Add space separated argument parser with quoting

6 files changed, 196 insertions, 1 deletions

diff --git a/src/Makefile.am b/src/Makefile.am
index 5a779fa..54ff40d 100644
--- a/src/Makefile.am
+++ b/src/Makefile.am
@@ -12,6 +12,7 @@ libcgi_la_SOURCES = cgi.hh common.hh cgi.cc \
 					query_parser.hh query_parser.cc \
 					header_parser.hh header_parser.cc \
 					strutils.hh strutils.cc \
+					args.hh args.cc \
 				    multipart_formdata_parser.hh multipart_formdata_parser.cc
 libcgi_la_CPPFLAGS = $(AM_CPPFLAGS) @FASTCGI_CFLAGS@
 libcgi_la_LIBADD = @FASTCGI_LIBS@
diff --git a/src/args.cc b/src/args.cc
new file mode 100644
index 0000000..1c0ce17
--- /dev/null
+++ b/src/args.cc
@@ -0,0 +1,63 @@
+#include "common.hh"
+
+#include "args.hh"
+
+namespace stuff {
+
+// static
+bool Args::parse(const std::string& input, std::vector<std::string>* output,
+                 bool nice) {
+    size_t last = 0, i = 0;
+    std::string arg;
+    output->clear();
+    while (i < input.size() && input[i] == ' ') ++i;
+    if (i == input.size()) return true;
+    while (i < input.size()) {
+        if (input[i] == ' ') {
+            arg.append(input.substr(last, i - last));
+            output->push_back(arg);
+            arg.clear();
+            ++i;
+            while (i < input.size() && input[i] == ' ') ++i;
+            if (i == input.size()) return true;
+            last = i;
+        } else if (input[i] == '\'' || input[i] == '"') {
+            const char closing = input[i];
+            std::string text;
+            size_t j = i + 1;
+            size_t last_j = j;
+            while (j < input.size()) {
+                if (input[j] == closing) {
+                    break;
+                } else if (input[j] == '\\') {
+                    text.append(input.substr(last_j, j - last_j));
+                    if (j == input.size()) {
+                        if (!nice) return false;
+                        break;
+                    }
+                    text.push_back(input[++j]);
+                    last_j = ++j;
+                } else {
+                    ++j;
+                }
+            }
+            if (j < input.size()) {
+                arg.append(input.substr(last, i - last));
+                arg.append(text);
+                arg.append(input.substr(last_j, j - last_j));
+                last = i = j + 1;
+            } else {
+                // No closing char
+                if (!nice) return false;
+                ++i;
+            }
+        } else {
+            ++i;
+        }
+    }
+    arg.append(input.substr(last, i - last));
+    output->push_back(arg);
+    return true;
+}
+
+}  // namespace stuff
diff --git a/src/args.hh b/src/args.hh
new file mode 100644
index 0000000..7a0f4fc
--- /dev/null
+++ b/src/args.hh
@@ -0,0 +1,22 @@
+#ifndef ARGS_HH
+#define ARGS_HH
+
+#include <string>
+#include <vector>
+
+namespace stuff {
+
+class Args {
+public:
+    static bool parse(const std::string& input,
+                      std::vector<std::string>* output,
+                      bool nice = true);
+
+private:
+    Args() = delete;
+    ~Args() = delete;
+};
+
+}  // namespace stuff
+
+#endif /* ARGS_HH */
diff --git a/test/.gitignore b/test/.gitignore
index 8ab7990..d859557 100644
--- a/test/.gitignore
+++ b/test/.gitignore
@@ -3,3 +3,4 @@
 /test-header-parser
 /test-multipart-formdata-parser
 /test-query-parser
+/test-args
diff --git a/test/Makefile.am b/test/Makefile.am
index 3017b6f..5711aa4 100644
--- a/test/Makefile.am
+++ b/test/Makefile.am
@@ -2,7 +2,8 @@ MAINTAINERCLEANFILES = Makefile.in
 
 AM_CPPFLAGS = @DEFINES@ -I$(top_builddir)/src
 
-TESTS = test-query-parser test-header-parser test-multipart-formdata-parser
+TESTS = test-query-parser test-header-parser test-multipart-formdata-parser \
+		test-args
 
 check_PROGRAMS = $(TESTS)
 
@@ -14,3 +15,6 @@ test_header_parser_LDADD = $(top_srcdir)/src/libcgi.la
 
 test_multipart_formdata_parser_SOURCES = test-multipart-formdata-parser.cc
 test_multipart_formdata_parser_LDADD = $(top_srcdir)/src/libcgi.la
+
+test_args_SOURCES = test-args.cc
+test_args_LDADD = $(top_srcdir)/src/libcgi.la
diff --git a/test/test-args.cc b/test/test-args.cc
new file mode 100644
index 0000000..a9d4328
--- /dev/null
+++ b/test/test-args.cc
@@ -0,0 +1,104 @@
+#include "common.hh"
+
+#include <cstdarg>
+#include <iostream>
+
+#include "args.hh"
+
+using namespace stuff;
+
+namespace {
+
+bool compare(const std::string& in, const std::vector<std::string>& out,
+             va_list args) {
+    for (const auto& arg : out) {
+        const char* str = va_arg(args, const char*);
+        if (!str) {
+            std::cerr << "expected less arguments: " << in
+                      << ", first extra: " << arg << std::endl;
+            return false;
+        }
+        if (arg.compare(str) != 0) {
+            std::cerr << "expected " << str << " in " << in
+                      << ", got: " << arg << std::endl;
+            return false;
+        }
+    }
+    const char* str = va_arg(args, const char*);
+    if (str) {
+        std::cerr << "expected more arguments: " << in
+                  << ", missing: " << str << std::endl;
+        return false;
+    }
+    return true;
+}
+
+bool test(const char* in, ...) {
+    std::vector<std::string> out;
+    if (!Args::parse(in, &out, false)) {
+        std::cerr << "expected success: " << in << std::endl;
+        return false;
+    }
+    va_list args;
+    va_start(args, in);
+    bool ret = compare(in, out, args);
+    va_end(args);
+    return ret;
+}
+
+bool test_nice(const char* in, ...) {
+    std::vector<std::string> out;
+    if (!Args::parse(in, &out, true)) {
+        std::cerr << "expected success: " << in << std::endl;
+        return false;
+    }
+    va_list args;
+    va_start(args, in);
+    bool ret = compare(in, out, args);
+    va_end(args);
+    return ret;
+}
+
+bool test_fail(const std::string& in) {
+    std::vector<std::string> out;
+    if (Args::parse(in, &out, false)) {
+        std::cerr << "expected fail: " << in << std::endl;
+        return false;
+    }
+    return true;
+}
+
+}  // namespace
+
+int main() {
+    unsigned int ok = 0, tot = 0;
+
+    tot++; if (test("", NULL)) ok++;
+    tot++; if (test("foo", "foo", NULL)) ok++;
+    tot++; if (test("hello world", "hello", "world", NULL)) ok++;
+    tot++; if (test("\"hello world\"", "hello world", NULL)) ok++;
+    tot++; if (test("'hello world'", "hello world", NULL)) ok++;
+    tot++; if (test("a' b'", "a b", NULL)) ok++;
+    tot++; if (test("a'b'ba", "abba", NULL)) ok++;
+    tot++; if (test("a'b'b '' a", "abb", "", "a", NULL)) ok++;
+    tot++; if (test("'\"'", "\"", NULL)) ok++;
+    tot++; if (test("'\\\"'", "\"", NULL)) ok++;
+    tot++; if (test("'\\\\'", "\\", NULL)) ok++;
+    tot++; if (test("'\\''", "'", NULL)) ok++;
+    tot++; if (test("'a\\'b'", "a'b", NULL)) ok++;
+    tot++; if (test("a'\\''b", "a'b", NULL)) ok++;
+
+    tot++; if (test_nice("a'b", "a'b", NULL)) ok++;
+    tot++; if (test_nice("a'b foo", "a'b", "foo", NULL)) ok++;
+    tot++; if (test_nice("'\\", "'\\", NULL)) ok++;
+    tot++; if (test_nice("'\\'", "'\\'", NULL)) ok++;
+
+    tot++; if (test_fail("'")) ok++;
+    tot++; if (test_fail("'''")) ok++;
+    tot++; if (test_fail("\"")) ok++;
+    tot++; if (test_fail("'\\")) ok++;
+    tot++; if (test_fail("'\\'")) ok++;
+
+    std::cout << "OK " << ok << "/" << tot << std::endl;
+    return ok == tot ? EXIT_SUCCESS : EXIT_FAILURE;
+}