ctx_parser 0.2.5 → 0.2.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: f5bf7a928b6287113d2db138fbd46ad11c22932b15ef99473caeedee6517dd07
-  data.tar.gz: 5acef4c71f501bd931309afec028334831b86bafe67fd3e03411887bfdb401df
+  metadata.gz: 488dca1bfd052afe8a7719c6e5c89d4b2b5c86f189243c02b651438ce45a69c1
+  data.tar.gz: 33e7aa374cc8285c6dbb76ff4c9d87849116c3fa1a61ddb1ee720d108baed8ee
 SHA512:
-  metadata.gz: 0ff61300ef0fc301eb315bf5d0e667534e996884ae6e83b2c08c0a630bf5f55da4a2faf2c7f97b23cba0520a71807c6393bb6a3fb8350b406632fd1930b1732d
-  data.tar.gz: 180f90050ea5f2794d638eb9e64efc680b1ce0d07c786318eba5c22d2b48c6177c26156678777c4ae3e5e33fe5b3611c9cb7b8d2e1fd48f1e66be9c7c7020694
+  metadata.gz: 8459a636b16bdbe34e4446530714b1f686daa9bab80ca0cb713fdff9422066a388a7410dae7d083126c92415c8cb97f479a321da91d6113feb947cd03158bff6
+  data.tar.gz: ae3de9681c49c1d4be422c3b4b7c1fc47aeb0c152cb740370846fc3f80122435b184de6f2c977c3dac06d848c04a2c2d7191ad54d286214eee027c67d0b48abc

data/CHANGELOG.md

@@ -4,6 +4,14 @@ Changes to the public contract are recorded here. Versions follow Semantic Versi
 
 ## Unreleased
 
+## 0.2.6 (2026-10-21)
+
+- Build immutable hash snapshots without temporary field-pair arrays, retaining ordinary key equality, cycle rejection and deep-tree support.
+
+- Add repository-specific AGENTS.md guidance for responsibilities, Ruby conventions, evidence boundaries and validation.
+- Snapshot requested profile views before parser callbacks can mutate caller selections.
+- Avoid redundant UTF-8 conversion and line-ending replacement for already normalized inputs while preserving mutable detached return values.
+
 ## 0.2.5 (2026-10-19)
 
 - Keep offline example facts and configuration benchmark inputs independent of Ruby's global input and output separators.

data/docs/input-performance.md

@@ -0,0 +1,19 @@
+# 输入归一化验证（2026-10-21）
+
+`Input.text` 的常见路径原先会对 UTF-8 输入再次执行 `encode(UTF_8)`，并在 LF 正文上执行无匹配的 `gsub`。现在跳过相同编码的转换，在没有 CR 时跳过替换；需要修改时只操作归一化过程拥有的副本。返回值仍可变且独立于调用方；原文、编码策略、逻辑 NUL 校验和转码前字节预算不变。
+
+在 macOS arm64、Ruby 4.0.6 上使用 `benchmark/input.rb`，对修改前的 `b8746d9` 与本次源码执行相同脚本。每轮重复归一化同一正文 30 次，共三轮；语义和对象所有权断言在计时外执行。下表是三轮中位数，属于重复输入的本地微基准，不代表设备采集、完整配置解析或网络吞吐。
+
+| 正文 | 输入字节 | 修改前耗时 | 修改后耗时 | 修改前分配 | 修改后分配 |
+| --- | ---: | ---: | ---: | ---: | ---: |
+| UTF-8 / LF | 950000 | 67.354 ms | 0.866 ms | 120 | 30 |
+| UTF-8 / CRLF | 1000000 | 177.798 ms | 162.056 ms | 150 | 150 |
+| ISO-8859-1 / LF | 850000 | 169.909 ms | 90.748 ms | 150 | 60 |
+
+复测当前代码：
+
+```sh
+bundle exec ruby -Ilib benchmark/input.rb
+```
+
+比较基线时，把目标提交的 `lib/ctx_parser/input.rb` 放入独立临时目录的 `ctx_parser/input.rb`，然后使用 `bundle exec ruby -I<临时目录> benchmark/input.rb`。脚本输出实际加载的 Input 来源，避免误测已安装版本。不要覆盖工作区文件来切换基线。

data/lib/ctx_parser/immutable.rb

@@ -28,7 +28,9 @@ module CtxParser
           return value
         end
         copies[value] = if value.is_a?(Hash)
-          value.to_h { |key, item| [ copy_value(key, copies, depth + 1), copy_value(item, copies, depth + 1) ] }.freeze
+          result = {}
+          value.each_pair { |key, item| result[copy_value(key, copies, depth + 1)] = copy_value(item, copies, depth + 1) }
+          result.freeze
         else
           value.map { |item| copy_value(item, copies, depth + 1) }.freeze
         end
@@ -69,7 +71,9 @@ module CtxParser
           copies[source] = if reusable?(source, copies, 0)
             source
           elsif source.is_a?(Hash)
-            source.to_h { |key, item| [ snapshot(key, copies), snapshot(item, copies) ] }.freeze
+            result = {}
+            source.each_pair { |key, item| result[snapshot(key, copies)] = snapshot(item, copies) }
+            result.freeze
           else
             source.map { |item| snapshot(item, copies) }.freeze
           end

data/lib/ctx_parser/input.rb

@@ -18,14 +18,18 @@ module CtxParser
       value.force_encoding(Encoding::UTF_8) if encoding == :utf8 || value.encoding == Encoding::BINARY
       raise EncodingError, "CLI output encoding is invalid" unless value.valid_encoding?
 
-      value = value.encode(Encoding::UTF_8) if encoding == :transcode
+      value = value.encode(Encoding::UTF_8) if encoding == :transcode && value.encoding != Encoding::UTF_8
       # UTF-16/32 等编码不兼容 ASCII 字面量；仅在临时视图上检查逻辑字符。
       logical = value.encoding.ascii_compatible? ? value : value.encode(Encoding::UTF_8)
       raise ArgumentError, "CLI output contains a NUL byte" if !allow_nul && logical.include?("\0")
 
       normalized = case newlines
-      when :lf then logical.gsub(/\r\n?/, "\n")
-      when :delete_cr then logical.delete("\r")
+      when :lf
+        logical.gsub!(/\r\n?/, "\n") if logical.include?("\r")
+        logical
+      when :delete_cr
+        logical.delete!("\r")
+        logical
       else return value
       end
       normalized.encoding == value.encoding ? normalized : normalized.encode(value.encoding)

data/lib/ctx_parser/profile.rb

@@ -102,6 +102,8 @@ module CtxParser
       unless dimensions.is_a?(Array) && dimensions.any? && dimensions.uniq == dimensions
         raise ArgumentError, "Parser dimensions must be a nonempty unique array"
       end
+      # 扩展解析器可调用外部代码；结果投影始终使用入口处的请求。
+      dimensions = dimensions.dup.freeze
       query = select_command(dimensions, command)
 
       # 查询已覆盖全部请求维度，无需再次遍历第一个维度的完整命令列表。

data/lib/ctx_parser/version.rb

@@ -2,5 +2,5 @@
 
 module CtxParser
   # 公开包版本，供 RubyGems 元数据和运行时调用方共用。
-  VERSION = "0.2.5"
+  VERSION = "0.2.6"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: ctx_parser
 version: !ruby/object:Gem::Version
-  version: 0.2.5
+  version: 0.2.6
 platform: ruby
 authors:
 - Gatework contributors
@@ -48,6 +48,7 @@ files:
 - README-zh.md
 - README.md
 - docs/architecture-zh.md
+- docs/input-performance.md
 - docs/release.md
 - examples/offline.rb
 - lib/ctx_parser.rb