser) isSnippet() (bool, string) { + keys := p.block.Keys + // A snippet block is a single key with parens. Nothing else qualifies. + if len(keys) == 1 && strings.HasPrefix(keys[0], "(") && strings.HasSuffix(keys[0], ")") { + return true, strings.TrimSuffix(keys[0][1:], ")") + } + return false, "" +} + +// read and store everything in a block for later replay. +func (p *parser) snippetTokens() ([]Token, error) { + // snippet must have curlies. + err := p.openCurlyBrace() + if err != nil { + return nil, err + } + count := 1 + tokens := []Token{} + for p.Next() { + if p.Val() == "}" { + count-- + if count == 0 { + break + } + } + if p.Val() == "{" { + count++ + } + tokens = append(tokens, p.tokens[p.cursor]) + } + // make sure we're matched up + if count != 0 { + return nil, p.SyntaxErr("}") + } + return tokens, nil +} diff --git a/caddyconfig/caddyfile/parse_test.go b/caddyconfig/caddyfile/parse_test.go new file mode 100755 index 0000000..654c68d --- /dev/null +++ b/caddyconfig/caddyfile/parse_test.go @@ -0,0 +1,718 @@ +// Copyright 2015 Light Code Labs, LLC +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +package caddyfile + +import ( + "io/ioutil" + "os" + "path/filepath" + "strings" + "testing" +) + +func TestAllTokens(t *testing.T) { + input := strings.NewReader("a b c\nd e") + expected := []string{"a", "b", "c", "d", "e"} + tokens, err := allTokens(input) + + if err != nil { + t.Fatalf("Expected no error, got %v", err) + } + if len(tokens) != len(expected) { + t.Fatalf("Expected %d tokens, got %d", len(expected), len(tokens)) + } + + for i, val := range expected { + if tokens[i].Text != val { + t.Errorf("Token %d should be '%s' but was '%s'", i, val, tokens[i].Text) + } + } +} + +func TestParseOneAndImport(t *testing.T) { + testParseOne := func(input string) (ServerBlock, error) { + p := testParser(input) + p.Next() // parseOne doesn't call Next() to start, so we must + err := p.parseOne() + return p.block, err + } + + for i, test := range []struct { + input string + shouldErr bool + keys []string + tokens map[string]int // map of directive name to number of tokens expected + }{ + {`localhost`, false, []string{ + "localhost", + }, map[string]int{}}, + + {`localhost + dir1`, false, []string{ + "localhost", + }, map[string]int{ + "dir1": 1, + }}, + + {`localhost:1234 + dir1 foo bar`, false, []string{ + "localhost:1234", + }, map[string]int{ + "dir1": 3, + }}, + + {`localhost { + dir1 + }`, false, []string{ + "localhost", + }, map[string]int{ + "dir1": 1, + }}, + + {`localhost:1234 { + dir1 foo bar + dir2 + }`, false, []string{ + "localhost:1234", + }, map[string]int{ + "dir1": 3, + "dir2": 1, + }}, + + {`http://localhost https://localhost + dir1 foo bar`, false, []string{ + "http://localhost", + "https://localhost", + }, map[string]int{ + "dir1": 3, + }}, + + {`http://localhost https://localhost { + dir1 foo bar + }`, false, []string{ + "http://localhost", + "https://localhost", + }, map[string]int{ + "dir1": 3, + }}, + + {`http://localhost, https://localhost { + dir1 foo bar + }`, false, []string{ + "http://localhost", + "https://localhost", + }, map[string]int{ + "dir1": 3, + }}, + + {`http://localhost, { + }`, true, []string{ + "http://localhost", + }, map[string]int{}}, + + {`host1:80, http://host2.com + dir1 foo bar + dir2 baz`, false, []string{ + "host1:80", + "http://host2.com", + }, map[string]int{ + "dir1": 3, + "dir2": 2, + }}, + + {`http://host1.com, + http://host2.com, + https://host3.com`, false, []string{ + "http://host1.com", + "http://host2.com", + "https://host3.com", + }, map[string]int{}}, + + {`http://host1.com:1234, https://host2.com + dir1 foo { + bar baz + } + dir2`, false, []string{ + "http://host1.com:1234", + "https://host2.com", + }, map[string]int{ + "dir1": 6, + "dir2": 1, + }}, + + {`127.0.0.1 + dir1 { + bar baz + } + dir2 { + foo bar + }`, false, []string{ + "127.0.0.1", + }, map[string]int{ + "dir1": 5, + "dir2": 5, + }}, + + {`localhost + dir1 { + foo`, true, []string{ + "localhost", + }, map[string]int{ + "dir1": 3, + }}, + + {`localhost + dir1 { + }`, false, []string{ + "localhost", + }, map[string]int{ + "dir1": 3, + }}, + + {`localhost + dir1 { + } }`, true, []string{ + "localhost", + }, map[string]int{ + "dir1": 3, + }}, + + {`localhost + dir1 { + nested { + foo + } + } + dir2 foo bar`, false, []string{ + "localhost", + }, map[string]int{ + "dir1": 7, + "dir2": 3, + }}, + + {``, false, []string{}, map[string]int{}}, + + {`localhost + dir1 arg1 + import testdata/import_test1.txt`, false, []string{ + "localhost", + }, map[string]int{ + "dir1": 2, + "dir2": 3, + "dir3": 1, + }}, + + {`import testdata/import_test2.txt`, false, []string{ + "host1", + }, map[string]int{ + "dir1": 1, + "dir2": 2, + }}, + + {`import testdata/import_test1.txt testdata/import_test2.txt`, true, []string{}, map[string]int{}}, + + {`import testdata/not_found.txt`, true, []string{}, map[string]int{}}, + + {`""`, false, []string{}, map[string]int{}}, + + {``, false, []string{}, map[string]int{}}, + + // test cases found by fuzzing! + {`import }{$"`, true, []string{}, map[string]int{}}, + {`import /*/*.txt`, true, []string{}, map[string]int{}}, + {`import /???/?*?o`, true, []string{}, map[string]int{}}, + {`import /??`, true, []string{}, map[string]int{}}, + {`import /[a-z]`, true, []string{}, map[string]int{}}, + {`import {$}`, true, []string{}, map[string]int{}}, + {`import {%}`, true, []string{}, map[string]int{}}, + {`import {$$}`, true, []string{}, map[string]int{}}, + {`import {%%}`, true, []string{}, map[string]int{}}, + } { + result, err := testParseOne(test.input) + + if test.shouldErr && err == nil { + t.Errorf("Test %d: Expected an error, but didn't get one", i) + } + if !test.shouldErr && err != nil { + t.Errorf("Test %d: Expected no error, but got: %v", i, err) + } + + if len(result.Keys) != len(test.keys) { + t.Errorf("Test %d: Expected %d keys, got %d", + i, len(test.keys), len(result.Keys)) + continue + } + for j, addr := range result.Keys { + if addr != test.keys[j] { + t.Errorf("Test %d, key %d: Expected '%s', but was '%s'", + i, j, test.keys[j], addr) + } + } + + if len(result.Tokens) != len(test.tokens) { + t.Errorf("Test %d: Expected %d directives, had %d", + i, len(test.tokens), len(result.Tokens)) + continue + } + for directive, tokens := range result.Tokens { + if len(tokens) != test.tokens[directive] { + t.Errorf("Test %d, directive '%s': Expected %d tokens, counted %d", + i, directive, test.tokens[directive], len(tokens)) + continue + } + } + } +} + +func TestRecursiveImport(t *testing.T) { + testParseOne := func(input string) (ServerBlock, error) { + p := testParser(input) + p.Next() // parseOne doesn't call Next() to start, so we must + err := p.parseOne() + return p.block, err + } + + isExpected := func(got ServerBlock) bool { + if len(got.Keys) != 1 || got.Keys[0] != "localhost" { + t.Errorf("got keys unexpected: expect localhost, got %v", got.Keys) + return false + } + if len(got.Tokens) != 2 { + t.Errorf("got wrong number of tokens: expect 2, got %d", len(got.Tokens)) + return false + } + if len(got.Tokens["dir1"]) != 1 || len(got.Tokens["dir2"]) != 2 { + t.Errorf("got unexpect tokens: %v", got.Tokens) + return false + } + return true + } + + recursiveFile1, err := filepath.Abs("testdata/recursive_import_test1") + if err != nil { + t.Fatal(err) + } + recursiveFile2, err := filepath.Abs("testdata/recursive_import_test2") + if err != nil { + t.Fatal(err) + } + + // test relative recursive import + err = ioutil.WriteFile(recursiveFile1, []byte( + `localhost + dir1 + import recursive_import_test2`), 0644) + if err != nil { + t.Fatal(err) + } + defer os.Remove(recursiveFile1) + + err = ioutil.WriteFile(recursiveFile2, []byte("dir2 1"), 0644) + if err != nil { + t.Fatal(err) + } + defer os.Remove(recursiveFile2) + + // import absolute path + result, err := testParseOne("import " + recursiveFile1) + if err != nil { + t.Fatal(err) + } + if !isExpected(result) { + t.Error("absolute+relative import failed") + } + + // import relative path + result, err = testParseOne("import testdata/recursive_import_test1") + if err != nil { + t.Fatal(err) + } + if !isExpected(result) { + t.Error("relative+relative import failed") + } + + // test absolute recursive import + err = ioutil.WriteFile(recursiveFile1, []byte( + `localhost + dir1 + import `+recursiveFile2), 0644) + if err != nil { + t.Fatal(err) + } + + // import absolute path + result, err = testParseOne("import " + recursiveFile1) + if err != nil { + t.Fatal(err) + } + if !isExpected(result) { + t.Error("absolute+absolute import failed") + } + + // import relative path + result, err = testParseOne("import testdata/recursive_import_test1") + if err != nil { + t.Fatal(err) + } + if !isExpected(result) { + t.Error("relative+absolute import failed") + } +} + +func TestDirectiveImport(t *testing.T) { + testParseOne := func(input string) (ServerBlock, error) { + p := testParser(input) + p.Next() // parseOne doesn't call Next() to start, so we must + err := p.parseOne() + return p.block, err + } + + isExpected := func(got ServerBlock) bool { + if len(got.Keys) != 1 || got.Keys[0] != "localhost" { + t.Errorf("got keys unexpected: expect localhost, got %v", got.Keys) + return false + } + if len(got.Tokens) != 2 { + t.Errorf("got wrong number of tokens: expect 2, got %d", len(got.Tokens)) + return false + } + if len(got.Tokens["dir1"]) != 1 || len(got.Tokens["proxy"]) != 8 { + t.Errorf("got unexpect tokens: %v", got.Tokens) + return false + } + return true + } + + directiveFile, err := filepath.Abs("testdata/directive_import_test") + if err != nil { + t.Fatal(err) + } + + err = ioutil.WriteFile(directiveFile, []byte(`prop1 1 + prop2 2`), 0644) + if err != nil { + t.Fatal(err) + } + defer os.Remove(directiveFile) + + // import from existing file + result, err := testParseOne(`localhost + dir1 + proxy { + import testdata/directive_import_test + transparent + }`) + if err != nil { + t.Fatal(err) + } + if !isExpected(result) { + t.Error("directive import failed") + } + + // import from nonexistent file + _, err = testParseOne(`localhost + dir1 + proxy { + import testdata/nonexistent_file + transparent + }`) + if err == nil { + t.Fatal("expected error when importing a nonexistent file") + } +} + +func TestParseAll(t *testing.T) { + for i, test := range []struct { + input string + shouldErr bool + keys [][]string // keys per server block, in order + }{ + {`localhost`, false, [][]string{ + {"localhost"}, + }}, + + {`localhost:1234`, false, [][]string{ + {"localhost:1234"}, + }}, + + {`localhost:1234 { + } + localhost:2015 { + }`, false, [][]string{ + {"localhost:1234"}, + {"localhost:2015"}, + }}, + + {`localhost:1234, http://host2`, false, [][]string{ + {"localhost:1234", "http://host2"}, + }}, + + {`localhost:1234, http://host2,`, true, [][]string{}}, + + {`http://host1.com, http://host2.com { + } + https://host3.com, https://host4.com { + }`, false, [][]string{ + {"http://host1.com", "http://host2.com"}, + {"https://host3.com", "https://host4.com"}, + }}, + + {`import testdata/import_glob*.txt`, false, [][]string{ + {"glob0.host0"}, + {"glob0.host1"}, + {"glob1.host0"}, + {"glob2.host0"}, + }}, + + {`import notfound/*`, false, [][]string{}}, // glob needn't error with no matches + {`import notfound/file.conf`, true, [][]string{}}, // but a specific file should + } { + p := testParser(test.input) + blocks, err := p.parseAll() + + if test.shouldErr && err == nil { + t.Errorf("Test %d: Expected an error, but didn't get one", i) + } + if !test.shouldErr && err != nil { + t.Errorf("Test %d: Expected no error, but got: %v", i, err) + } + + if len(blocks) != len(test.keys) { + t.Errorf("Test %d: Expected %d server blocks, got %d", + i, len(test.keys), len(blocks)) + continue + } + for j, block := range blocks { + if len(block.Keys) != len(test.keys[j]) { + t.Errorf("Test %d: Expected %d keys in block %d, got %d", + i, len(test.keys[j]), j, len(block.Keys)) + continue + } + for k, addr := range block.Keys { + if addr != test.keys[j][k] { + t.Errorf("Test %d, block %d, key %d: Expected '%s', but got '%s'", + i, j, k, test.keys[j][k], addr) + } + } + } + } +} + +func TestEnvironmentReplacement(t *testing.T) { + os.Setenv("PORT", "8080") + os.Setenv("ADDRESS", "servername.com") + os.Setenv("FOOBAR", "foobar") + os.Setenv("PARTIAL_DIR", "r1") + + // basic test; unix-style env vars + p := testParser(`{$ADDRESS}`) + blocks, _ := p.parseAll() + if actual, expected := blocks[0].Keys[0], "servername.com"; expected != actual { + t.Errorf("Expected key to be '%s' but was '%s'", expected, actual) + } + + // basic test; unix-style env vars + p = testParser(`di{$PARTIAL_DIR}`) + blocks, _ = p.parseAll() + if actual, expected := blocks[0].Keys[0], "dir1"; expected != actual { + t.Errorf("Expected key to be '%s' but was '%s'", expected, actual) + } + + // multiple vars per token + p = testParser(`{$ADDRESS}:{$PORT}`) + blocks, _ = p.parseAll() + if actual, expected := blocks[0].Keys[0], "servername.com:8080"; expected != actual { + t.Errorf("Expected key to be '%s' but was '%s'", expected, actual) + } + + // windows-style var and unix style in same token + p = testParser(`{%ADDRESS%}:{$PORT}`) + blocks, _ = p.parseAll() + if actual, expected := blocks[0].Keys[0], "servername.com:8080"; expected != actual { + t.Errorf("Expected key to be '%s' but was '%s'", expected, actual) + } + + // reverse order + p = testParser(`{$ADDRESS}:{%PORT%}`) + blocks, _ = p.parseAll() + if actual, expected := blocks[0].Keys[0], "servername.com:8080"; expected != actual { + t.Errorf("Expected key to be '%s' but was '%s'", expected, actual) + } + + // env var in server block body as argument + p = testParser(":{%PORT%}\ndir1 {$FOOBAR}") + blocks, _ = p.parseAll() + if actual, expected := blocks[0].Keys[0], ":8080"; expected != actual { + t.Errorf("Expected key to be '%s' but was '%s'", expected, actual) + } + if actual, expected := blocks[0].Tokens["dir1"][1].Text, "foobar"; expected != actual { + t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) + } + + // combined windows env vars in argument + p = testParser(":{%PORT%}\ndir1 {%ADDRESS%}/{%FOOBAR%}") + blocks, _ = p.parseAll() + if actual, expected := blocks[0].Tokens["dir1"][1].Text, "servername.com/foobar"; expected != actual { + t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) + } + + // malformed env var (windows) + p = testParser(":1234\ndir1 {%ADDRESS}") + blocks, _ = p.parseAll() + if actual, expected := blocks[0].Tokens["dir1"][1].Text, "{%ADDRESS}"; expected != actual { + t.Errorf("Expected host to be '%s' but was '%s'", expected, actual) + } + + // malformed (non-existent) env var (unix) + p = testParser(`:{$PORT$}`) + blocks, _ = p.parseAll() + if actual, expected := blocks[0].Keys[0], ":"; expected != actual { + t.Errorf("Expected key to be '%s' but was '%s'", expected, actual) + } + + // in quoted field + p = testParser(":1234\ndir1 \"Test {$FOOBAR} test\"") + blocks, _ = p.parseAll() + if actual, expected := blocks[0].Tokens["dir1"][1].Text, "Test foobar test"; expected != actual { + t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) + } + + // after end token + p = testParser(":1234\nanswer \"{{ .Name }} {$FOOBAR}\"") + blocks, _ = p.parseAll() + if actual, expected := blocks[0].Tokens["answer"][1].Text, "{{ .Name }} foobar"; expected != actual { + t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) + } +} + +func testParser(input string) parser { + return parser{Dispenser: newTestDispenser(input)} +} + +func TestSnippets(t *testing.T) { + p := testParser(` + (common) { + gzip foo + errors stderr + } + http://example.com { + import common + } + `) + blocks, err := p.parseAll() + if err != nil { + t.Fatal(err) + } + for _, b := range blocks { + t.Log(b.Keys) + t.Log(b.Tokens) + } + if len(blocks) != 1 { + t.Fatalf("Expect exactly one server block. Got %d.", len(blocks)) + } + if actual, expected := blocks[0].Keys[0], "http://example.com"; expected != actual { + t.Errorf("Expected server name to be '%s' but was '%s'", expected, actual) + } + if len(blocks[0].Tokens) != 2 { + t.Fatalf("Server block should have tokens from import") + } + if actual, expected := blocks[0].Tokens["gzip"][0].Text, "gzip"; expected != actual { + t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) + } + if actual, expected := blocks[0].Tokens["errors"][1].Text, "stderr"; expected != actual { + t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) + } + +} + +func writeStringToTempFileOrDie(t *testing.T, str string) (pathToFile string) { + file, err := ioutil.TempFile("", t.Name()) + if err != nil { + panic(err) // get a stack trace so we know where this was called from. + } + if _, err := file.WriteString(str); err != nil { + panic(err) + } + if err := file.Close(); err != nil { + panic(err) + } + return file.Name() +} + +func TestImportedFilesIgnoreNonDirectiveImportTokens(t *testing.T) { + fileName := writeStringToTempFileOrDie(t, ` + http://example.com { + # This isn't an import directive, it's just an arg with value 'import' + basicauth / import password + } + `) + // Parse the root file that imports the other one. + p := testParser(`import ` + fileName) + blocks, err := p.parseAll() + if err != nil { + t.Fatal(err) + } + for _, b := range blocks { + t.Log(b.Keys) + t.Log(b.Tokens) + } + auth := blocks[0].Tokens["basicauth"] + line := auth[0].Text + " " + auth[1].Text + " " + auth[2].Text + " " + auth[3].Text + if line != "basicauth / import password" { + // Previously, it would be changed to: + // basicauth / import /path/to/test/dir/password + // referencing a file that (probably) doesn't exist and changing the + // password! + t.Errorf("Expected basicauth tokens to be 'basicauth / import password' but got %#q", line) + } +} + +func TestSnippetAcrossMultipleFiles(t *testing.T) { + // Make the derived Caddyfile that expects (common) to be defined. + fileName := writeStringToTempFileOrDie(t, ` + http://example.com { + import common + } + `) + + // Parse the root file that defines (common) and then imports the other one. + p := testParser(` + (common) { + gzip foo + } + import ` + fileName + ` + `) + + blocks, err := p.parseAll() + if err != nil { + t.Fatal(err) + } + for _, b := range blocks { + t.Log(b.Keys) + t.Log(b.Tokens) + } + if len(blocks) != 1 { + t.Fatalf("Expect exactly one server block. Got %d.", len(blocks)) + } + if actual, expected := blocks[0].Keys[0], "http://example.com"; expected != actual { + t.Errorf("Expected server name to be '%s' but was '%s'", expected, actual) + } + if len(blocks[0].Tokens) != 1 { + t.Fatalf("Server block should have tokens from import") + } + if actual, expected := blocks[0].Tokens["gzip"][0].Text, "gzip"; expected != actual { + t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) + } +} diff --git a/caddyconfig/caddyfile/testdata/import_glob0.txt b/caddyconfig/caddyfile/testdata/import_glob0.txt new file mode 100755 index 0000000..e610b5e --- /dev/null +++ b/caddyconfig/caddyfile/testdata/import_glob0.txt @@ -0,0 +1,6 @@ +glob0.host0 { + dir2 arg1 +} + +glob0.host1 { +} diff --git a/caddyconfig/caddyfile/testdata/import_glob1.txt b/caddyconfig/caddyfile/testdata/import_glob1.txt new file mode 100755 index 0000000..111eb04 --- /dev/null +++ b/caddyconfig/caddyfile/testdata/import_glob1.txt @@ -0,0 +1,4 @@ +glob1.host0 { + dir1 + dir2 arg1 +} diff --git a/caddyconfig/caddyfile/testdata/import_glob2.txt b/caddyconfig/caddyfile/testdata/import_glob2.txt new file mode 100755 index 0000000..c09f784 --- /dev/null +++ b/caddyconfig/caddyfile/testdata/import_glob2.txt @@ -0,0 +1,3 @@ +glob2.host0 { + dir2 arg1 +} diff --git a/caddyconfig/caddyfile/testdata/import_test1.txt b/caddyconfig/caddyfile/testdata/import_test1.txt new file mode 100755 index 0000000..dac7b29 --- /dev/null +++ b/caddyconfig/caddyfile/testdata/import_test1.txt @@ -0,0 +1,2 @@ +dir2 arg1 arg2 +dir3 \ No newline at end of file diff --git a/caddyconfig/caddyfile/testdata/import_test2.txt b/caddyconfig/caddyfile/testdata/import_test2.txt new file mode 100755 index 0000000..140c879 --- /dev/null +++ b/caddyconfig/caddyfile/testdata/import_test2.txt @@ -0,0 +1,4 @@ +host1 { + dir1 + dir2 arg1 +} \ No newline at end of file -- cgit v1.2.3 From c9980fd3671d873a7197a5ac4d6ac9d6b046abb6 Mon Sep 17 00:00:00 2001 From: Matthew Holt Date: Wed, 21 Aug 2019 10:46:35 -0600 Subject: Refactor Caddyfile adapter and module registration Use piles from which to draw config values. Module values can return their name, so now we can do two-way mapping from value to name and name to value; whereas before we could only map name to value. This was problematic with the Caddyfile adapter since it receives values and needs to know the name to put in the config. --- caddyconfig/caddyfile/adapter.go | 8 +- caddyconfig/caddyfile/dispenser.go | 30 +++--- caddyconfig/caddyfile/parse.go | 112 ++++++++++++++-------- caddyconfig/caddyfile/parse_test.go | 181 ++++++++++++++---------------------- 4 files changed, 163 insertions(+), 168 deletions(-) (limited to 'caddyconfig/caddyfile') diff --git a/caddyconfig/caddyfile/adapter.go b/caddyconfig/caddyfile/adapter.go index ab4905a..377f77b 100644 --- a/caddyconfig/caddyfile/adapter.go +++ b/caddyconfig/caddyfile/adapter.go @@ -37,14 +37,12 @@ func (a Adapter) Adapt(body []byte, options map[string]string) ([]byte, []caddyc options = make(map[string]string) } - directives := a.ServerType.ValidDirectives() - filename := options["filename"] if filename == "" { filename = "Caddyfile" } - serverBlocks, err := Parse(filename, bytes.NewReader(body), directives) + serverBlocks, err := Parse(filename, bytes.NewReader(body)) if err != nil { return nil, nil, err } @@ -77,10 +75,6 @@ type Unmarshaler interface { // ServerType is a type that can evaluate a Caddyfile and set up a caddy config. type ServerType interface { - // ValidDirectives returns a list of the - // server type's recognized directives. - ValidDirectives() []string - // Setup takes the server blocks which // contain tokens, as well as options // (e.g. CLI flags) and creates a Caddy diff --git a/caddyconfig/caddyfile/dispenser.go b/caddyconfig/caddyfile/dispenser.go index 1cf5d04..0d2c789 100755 --- a/caddyconfig/caddyfile/dispenser.go +++ b/caddyconfig/caddyfile/dispenser.go @@ -31,6 +31,7 @@ type Dispenser struct { } // NewDispenser returns a Dispenser filled with the given tokens. +// TODO: Get rid of the filename argument; it seems pointless here func NewDispenser(filename string, tokens []Token) *Dispenser { return &Dispenser{ filename: filename, @@ -51,15 +52,15 @@ func (d *Dispenser) Next() bool { } // Prev moves to the previous token. It does the inverse -// of Next(). Generally, this should only be used in -// special cases such as deleting a token from the slice -// that d is iterating. In that case, without using Prev(), -// the dispenser would be pointing at the wrong token since -// deleting a token implicitly advances the cursor. +// of Next(), except this function may decrement the cursor +// to -1 so that the next call to Next() points to the +// first token; this allows dispensing to "start over". This +// method returns true if the cursor ends up pointing to a +// valid token. func (d *Dispenser) Prev() bool { - if d.cursor > 0 { + if d.cursor > -1 { d.cursor-- - return true + return d.cursor > -1 } return false } @@ -223,8 +224,7 @@ func (d *Dispenser) RemainingArgs() []string { // "directive" whether that be to the end of the line or // the end of a block that starts at the end of the line. func (d *Dispenser) NewFromNextTokens() *Dispenser { - var tkns []Token - tkns = append(tkns, d.Token()) + tkns := []Token{d.Token()} for d.NextArg() { tkns = append(tkns, d.Token()) } @@ -245,10 +245,14 @@ func (d *Dispenser) NewFromNextTokens() *Dispenser { // Token returns the current token. func (d *Dispenser) Token() Token { - if d.cursor < 0 || d.cursor >= len(d.tokens) { + return d.TokenAt(d.cursor) +} + +func (d *Dispenser) TokenAt(cursor int) Token { + if cursor < 0 || cursor >= len(d.tokens) { return Token{} } - return d.tokens[d.cursor] + return d.tokens[cursor] } // Cursor returns the current cursor (token index). @@ -256,6 +260,10 @@ func (d *Dispenser) Cursor() int { return d.cursor } +func (d *Dispenser) Reset() { + d.cursor = -1 +} + // ArgErr returns an argument error, meaning that another // argument was expected but not found. In other words, // a line break or open curly brace was encountered instead of diff --git a/caddyconfig/caddyfile/parse.go b/caddyconfig/caddyfile/parse.go index cc91e3d..e5b25fc 100755 --- a/caddyconfig/caddyfile/parse.go +++ b/caddyconfig/caddyfile/parse.go @@ -28,12 +28,12 @@ import ( // Directives that do not appear in validDirectives will cause // an error. If you do not want to check for valid directives, // pass in nil instead. -func Parse(filename string, input io.Reader, validDirectives []string) ([]ServerBlock, error) { +func Parse(filename string, input io.Reader) ([]ServerBlock, error) { tokens, err := allTokens(input) if err != nil { return nil, err } - p := parser{Dispenser: NewDispenser(filename, tokens), validDirectives: validDirectives} + p := parser{Dispenser: NewDispenser(filename, tokens)} return p.parseAll() } @@ -56,9 +56,9 @@ func allTokens(input io.Reader) ([]Token, error) { type parser struct { *Dispenser block ServerBlock // current server block being parsed - validDirectives []string // a directive must be valid or it's an error eof bool // if we encounter a valid EOF in a hard place definedSnippets map[string][]Token + nesting int } func (p *parser) parseAll() ([]ServerBlock, error) { @@ -72,14 +72,16 @@ func (p *parser) parseAll() ([]ServerBlock, error) { if len(p.block.Keys) > 0 { blocks = append(blocks, p.block) } + if p.nesting > 0 { + return blocks, p.EOFErr() + } } return blocks, nil } func (p *parser) parseOne() error { - p.block = ServerBlock{Tokens: make(map[string][]Token)} - + p.block = ServerBlock{} return p.begin() } @@ -186,7 +188,7 @@ func (p *parser) blockContents() error { return err } - // Only look for close curly brace if there was an opening + // only look for close curly brace if there was an opening if errOpenCurlyBrace == nil { err = p.closeCurlyBrace() if err != nil { @@ -205,6 +207,7 @@ func (p *parser) directives() error { for p.Next() { // end of server block if p.Val() == "}" { + // p.nesting has already been decremented break } @@ -218,11 +221,15 @@ func (p *parser) directives() error { continue } - // normal case: parse a directive on this line + // normal case: parse a directive as a new segment + // (a "segment" is a line which starts with a directive + // and which ends at the end of the line or at the end of + // the block that is opened at the end of the line) if err := p.directive(); err != nil { return err } } + return nil } @@ -345,25 +352,24 @@ func (p *parser) doSingleImport(importFile string) ([]Token, error) { // are loaded into the current server block for later use // by directive setup functions. func (p *parser) directive() error { - dir := replaceEnvVars(p.Val()) - nesting := 0 + // evaluate any env vars in directive token + p.tokens[p.cursor].Text = replaceEnvVars(p.tokens[p.cursor].Text) - if !p.validDirective(dir) { - return p.Errf("Unknown directive '%s'", dir) - } + // a segment is a list of tokens associated with this directive + var segment Segment - // The directive itself is appended as a relevant token - p.block.Tokens[dir] = append(p.block.Tokens[dir], p.tokens[p.cursor]) + // the directive itself is appended as a relevant token + segment = append(segment, p.Token()) for p.Next() { if p.Val() == "{" { - nesting++ - } else if p.isNewLine() && nesting == 0 { + p.nesting++ + } else if p.isNewLine() && p.nesting == 0 { p.cursor-- // read too far break - } else if p.Val() == "}" && nesting > 0 { - nesting-- - } else if p.Val() == "}" && nesting == 0 { + } else if p.Val() == "}" && p.nesting > 0 { + p.nesting-- + } else if p.Val() == "}" && p.nesting == 0 { return p.Err("Unexpected '}' because no matching opening brace") } else if p.Val() == "import" && p.isNewLine() { if err := p.doImport(); err != nil { @@ -373,12 +379,15 @@ func (p *parser) directive() error { continue } p.tokens[p.cursor].Text = replaceEnvVars(p.tokens[p.cursor].Text) - p.block.Tokens[dir] = append(p.block.Tokens[dir], p.tokens[p.cursor]) + segment = append(segment, p.Token()) } - if nesting > 0 { + p.block.Segments = append(p.block.Segments, segment) + + if p.nesting > 0 { return p.EOFErr() } + return nil } @@ -404,19 +413,6 @@ func (p *parser) closeCurlyBrace() error { return nil } -// validDirective returns true if dir is in p.validDirectives. -func (p *parser) validDirective(dir string) bool { - if p.validDirectives == nil { - return true - } - for _, d := range p.validDirectives { - if d == dir { - return true - } - } - return false -} - // replaceEnvVars replaces environment variables that appear in the token // and understands both the $UNIX and %WINDOWS% syntaxes. func replaceEnvVars(s string) string { @@ -447,13 +443,6 @@ func replaceEnvReferences(s, refStart, refEnd string) string { return s } -// ServerBlock associates any number of keys (usually addresses -// of some sort) with tokens (grouped by directive name). -type ServerBlock struct { - Keys []string - Tokens map[string][]Token -} - func (p *parser) isSnippet() (bool, string) { keys := p.block.Keys // A snippet block is a single key with parens. Nothing else qualifies. @@ -480,6 +469,7 @@ func (p *parser) snippetTokens() ([]Token, error) { } } if p.Val() == "{" { + p.nesting++ count++ } tokens = append(tokens, p.tokens[p.cursor]) @@ -490,3 +480,43 @@ func (p *parser) snippetTokens() ([]Token, error) { } return tokens, nil } + +// ServerBlock associates any number of keys from the +// head of the server block with tokens, which are +// grouped by segments. +type ServerBlock struct { + Keys []string + Segments []Segment +} + +// DispenseDirective returns a dispenser that contains +// all the tokens in the server block. +func (sb ServerBlock) DispenseDirective(dir string) *Dispenser { + var tokens []Token + for _, seg := range sb.Segments { + if len(seg) > 0 && seg[0].Text == dir { + tokens = append(tokens, seg...) + } + } + return NewDispenser("", tokens) +} + +// Segment is a list of tokens which begins with a directive +// and ends at the end of the directive (either at the end of +// the line, or at the end of a block it opens). +type Segment []Token + +// Directive returns the directive name for the segment. +// The directive name is the text of the first token. +func (s Segment) Directive() string { + if len(s) > 0 { + return s[0].Text + } + return "" +} + +// NewDispenser returns a dispenser for this +// segment's tokens. +func (s Segment) NewDispenser() *Dispenser { + return NewDispenser("", s) +} diff --git a/caddyconfig/caddyfile/parse_test.go b/caddyconfig/caddyfile/parse_test.go index 654c68d..19959de 100755 --- a/caddyconfig/caddyfile/parse_test.go +++ b/caddyconfig/caddyfile/parse_test.go @@ -22,6 +22,8 @@ import ( "testing" ) +// TODO: re-enable all tests + func TestAllTokens(t *testing.T) { input := strings.NewReader("a b c\nd e") expected := []string{"a", "b", "c", "d", "e"} @@ -53,84 +55,67 @@ func TestParseOneAndImport(t *testing.T) { input string shouldErr bool keys []string - tokens map[string]int // map of directive name to number of tokens expected + numTokens []int // number of tokens to expect in each segment }{ {`localhost`, false, []string{ "localhost", - }, map[string]int{}}, + }, []int{}}, {`localhost dir1`, false, []string{ "localhost", - }, map[string]int{ - "dir1": 1, - }}, + }, []int{1}}, {`localhost:1234 dir1 foo bar`, false, []string{ "localhost:1234", - }, map[string]int{ - "dir1": 3, - }}, + }, []int{3}, + }, {`localhost { dir1 }`, false, []string{ "localhost", - }, map[string]int{ - "dir1": 1, - }}, + }, []int{1}}, {`localhost:1234 { dir1 foo bar dir2 }`, false, []string{ "localhost:1234", - }, map[string]int{ - "dir1": 3, - "dir2": 1, - }}, + }, []int{3, 1}}, {`http://localhost https://localhost dir1 foo bar`, false, []string{ "http://localhost", "https://localhost", - }, map[string]int{ - "dir1": 3, - }}, + }, []int{3}}, {`http://localhost https://localhost { dir1 foo bar }`, false, []string{ "http://localhost", "https://localhost", - }, map[string]int{ - "dir1": 3, - }}, + }, []int{3}}, {`http://localhost, https://localhost { dir1 foo bar }`, false, []string{ "http://localhost", "https://localhost", - }, map[string]int{ - "dir1": 3, - }}, + }, []int{3}}, {`http://localhost, { }`, true, []string{ "http://localhost", - }, map[string]int{}}, + }, []int{}}, {`host1:80, http://host2.com dir1 foo bar dir2 baz`, false, []string{ "host1:80", "http://host2.com", - }, map[string]int{ - "dir1": 3, - "dir2": 2, - }}, + }, []int{3, 2}}, {`http://host1.com, http://host2.com, @@ -138,7 +123,7 @@ func TestParseOneAndImport(t *testing.T) { "http://host1.com", "http://host2.com", "https://host3.com", - }, map[string]int{}}, + }, []int{}}, {`http://host1.com:1234, https://host2.com dir1 foo { @@ -147,10 +132,7 @@ func TestParseOneAndImport(t *testing.T) { dir2`, false, []string{ "http://host1.com:1234", "https://host2.com", - }, map[string]int{ - "dir1": 6, - "dir2": 1, - }}, + }, []int{6, 1}}, {`127.0.0.1 dir1 { @@ -160,34 +142,25 @@ func TestParseOneAndImport(t *testing.T) { foo bar }`, false, []string{ "127.0.0.1", - }, map[string]int{ - "dir1": 5, - "dir2": 5, - }}, + }, []int{5, 5}}, {`localhost dir1 { foo`, true, []string{ "localhost", - }, map[string]int{ - "dir1": 3, - }}, + }, []int{3}}, {`localhost dir1 { }`, false, []string{ "localhost", - }, map[string]int{ - "dir1": 3, - }}, + }, []int{3}}, {`localhost dir1 { } }`, true, []string{ "localhost", - }, map[string]int{ - "dir1": 3, - }}, + }, []int{}}, {`localhost dir1 { @@ -197,48 +170,38 @@ func TestParseOneAndImport(t *testing.T) { } dir2 foo bar`, false, []string{ "localhost", - }, map[string]int{ - "dir1": 7, - "dir2": 3, - }}, + }, []int{7, 3}}, - {``, false, []string{}, map[string]int{}}, + {``, false, []string{}, []int{}}, {`localhost dir1 arg1 import testdata/import_test1.txt`, false, []string{ "localhost", - }, map[string]int{ - "dir1": 2, - "dir2": 3, - "dir3": 1, - }}, + }, []int{2, 3, 1}}, {`import testdata/import_test2.txt`, false, []string{ "host1", - }, map[string]int{ - "dir1": 1, - "dir2": 2, - }}, + }, []int{1, 2}}, - {`import testdata/import_test1.txt testdata/import_test2.txt`, true, []string{}, map[string]int{}}, + {`import testdata/import_test1.txt testdata/import_test2.txt`, true, []string{}, []int{}}, - {`import testdata/not_found.txt`, true, []string{}, map[string]int{}}, + {`import testdata/not_found.txt`, true, []string{}, []int{}}, - {`""`, false, []string{}, map[string]int{}}, + {`""`, false, []string{}, []int{}}, - {``, false, []string{}, map[string]int{}}, + {``, false, []string{}, []int{}}, // test cases found by fuzzing! - {`import }{$"`, true, []string{}, map[string]int{}}, - {`import /*/*.txt`, true, []string{}, map[string]int{}}, - {`import /???/?*?o`, true, []string{}, map[string]int{}}, - {`import /??`, true, []string{}, map[string]int{}}, - {`import /[a-z]`, true, []string{}, map[string]int{}}, - {`import {$}`, true, []string{}, map[string]int{}}, - {`import {%}`, true, []string{}, map[string]int{}}, - {`import {$$}`, true, []string{}, map[string]int{}}, - {`import {%%}`, true, []string{}, map[string]int{}}, + {`import }{$"`, true, []string{}, []int{}}, + {`import /*/*.txt`, true, []string{}, []int{}}, + {`import /???/?*?o`, true, []string{}, []int{}}, + {`import /??`, true, []string{}, []int{}}, + {`import /[a-z]`, true, []string{}, []int{}}, + {`import {$}`, true, []string{}, []int{}}, + {`import {%}`, true, []string{}, []int{}}, + {`import {$$}`, true, []string{}, []int{}}, + {`import {%%}`, true, []string{}, []int{}}, } { result, err := testParseOne(test.input) @@ -261,15 +224,16 @@ func TestParseOneAndImport(t *testing.T) { } } - if len(result.Tokens) != len(test.tokens) { - t.Errorf("Test %d: Expected %d directives, had %d", - i, len(test.tokens), len(result.Tokens)) + if len(result.Segments) != len(test.numTokens) { + t.Errorf("Test %d: Expected %d segments, had %d", + i, len(test.numTokens), len(result.Segments)) continue } - for directive, tokens := range result.Tokens { - if len(tokens) != test.tokens[directive] { - t.Errorf("Test %d, directive '%s': Expected %d tokens, counted %d", - i, directive, test.tokens[directive], len(tokens)) + + for j, seg := range result.Segments { + if len(seg) != test.numTokens[j] { + t.Errorf("Test %d, segment %d: Expected %d tokens, counted %d", + i, j, test.numTokens[j], len(seg)) continue } } @@ -289,12 +253,12 @@ func TestRecursiveImport(t *testing.T) { t.Errorf("got keys unexpected: expect localhost, got %v", got.Keys) return false } - if len(got.Tokens) != 2 { - t.Errorf("got wrong number of tokens: expect 2, got %d", len(got.Tokens)) + if len(got.Segments) != 2 { + t.Errorf("got wrong number of segments: expect 2, got %d", len(got.Segments)) return false } - if len(got.Tokens["dir1"]) != 1 || len(got.Tokens["dir2"]) != 2 { - t.Errorf("got unexpect tokens: %v", got.Tokens) + if len(got.Segments[0]) != 1 || len(got.Segments[1]) != 2 { + t.Errorf("got unexpect tokens: %v", got.Segments) return false } return true @@ -384,12 +348,12 @@ func TestDirectiveImport(t *testing.T) { t.Errorf("got keys unexpected: expect localhost, got %v", got.Keys) return false } - if len(got.Tokens) != 2 { - t.Errorf("got wrong number of tokens: expect 2, got %d", len(got.Tokens)) + if len(got.Segments) != 2 { + t.Errorf("got wrong number of segments: expect 2, got %d", len(got.Segments)) return false } - if len(got.Tokens["dir1"]) != 1 || len(got.Tokens["proxy"]) != 8 { - t.Errorf("got unexpect tokens: %v", got.Tokens) + if len(got.Segments[0]) != 1 || len(got.Segments[1]) != 8 { + t.Errorf("got unexpect tokens: %v", got.Segments) return false } return true @@ -557,21 +521,21 @@ func TestEnvironmentReplacement(t *testing.T) { if actual, expected := blocks[0].Keys[0], ":8080"; expected != actual { t.Errorf("Expected key to be '%s' but was '%s'", expected, actual) } - if actual, expected := blocks[0].Tokens["dir1"][1].Text, "foobar"; expected != actual { + if actual, expected := blocks[0].Segments[0][1].Text, "foobar"; expected != actual { t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) } // combined windows env vars in argument p = testParser(":{%PORT%}\ndir1 {%ADDRESS%}/{%FOOBAR%}") blocks, _ = p.parseAll() - if actual, expected := blocks[0].Tokens["dir1"][1].Text, "servername.com/foobar"; expected != actual { + if actual, expected := blocks[0].Segments[0][1].Text, "servername.com/foobar"; expected != actual { t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) } // malformed env var (windows) p = testParser(":1234\ndir1 {%ADDRESS}") blocks, _ = p.parseAll() - if actual, expected := blocks[0].Tokens["dir1"][1].Text, "{%ADDRESS}"; expected != actual { + if actual, expected := blocks[0].Segments[0][1].Text, "{%ADDRESS}"; expected != actual { t.Errorf("Expected host to be '%s' but was '%s'", expected, actual) } @@ -585,22 +549,18 @@ func TestEnvironmentReplacement(t *testing.T) { // in quoted field p = testParser(":1234\ndir1 \"Test {$FOOBAR} test\"") blocks, _ = p.parseAll() - if actual, expected := blocks[0].Tokens["dir1"][1].Text, "Test foobar test"; expected != actual { + if actual, expected := blocks[0].Segments[0][1].Text, "Test foobar test"; expected != actual { t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) } // after end token p = testParser(":1234\nanswer \"{{ .Name }} {$FOOBAR}\"") blocks, _ = p.parseAll() - if actual, expected := blocks[0].Tokens["answer"][1].Text, "{{ .Name }} foobar"; expected != actual { + if actual, expected := blocks[0].Segments[0][1].Text, "{{ .Name }} foobar"; expected != actual { t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) } } -func testParser(input string) parser { - return parser{Dispenser: newTestDispenser(input)} -} - func TestSnippets(t *testing.T) { p := testParser(` (common) { @@ -617,7 +577,7 @@ func TestSnippets(t *testing.T) { } for _, b := range blocks { t.Log(b.Keys) - t.Log(b.Tokens) + t.Log(b.Segments) } if len(blocks) != 1 { t.Fatalf("Expect exactly one server block. Got %d.", len(blocks)) @@ -625,16 +585,15 @@ func TestSnippets(t *testing.T) { if actual, expected := blocks[0].Keys[0], "http://example.com"; expected != actual { t.Errorf("Expected server name to be '%s' but was '%s'", expected, actual) } - if len(blocks[0].Tokens) != 2 { - t.Fatalf("Server block should have tokens from import") + if len(blocks[0].Segments) != 2 { + t.Fatalf("Server block should have tokens from import, got: %+v", blocks[0]) } - if actual, expected := blocks[0].Tokens["gzip"][0].Text, "gzip"; expected != actual { + if actual, expected := blocks[0].Segments[0][0].Text, "gzip"; expected != actual { t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) } - if actual, expected := blocks[0].Tokens["errors"][1].Text, "stderr"; expected != actual { + if actual, expected := blocks[0].Segments[1][1].Text, "stderr"; expected != actual { t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) } - } func writeStringToTempFileOrDie(t *testing.T, str string) (pathToFile string) { @@ -666,9 +625,9 @@ func TestImportedFilesIgnoreNonDirectiveImportTokens(t *testing.T) { } for _, b := range blocks { t.Log(b.Keys) - t.Log(b.Tokens) + t.Log(b.Segments) } - auth := blocks[0].Tokens["basicauth"] + auth := blocks[0].Segments[0] line := auth[0].Text + " " + auth[1].Text + " " + auth[2].Text + " " + auth[3].Text if line != "basicauth / import password" { // Previously, it would be changed to: @@ -701,7 +660,7 @@ func TestSnippetAcrossMultipleFiles(t *testing.T) { } for _, b := range blocks { t.Log(b.Keys) - t.Log(b.Tokens) + t.Log(b.Segments) } if len(blocks) != 1 { t.Fatalf("Expect exactly one server block. Got %d.", len(blocks)) @@ -709,10 +668,14 @@ func TestSnippetAcrossMultipleFiles(t *testing.T) { if actual, expected := blocks[0].Keys[0], "http://example.com"; expected != actual { t.Errorf("Expected server name to be '%s' but was '%s'", expected, actual) } - if len(blocks[0].Tokens) != 1 { + if len(blocks[0].Segments) != 1 { t.Fatalf("Server block should have tokens from import") } - if actual, expected := blocks[0].Tokens["gzip"][0].Text, "gzip"; expected != actual { + if actual, expected := blocks[0].Segments[0][0].Text, "gzip"; expected != actual { t.Errorf("Expected argument to be '%s' but was '%s'", expected, actual) } } + +func testParser(input string) parser { + return parser{Dispenser: newTestDispenser(input)} +} -- cgit v1.2.3 From b2aa679c33f63ebec5bc1a21bca01f345dffebdd Mon Sep 17 00:00:00 2001 From: Matthew Holt Date: Wed, 21 Aug 2019 11:26:48 -0600 Subject: Fix snippet nesting bug --- caddyconfig/caddyfile/parse.go | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) (limited to 'caddyconfig/caddyfile') diff --git a/caddyconfig/caddyfile/parse.go b/caddyconfig/caddyfile/parse.go index e5b25fc..cc7ed25 100755 --- a/caddyconfig/caddyfile/parse.go +++ b/caddyconfig/caddyfile/parse.go @@ -459,23 +459,22 @@ func (p *parser) snippetTokens() ([]Token, error) { if err != nil { return nil, err } - count := 1 + nesting := 1 // count our own nesting in snippets tokens := []Token{} for p.Next() { if p.Val() == "}" { - count-- - if count == 0 { + nesting-- + if nesting == 0 { break } } if p.Val() == "{" { - p.nesting++ - count++ + nesting++ } tokens = append(tokens, p.tokens[p.cursor]) } // make sure we're matched up - if count != 0 { + if nesting != 0 { return nil, p.SyntaxErr("}") } return tokens, nil -- cgit v1.2.3