[haiku-commits] BRANCH HaikuPM-github.package-management [0e9ec70] src/kits/package
- From: HaikuPM-github.package-management <community@xxxxxxxxxxxx>
- To: haiku-commits@xxxxxxxxxxxxx
- Date: Tue, 9 Jul 2013 21:45:33 +0200 (CEST)
added 2 changesets to branch 'refs/remotes/HaikuPM-github/package-management'
old head: 0361a1da73a964be9d432cab0e123a51b37a3b34
new head: 0e9ec703dd88ade7b1adc8d28a3bec4484bcd63c
overview: https://github.com/haiku/HaikuPM/compare/0361a1d...0e9ec70
----------------------------------------------------------------------------
276c321: BPackageInfo::Parser::_ParseFlags(): fix indentation
0e9ec70: package info parser: improve string parsing
* Instead of two string token types (TOKEN_WORD, TOKEN_QUOTED_STRING),
there's now only one (TOKEN_STRING). Whether the string meets the
criteria is checked where needed. In most cases the check was already
done or not necessary anyway.
* Strings can now consist of an arbitrary sequence of quoted and
unquoted strings and escaping is also supported in unquoted string
segments.
* Among other things this fixes incorrect restrictions for resolvable
names and should also make quoting paths superfluous (unless they
contain separator characters).
[ Ingo Weinhold <ingo_weinhold@xxxxxx> ]
----------------------------------------------------------------------------
2 files changed, 137 insertions(+), 122 deletions(-)
src/kits/package/PackageInfoParser.cpp | 214 +++++++++++++++++------------
src/kits/package/PackageInfoParser.h | 45 ++----
############################################################################
Commit: 276c321bcdd0a579d3584198b1b41d95afc8275b
Author: Ingo Weinhold <ingo_weinhold@xxxxxx>
Date: Tue Jul 9 18:24:57 2013 UTC
BPackageInfo::Parser::_ParseFlags(): fix indentation
----------------------------------------------------------------------------
diff --git a/src/kits/package/PackageInfoParser.cpp
b/src/kits/package/PackageInfoParser.cpp
index 02ffe0e..7f15e8f 100644
--- a/src/kits/package/PackageInfoParser.cpp
+++ b/src/kits/package/PackageInfoParser.cpp
@@ -443,19 +443,19 @@ BPackageInfo::Parser::_ParseFlags()
virtual void operator()(const Token& token)
{
- if (token.type != TOKEN_WORD)
- throw ParseError("expected word (a flag)", token.pos);
-
- if (token.text.ICompare("approve_license") == 0)
- flags |= B_PACKAGE_FLAG_APPROVE_LICENSE;
- else if (token.text.ICompare("system_package") == 0)
- flags |= B_PACKAGE_FLAG_SYSTEM_PACKAGE;
- else {
+ if (token.type != TOKEN_WORD)
+ throw ParseError("expected word (a flag)",
token.pos);
+
+ if (token.text.ICompare("approve_license") == 0)
+ flags |= B_PACKAGE_FLAG_APPROVE_LICENSE;
+ else if (token.text.ICompare("system_package") == 0)
+ flags |= B_PACKAGE_FLAG_SYSTEM_PACKAGE;
+ else {
throw ParseError(
"expected 'approve_license' or
'system_package'",
- token.pos);
+ token.pos);
+ }
}
- }
} flagParser;
_ParseList(flagParser, true);
############################################################################
Commit: 0e9ec703dd88ade7b1adc8d28a3bec4484bcd63c
Author: Ingo Weinhold <ingo_weinhold@xxxxxx>
Date: Tue Jul 9 19:41:45 2013 UTC
package info parser: improve string parsing
* Instead of two string token types (TOKEN_WORD, TOKEN_QUOTED_STRING),
there's now only one (TOKEN_STRING). Whether the string meets the
criteria is checked where needed. In most cases the check was already
done or not necessary anyway.
* Strings can now consist of an arbitrary sequence of quoted and
unquoted strings and escaping is also supported in unquoted string
segments.
* Among other things this fixes incorrect restrictions for resolvable
names and should also make quoting paths superfluous (unless they
contain separator characters).
----------------------------------------------------------------------------
diff --git a/src/kits/package/PackageInfoParser.cpp
b/src/kits/package/PackageInfoParser.cpp
index 7f15e8f..377e8e8 100644
--- a/src/kits/package/PackageInfoParser.cpp
+++ b/src/kits/package/PackageInfoParser.cpp
@@ -10,6 +10,7 @@
#include <stdint.h>
#include <algorithm>
+#include <string>
namespace BPackageKit {
@@ -84,7 +85,7 @@ BPackageInfo::Parser::ParseVersion(const BString&
versionString,
fPos = versionString.String();
try {
- Token token(TOKEN_WORD, fPos, versionString.Length());
+ Token token(TOKEN_STRING, fPos, versionString.Length());
_ParseVersionValue(token, &_version, revisionIsOptional);
} catch (const ParseError& error) {
if (fListener != NULL) {
@@ -175,38 +176,76 @@ BPackageInfo::Parser::_NextToken()
}
return Token(TOKEN_OPERATOR_GREATER, tokenPos, 1);
- case '"':
- case '\'':
- {
- char quoteChar = *fPos;
- fPos++;
- const char* start = fPos;
- // anything until the next quote is part of the value
- bool lastWasEscape = false;
- while ((*fPos != quoteChar || lastWasEscape) && *fPos
!= '\0') {
- if (lastWasEscape)
- lastWasEscape = false;
- else if (*fPos == '\\')
- lastWasEscape = true;
- fPos++;
- }
- if (*fPos != quoteChar)
- throw ParseError("unterminated quoted-string",
tokenPos);
- const char* end = fPos++;
- return Token(TOKEN_QUOTED_STRING, start, end - start);
- }
-
default:
{
- const char* start = fPos;
- while (isalnum(*fPos) || *fPos == '.' || *fPos == '-'
- || *fPos == '_' || *fPos == ':' || *fPos == '+'
- || *fPos == '~') {
- fPos++;
+ std::string string;
+ char quoteChar = '\0';
+
+ for (; *fPos != '\0'; fPos++) {
+ char c = *fPos;
+ if (quoteChar != '\0') {
+ // within a quoted string segment
+ if (c == quoteChar) {
+ quoteChar = '\0';
+ continue;
+ }
+
+ if (c == '\\') {
+ // next char is escaped
+ c = *++fPos;
+ if (c == '\0') {
+ throw
ParseError("unterminated quoted-string",
+ tokenPos);
+ }
+
+ if (c == 'n')
+ c = '\n';
+ else if (c == 't')
+ c = '\t';
+ }
+
+ string += c;
+ } else {
+ // unquoted string segment
+ switch (c) {
+ case '"':
+ case '\'':
+ // quoted string start
+ quoteChar = c;
+ continue;
+
+ case '{':
+ case '}':
+ case '<':
+ case '=':
+ case '!':
+ case '>':
+ // a separator
character -- this ends the string
+ break;
+
+ case '\\':
+ // next char is escaped
+ c = *++fPos;
+ if (c == '\0') {
+ throw
ParseError("'\\' at end of string",
+
tokenPos);
+ }
+ string += c;
+ continue;
+
+ default:
+ if (isspace(c))
+ break;
+ string += c;
+ continue;
+ }
+
+ break;
+ }
}
- if (fPos == start)
- break;
- return Token(TOKEN_WORD, start, fPos - start);
+
+ return Token(TOKEN_STRING, tokenPos, fPos - tokenPos,
+ string.c_str());
}
}
@@ -226,8 +265,8 @@ void
BPackageInfo::Parser::_ParseStringValue(BString* value, const char** _tokenPos)
{
Token string = _NextToken();
- if (string.type != TOKEN_QUOTED_STRING && string.type != TOKEN_WORD)
- throw ParseError("expected quoted-string or word", string.pos);
+ if (string.type != TOKEN_STRING)
+ throw ParseError("expected string", string.pos);
*value = string.text;
if (_tokenPos != NULL)
@@ -239,7 +278,7 @@ void
BPackageInfo::Parser::_ParseArchitectureValue(BPackageArchitecture* value)
{
Token arch = _NextToken();
- if (arch.type == TOKEN_WORD) {
+ if (arch.type == TOKEN_STRING) {
for (int i = 0; i < B_PACKAGE_ARCHITECTURE_ENUM_COUNT; ++i) {
if
(arch.text.ICompare(BPackageInfo::kArchitectureNames[i]) == 0) {
*value = (BPackageArchitecture)i;
@@ -272,8 +311,8 @@ BPackageInfo::Parser::_ParseVersionValue(BPackageVersion*
value,
BPackageInfo::Parser::_ParseVersionValue(Token& word, BPackageVersion* value,
bool revisionIsOptional)
{
- if (word.type != TOKEN_WORD)
- throw ParseError("expected word (a version)", word.pos);
+ if (word.type != TOKEN_STRING)
+ throw ParseError("expected string (a version)", word.pos);
// get the revision number
uint32 revision = 0;
@@ -388,33 +427,33 @@ BPackageInfo::Parser::_ParseList(ListElementParser&
elementParser,
void
BPackageInfo::Parser::_ParseStringList(BStringList* value,
- bool allowQuotedStrings, bool convertToLowerCase)
+ bool requireResolvableName, bool convertToLowerCase)
{
struct StringParser : public ListElementParser {
BStringList* value;
- bool allowQuotedStrings;
+ bool requireResolvableName;
bool convertToLowerCase;
- StringParser(BStringList* value, bool allowQuotedStrings,
+ StringParser(BStringList* value, bool requireResolvableName,
bool convertToLowerCase)
:
value(value),
- allowQuotedStrings(allowQuotedStrings),
+ requireResolvableName(requireResolvableName),
convertToLowerCase(convertToLowerCase)
{
}
virtual void operator()(const Token& token)
{
- if (allowQuotedStrings) {
- if (token.type != TOKEN_QUOTED_STRING
- && token.type != TOKEN_WORD) {
- throw ParseError("expected
quoted-string or word",
- token.pos);
+ if (token.type != TOKEN_STRING)
+ throw ParseError("expected string", token.pos);
+
+ if (requireResolvableName) {
+ int32 errorPos;
+ if (!_IsValidResolvableName(token.text,
&errorPos)) {
+ throw ParseError("invalid character in
resolvable name",
+ token.pos + errorPos);
}
- } else {
- if (token.type != TOKEN_WORD)
- throw ParseError("expected word",
token.pos);
}
BString element(token.text);
@@ -423,7 +462,7 @@ BPackageInfo::Parser::_ParseStringList(BStringList* value,
value->Add(element);
}
- } stringParser(value, allowQuotedStrings, convertToLowerCase);
+ } stringParser(value, requireResolvableName, convertToLowerCase);
_ParseList(stringParser, true);
}
@@ -443,7 +482,7 @@ BPackageInfo::Parser::_ParseFlags()
virtual void operator()(const Token& token)
{
- if (token.type != TOKEN_WORD)
+ if (token.type != TOKEN_STRING)
throw ParseError("expected word (a flag)",
token.pos);
if (token.text.ICompare("approve_license") == 0)
@@ -482,7 +521,7 @@ BPackageInfo::Parser::_ParseResolvableList(
virtual void operator()(const Token& token)
{
- if (token.type != TOKEN_WORD) {
+ if (token.type != TOKEN_STRING) {
throw ParseError("expected word (a resolvable
name)",
token.pos);
}
@@ -507,7 +546,7 @@ BPackageInfo::Parser::_ParseResolvableList(
// parse compatible version
BPackageVersion compatibleVersion;
Token compatible = parser._NextToken();
- if (compatible.type == TOKEN_WORD
+ if (compatible.type == TOKEN_STRING
&& (compatible.text == "compat"
|| compatible.text == "compatible")) {
op = parser._NextToken();
@@ -548,7 +587,7 @@ BPackageInfo::Parser::_ParseResolvableExprList(
virtual void operator()(const Token& token)
{
- if (token.type != TOKEN_WORD) {
+ if (token.type != TOKEN_STRING) {
throw ParseError("expected word (a resolvable
name)",
token.pos);
}
@@ -571,7 +610,7 @@ BPackageInfo::Parser::_ParseResolvableExprList(
if (basePackage != NULL) {
Token base = parser._NextToken();
- if (base.type == TOKEN_WORD &&
base.text == "base") {
+ if (base.type == TOKEN_STRING &&
base.text == "base") {
if (!basePackage->IsEmpty()) {
throw ParseError(
"multiple packages
marked as base package",
@@ -621,8 +660,8 @@ BPackageInfo::Parser::_ParseGlobalWritableFileInfos(
virtual void operator()(const Token& token)
{
- if (token.type != TOKEN_WORD && token.type !=
TOKEN_QUOTED_STRING) {
- throw ParseError("expected string (a settings
file path)",
+ if (token.type != TOKEN_STRING) {
+ throw ParseError("expected string (a file
path)",
token.pos);
}
@@ -631,12 +670,13 @@ BPackageInfo::Parser::_ParseGlobalWritableFileInfos(
bool isDirectory = false;
Token nextToken = parser._NextToken();
- if (nextToken.type == TOKEN_WORD && nextToken.text ==
"directory") {
+ if (nextToken.type == TOKEN_STRING
+ && nextToken.text == "directory") {
isDirectory = true;
nextToken = parser._NextToken();
}
- if (nextToken.type == TOKEN_WORD) {
+ if (nextToken.type == TOKEN_STRING) {
const char* const* end =
kWritableFileUpdateTypes
+
B_WRITABLE_FILE_UPDATE_TYPE_ENUM_COUNT;
const char* const* found =
std::find(kWritableFileUpdateTypes,
@@ -685,7 +725,7 @@ BPackageInfo::Parser::_ParseUserSettingsFileInfos(
virtual void operator()(const Token& token)
{
- if (token.type != TOKEN_WORD && token.type !=
TOKEN_QUOTED_STRING) {
+ if (token.type != TOKEN_STRING) {
throw ParseError("expected string (a settings
file path)",
token.pos);
}
@@ -694,13 +734,13 @@ BPackageInfo::Parser::_ParseUserSettingsFileInfos(
bool isDirectory = false;
Token nextToken = parser._NextToken();
- if (nextToken.type == TOKEN_WORD && nextToken.text ==
"directory") {
+ if (nextToken.type == TOKEN_STRING
+ && nextToken.text == "directory") {
isDirectory = true;
- } else if (nextToken.type == TOKEN_WORD
+ } else if (nextToken.type == TOKEN_STRING
&& nextToken.text == "template") {
nextToken = parser._NextToken();
- if (nextToken.type != TOKEN_WORD
- && nextToken.type !=
TOKEN_QUOTED_STRING) {
+ if (nextToken.type != TOKEN_STRING) {
throw ParseError(
"expected string (a settings
template file path)",
nextToken.pos);
@@ -744,9 +784,9 @@ BPackageInfo::Parser::_ParseUsers(UserList* users)
virtual void operator()(const Token& token)
{
- if (token.type != TOKEN_WORD) {
- throw ParseError("expected a user name",
- token.pos);
+ if (token.type != TOKEN_STRING
+ || !BUser::IsValidUserName(token.text)) {
+ throw ParseError("expected a user name",
token.pos);
}
BString realName;
@@ -756,31 +796,28 @@ BPackageInfo::Parser::_ParseUsers(UserList* users)
for (;;) {
Token nextToken = parser._NextToken();
- if (nextToken.type != TOKEN_WORD) {
+ if (nextToken.type != TOKEN_STRING) {
parser._RewindTo(nextToken);
break;
}
if (nextToken.text == "real-name") {
nextToken = parser._NextToken();
- if (nextToken.type != TOKEN_WORD
- && nextToken.type !=
TOKEN_QUOTED_STRING) {
+ if (nextToken.type != TOKEN_STRING) {
throw ParseError("expected
string (a user real name)",
nextToken.pos);
}
realName = nextToken.text;
} else if (nextToken.text == "home") {
nextToken = parser._NextToken();
- if (nextToken.type != TOKEN_WORD
- && nextToken.type !=
TOKEN_QUOTED_STRING) {
+ if (nextToken.type != TOKEN_STRING) {
throw ParseError("expected
string (a home path)",
nextToken.pos);
}
home = nextToken.text;
} else if (nextToken.text == "shell") {
nextToken = parser._NextToken();
- if (nextToken.type != TOKEN_WORD
- && nextToken.type !=
TOKEN_QUOTED_STRING) {
+ if (nextToken.type != TOKEN_STRING) {
throw ParseError("expected
string (a shell path)",
nextToken.pos);
}
@@ -788,7 +825,8 @@ BPackageInfo::Parser::_ParseUsers(UserList* users)
} else if (nextToken.text == "groups") {
for (;;) {
nextToken = parser._NextToken();
- if (nextToken.type ==
TOKEN_WORD) {
+ if (nextToken.type ==
TOKEN_STRING
+ &&
BUser::IsValidUserName(nextToken.text)) {
if
(!groups.Add(nextToken.text))
throw
std::bad_alloc();
} else if (nextToken.type ==
TOKEN_ITEM_SEPARATOR
@@ -811,10 +849,10 @@ BPackageInfo::Parser::_ParseUsers(UserList* users)
BString templatePath;
Token nextToken = parser._NextToken();
- if (nextToken.type == TOKEN_WORD && nextToken.text ==
"template") {
+ if (nextToken.type == TOKEN_STRING
+ && nextToken.text == "template") {
nextToken = parser._NextToken();
- if (nextToken.type != TOKEN_WORD
- && nextToken.type !=
TOKEN_QUOTED_STRING) {
+ if (nextToken.type != TOKEN_STRING) {
throw ParseError(
"expected string (a settings
template file path)",
nextToken.pos);
@@ -853,8 +891,8 @@ BPackageInfo::Parser::_Parse(BPackageInfo* packageInfo)
if (t.type == TOKEN_ITEM_SEPARATOR)
continue;
- if (t.type != TOKEN_WORD)
- throw ParseError("expected word (a variable name)",
t.pos);
+ if (t.type != TOKEN_STRING)
+ throw ParseError("expected string (a variable name)",
t.pos);
BPackageInfoAttributeID attribute = B_PACKAGE_INFO_ENUM_COUNT;
for (int i = 0; i < B_PACKAGE_INFO_ENUM_COUNT; i++) {
@@ -986,7 +1024,7 @@ BPackageInfo::Parser::_Parse(BPackageInfo* packageInfo)
break;
case B_PACKAGE_INFO_REPLACES:
- _ParseStringList(&packageInfo->fReplacesList,
false, true);
+ _ParseStringList(&packageInfo->fReplacesList,
true);
break;
case B_PACKAGE_INFO_FLAGS:
diff --git a/src/kits/package/PackageInfoParser.h
b/src/kits/package/PackageInfoParser.h
index 3c94b00..613ec68 100644
--- a/src/kits/package/PackageInfoParser.h
+++ b/src/kits/package/PackageInfoParser.h
@@ -34,8 +34,7 @@ private:
friend struct ListElementParser;
enum TokenType {
- TOKEN_WORD,
- TOKEN_QUOTED_STRING,
+ TOKEN_STRING,
TOKEN_OPERATOR_ASSIGN,
TOKEN_OPERATOR_LESS,
TOKEN_OPERATOR_LESS_EQUAL,
@@ -67,7 +66,7 @@ private:
void
_ParseList(ListElementParser& elementParser,
bool
allowSingleNonListElement);
void
_ParseStringList(BStringList* value,
- bool
allowQuotedStrings = true,
+ bool
requireResolvableName = false,
bool
convertToLowerCase = false);
void _ParseResolvableList(
BObjectList<BPackageResolvable>* value);
@@ -118,38 +117,16 @@ struct BPackageInfo::Parser::Token {
BString text;
const char* pos;
- Token(TokenType _type, const char* _pos, int length = 0)
- : type(_type), pos(_pos)
+ Token(TokenType _type, const char* _pos, int length = 0,
+ const char* text = NULL)
+ :
+ type(_type),
+ pos(_pos)
{
- if (length != 0) {
- text.SetTo(pos, length);
-
- if (type == TOKEN_QUOTED_STRING) {
- // unescape value of quoted string
- char* value = text.LockBuffer(length);
- if (value == NULL)
- return;
- int index = 0;
- int newIndex = 0;
- bool lastWasEscape = false;
- while (char c = value[index++]) {
- if (lastWasEscape) {
- lastWasEscape = false;
- // map \n to newline and \t to
tab
- if (c == 'n')
- c = '\n';
- else if (c == 't')
- c = '\t';
- } else if (c == '\\') {
- lastWasEscape = true;
- continue;
- }
- value[newIndex++] = c;
- }
- value[newIndex] = '\0';
- text.UnlockBuffer(newIndex);
- }
- }
+ if (text != NULL)
+ this->text = text;
+ else if (length != 0)
+ this->text.SetTo(pos, length);
}
operator bool() const
Other related posts:
- » [haiku-commits] BRANCH HaikuPM-github.package-management [0e9ec70] src/kits/package - HaikuPM-github . package-management
