picohttp 0.3.1 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 664158c1afacf094df70190d2b86e9a132e8ad4c5326bf57d26e8ba31050caab
-  data.tar.gz: 4db750f953e50493e2dbbc87e26619aa4078d125ad2f45ca567562efaac37176
+  metadata.gz: 32830c406cc611f0808d7b360b22294b03a7a3a76cec684175e6bd058c2d2d6e
+  data.tar.gz: f929ecc68df200572fd38506cf00db14cc8b80baaa5ad01a1d51a5b5f11f2805
 SHA512:
-  metadata.gz: 6379e766625fcd567fd47aaa7387514d8e7447b85488876e7dea3365fda1cda2c314e744885c44ab425e8947a8695178a35eb8b907701c3a7b85c357f82245d1
-  data.tar.gz: 9505e4c27f13c9b0253c4b910013157fdf945700d7e0f0b34e8d4c1ee0450b223508d6e060187d93c380017be153703eef9cd9cf77ec40686075246b7f6dd3ef
+  metadata.gz: f0cc7ecee0bcbb834c2a44ac85c8a33e66278e87efee338fa91c9d392b07337be3d1c7feb2f2b98b015f2bbae574022136702a98b266474b68c1263373520940
+  data.tar.gz: 2bbdee0a6887a7fb0d068f80d61e934edb3fd703e67e75f4dbf0fbb271b9f3ffa03a55f3b0995f28e5316daaf37a31f0eb16bc84d88b8149429fb7927d6781df

data/README.md

@@ -34,6 +34,49 @@ env = Picohttp.parse_request_env(request)
 
 Returns `nil` for incomplete requests, raises `Picohttp::ParseError` for invalid ones.
 
+### Environment fields
+
+The produced hash uses CGI/Rack-style keys:
+
+| Key               | Value                                                                                                | Set when |
+| ----------------- | --------------------------------------------------------------------------------------------------- | -------- |
+| `REQUEST_METHOD`  | The request method, e.g. `"GET"`, `"POST"`.                                                          | always   |
+| `SCRIPT_NAME`     | Always `""`.                                                                                         | always   |
+| `PATH_INFO`       | The request path with any query string removed, e.g. `"/api/users"`.                                | always   |
+| `QUERY_STRING`    | The part of the target after `?`, or `""` when there is none.                                        | always   |
+| `REQUEST_URI`     | The full request target, e.g. `"/api/users?limit=10"`.                                               | always   |
+| `SERVER_PROTOCOL` | `"HTTP/1.0"` or `"HTTP/1.1"`.                                                                        | always   |
+| `SERVER_NAME`     | The host portion of the `Host` header, with brackets kept for IPv6 literals (`"[::1]"`).             | `Host` present |
+| `SERVER_PORT`     | The port from the `Host` header, e.g. `"8080"`.                                                      | `Host` has a port |
+| `CONTENT_TYPE`    | The `Content-Type` header (note: no `HTTP_` prefix).                                                 | header present |
+| `CONTENT_LENGTH`  | The `Content-Length` header (note: no `HTTP_` prefix).                                               | header present |
+
+Every other request header becomes an `HTTP_`-prefixed key: the name is
+uppercased and `-` becomes `_` (e.g. `User-Agent` → `HTTP_USER_AGENT`). Repeated
+headers are joined with `, `. A repeated `Host` header raises
+`Picohttp::ParseError` (RFC 7230 §5.4).
+
+### Starting from a template
+
+Servers usually add the same constant keys to every env. `parse_request_env_with_template` starts from a copy of a template hash instead, which is faster than merging the constants in afterwards:
+
+```ruby
+RACK_ENV_CONST = {
+  "rack.url_scheme" => "http",
+  "SERVER_SOFTWARE" => "MyServer/1.0",
+}.freeze
+
+env = Picohttp.parse_request_env_with_template(request, RACK_ENV_CONST)
+# => {
+#      "rack.url_scheme" => "http",
+#      "SERVER_SOFTWARE" => "MyServer/1.0",
+#      "REQUEST_METHOD" => "GET",
+#      ...
+#    }
+```
+
+The template is never modified and may be frozen. If a key is in both the template and the request, the request wins, but this isn't recommended.
+
 ## Development
 
 After checking out the repo, run `bin/setup` to install dependencies. Then, run `rake test` to run the tests. You can also run `bin/console` for an interactive prompt that will allow you to experiment.

data/ext/picohttp/picohttp.c

@@ -132,8 +132,8 @@ build_hash_with_combined_duplicates(VALUE *header_values, int count)
     return env;
 }
 
-static VALUE
-picohttp_parse_request_env(VALUE self, VALUE str)
+static int
+parse_request_env_pairs(VALUE str, VALUE *header_values)
 {
     Check_Type(str, T_STRING);
 
@@ -150,12 +150,11 @@ picohttp_parse_request_env(VALUE self, VALUE str)
 
     if (result < 0) {
         if (result == -2) {
-            return Qnil; // Incomplete request
+            return -1; // Incomplete request
         }
         rb_raise(rb_ePicohttpParseError, "Invalid HTTP request");
     }
 
-    VALUE header_values[(MAX_HTTP_HEADERS + EXTRA_RACK_HEADERS) * 2];
     int idx = 0;
 
     // Standard CGI/Rack environment variables
@@ -193,6 +192,7 @@ picohttp_parse_request_env(VALUE self, VALUE str)
     header_values[idx++] = rb_str_empty;
 
     // Convert headers to HTTP_ prefixed environment variables
+    int host_seen = 0;
     for (size_t i = 0; i < num_headers; i++) {
         if (headers[i].name == NULL) {
             rb_raise(rb_ePicohttpParseError, "HTTP line folding not supported");
@@ -201,21 +201,37 @@ picohttp_parse_request_env(VALUE self, VALUE str)
         header_values[idx++] = header_name_to_env_key(headers[i].name, headers[i].name_len);
         header_values[idx++] = rb_str_new(headers[i].value, headers[i].value_len);
 
-        // Extract SERVER_NAME/SERVER_PORT from Host header
         if (headers[i].name_len == 4 &&
             (headers[i].name[0] | 0x20) == 'h' &&
             (headers[i].name[1] | 0x20) == 'o' &&
             (headers[i].name[2] | 0x20) == 's' &&
             (headers[i].name[3] | 0x20) == 't') {
+            // More than one Host is a 400 (RFC 7230 5.4). Rejecting also keeps
+            // idx bounded: the SERVER_* extras are written at most once.
+            if (host_seen) {
+                rb_raise(rb_ePicohttpParseError, "Duplicate Host header");
+            }
+            host_seen = 1;
+
             const char *host = headers[i].value;
             size_t host_len = headers[i].value_len;
-            const char *colon = memchr(host, ':', host_len);
+            const char *host_end = host + host_len;
+
+            // IPv6 authority is "[::1]" or "[::1]:port"; the colons inside the
+            // brackets aren't the port separator, so scan only after the ']'.
+            const char *colon;
+            if (host_len > 0 && host[0] == '[') {
+                const char *bracket = memchr(host, ']', host_len);
+                colon = bracket ? memchr(bracket, ':', host_end - bracket) : NULL;
+            } else {
+                colon = memchr(host, ':', host_len);
+            }
 
             if (colon) {
                 header_values[idx++] = rb_str_server_name;
                 header_values[idx++] = rb_str_new(host, colon - host);
                 header_values[idx++] = rb_str_server_port;
-                header_values[idx++] = rb_str_new(colon + 1, host_len - (colon - host) - 1);
+                header_values[idx++] = rb_str_new(colon + 1, host_end - colon - 1);
             } else {
                 header_values[idx++] = rb_str_server_name;
                 header_values[idx++] = rb_str_new(host, host_len);
@@ -223,8 +239,19 @@ picohttp_parse_request_env(VALUE self, VALUE str)
         }
     }
 
+    return idx;
+}
+
+static VALUE
+picohttp_parse_request_env(VALUE self, VALUE str)
+{
+    VALUE header_values[(MAX_HTTP_HEADERS + EXTRA_RACK_HEADERS) * 2];
+    int idx = parse_request_env_pairs(str, header_values);
+    if (idx < 0) return Qnil;
+
 #ifdef HAVE_RB_HASH_NEW_CAPA
-    VALUE env = rb_hash_new_capa(idx / 2);
+    // Extra slack so servers can add their own env keys without a rehash
+    VALUE env = rb_hash_new_capa(idx / 2 + 8);
 #else
     VALUE env = rb_hash_new();
 #endif
@@ -239,6 +266,29 @@ picohttp_parse_request_env(VALUE self, VALUE str)
     return env;
 }
 
+static VALUE
+picohttp_parse_request_env_with_template(VALUE self, VALUE str, VALUE template)
+{
+    Check_Type(template, T_HASH);
+
+    VALUE header_values[(MAX_HTTP_HEADERS + EXTRA_RACK_HEADERS) * 2];
+    int idx = parse_request_env_pairs(str, header_values);
+    if (idx < 0) return Qnil;
+
+    VALUE env = rb_hash_dup(template);
+    size_t expected_size = RHASH_SIZE(env) + idx / 2;
+
+    rb_hash_bulk_insert(idx, header_values, env);
+
+    // Handle duplicate headers per RFC 7230
+    if (RHASH_SIZE(env) != expected_size) {
+        env = rb_hash_dup(template);
+        rb_hash_update_by(env, build_hash_with_combined_duplicates(header_values, idx), NULL);
+    }
+
+    return env;
+}
+
 static VALUE
 register_interned_string(const char *str)
 {
@@ -258,6 +308,7 @@ Init_picohttp(void)
     rb_ePicohttpParseError = rb_define_class_under(rb_mPicohttp, "ParseError", rb_eStandardError);
     rb_define_module_function(rb_mPicohttp, "parse_request", picohttp_parse_request, 1);
     rb_define_module_function(rb_mPicohttp, "parse_request_env", picohttp_parse_request_env, 1);
+    rb_define_module_function(rb_mPicohttp, "parse_request_env_with_template", picohttp_parse_request_env_with_template, 2);
 
     // Initialize interned string constants
     init_string_lookup();

data/lib/picohttp/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Picohttp
-  VERSION = "0.3.1"
+  VERSION = "0.4.0"
 end

metadata

@@ -1,60 +1,61 @@
 --- !ruby/object:Gem::Specification
 name: picohttp
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.4.0
 platform: ruby
 authors:
-- John Hawthorn
+  - John Hawthorn
 bindir: exe
 cert_chain: []
 date: 1980-01-02 00:00:00.000000000 Z
 dependencies: []
-description: Wraps picohttpparser to parse HTTP request strings into Rack-compatible
-  hashes
+description: Wraps picohttpparser to parse HTTP request strings into Rack-compatible hashes
 email:
-- john@hawthorn.email
+  - john@hawthorn.email
 executables: []
 extensions:
-- ext/picohttp/extconf.rb
+  - ext/picohttp/extconf.rb
 extra_rdoc_files: []
 files:
-- CODE_OF_CONDUCT.md
-- LICENSE.txt
-- README.md
-- Rakefile
-- ext/picohttp/extconf.rb
-- ext/picohttp/picohttp.c
-- ext/picohttp/picohttp.h
-- ext/picohttp/picohttpparser.c
-- ext/picohttp/picohttpparser.h
-- ext/picohttp/string_lookup.inc
-- lib/picohttp.rb
-- lib/picohttp/version.rb
-- tool/generate_lookup.rb
-homepage: https://github.com/jhawthorn/picohttp
+  - CODE_OF_CONDUCT.md
+  - LICENSE.txt
+  - README.md
+  - Rakefile
+  - ext/picohttp/extconf.rb
+  - ext/picohttp/picohttp.c
+  - ext/picohttp/picohttp.h
+  - ext/picohttp/picohttpparser.c
+  - ext/picohttp/picohttpparser.h
+  - ext/picohttp/string_lookup.inc
+  - lib/picohttp.rb
+  - lib/picohttp/version.rb
+  - tool/generate_lookup.rb
+homepage: "https://github.com/jhawthorn/picohttp"
 licenses:
-- MIT
+  - MIT
 metadata:
-  homepage_uri: https://github.com/jhawthorn/picohttp
-  source_code_uri: https://github.com/jhawthorn/picohttp
-  changelog_uri: https://github.com/jhawthorn/picohttp/blob/main/CHANGELOG.md
-  bug_tracker_uri: https://github.com/jhawthorn/picohttp/issues
-  documentation_uri: https://github.com/jhawthorn/picohttp#readme
+  homepage_uri: "https://github.com/jhawthorn/picohttp"
+  source_code_uri: "https://github.com/jhawthorn/picohttp"
+  changelog_uri: "https://github.com/jhawthorn/picohttp/blob/main/CHANGELOG.md"
+  bug_tracker_uri: "https://github.com/jhawthorn/picohttp/issues"
+  documentation_uri: "https://github.com/jhawthorn/picohttp#readme"
 rdoc_options: []
 require_paths:
-- lib
+  - lib
 required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ">="
-    - !ruby/object:Gem::Version
-      version: 3.2.0
+    -
+      - ">="
+      - !ruby/object:Gem::Version
+        version: 3.2.0
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ">="
-    - !ruby/object:Gem::Version
-      version: '0'
+    -
+      - ">="
+      - !ruby/object:Gem::Version
+        version: "0"
 requirements: []
-rubygems_version: 4.0.3
+rubygems_version: 4.1.0.beta1
 specification_version: 4
 summary: Fast HTTP request parser using picohttpparser
 test_files: []