fix: Fix the bug where an attempt is made to read data even after it has already been fully read.

[futz12]

原逻辑:

输入流: 19="1" 20="0h" 21="1"

第一次循环 (处理 19="1")

1. dr.scan("%d=", &id): 成功。

   • id 设置为 19。
   • 输入流指针移动到 19= 之后，指向 "。

2. dr.scan("%15[^,\n ]", vstr): 成功。

   • 它读取不包含空格的字符串，所以它只读取了 "1"。
   • vstr 被设置为 "1"。
   • 输入流指针移动到 "1" 之后，指向空格。

3. is_string 分支:

   • vstr_is_string(vstr) 因为 " 开头而返回 true。
   • vstr[0] == '"' 为 true。
   • 执行 nscan = dr.scan("%255[^\"\n]\"", vstr2)。
   • 这是关键点：
     • 此时的输入流是 20="0h" 21="1" (从空格开始)。
     • 格式串 [^\"\n] 会匹配任何不是 " 和 \n 的字符。
     • 因此，它会贪婪地读取 20=。
     • 当它读到 "0h" 的第一个 " 时，[^\"\n] 的匹配停止。
     • 此时，格式串的下一部分要求输入流中必须有一个 " 来作为结束符。当前输入流的字符恰好就是 "。
     • 匹配成功！
   • nscan 的值是 1。
   • vstr2 被设置为字符串 " 20="。

4. 字符串拼接:

   • 代码执行 d->params[id].s = std::string(&vstr[1]) + vstr2;
   • std::string(&vstr[1]) 的值是 "1" (去掉了开头的引号)。
   • 拼接后的结果是 "1 20="。

5. 最终处理:

   • d->params[19] 的 s 成员被设置为字符串 "1 20="。
   • 它的 type 被设置为 7。
   • 此时，输入流已经被消耗到了 "0h" 的第一个 " 之后，指针指向了 0。

---

第二次循环的尝试

1. while (dr.scan("%d=", &id) == 1):

   • 此时的输入流是 0h" 21="1"。
   • scan 函数尝试在当前位置匹配一个整数 (%d) 后面跟着一个等号 (=)。
   • 输入流的开头是 0h，这不符合 %d= 的模式。
   • scan 失败，返回 0。

2. 循环条件 dr.scan(...) == 1 为 false，整个 while 循环终止。

读取情况：
19 = 1 20=
20 =
21 =

[github-actions]

The binary size change of libncnn.so (bytes)

architecture	base size	pr size	difference
x86_64	15159560	15159560	0 😘
armhf	6182416	6186432	+4016 ⚠️
aarch64	9454784	9521008	+66224 ⚠️

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.57%. Comparing base (22aba56) to head (b0d970b).
⚠️ Report is 12 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #6337      +/-   ##
==========================================
- Coverage   95.90%   95.57%   -0.33%     
==========================================
  Files         840      840              
  Lines      265637   260958    -4679     
==========================================
- Hits       254747   249408    -5339     
- Misses      10890    11550     +660     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull Request Overview

Fixes a bug in string parsing logic where the scanner would continue reading input data even after a string had already been fully consumed. The issue occurred when a quoted string was complete (ending with a quote) but the parser would attempt to scan for additional content, causing it to consume data from subsequent parameters.

• Adds validation to check if a quoted string is already complete before scanning for additional content
• Prevents over-consumption of input data that should belong to subsequent parameters

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[tencent-adm]

Thank you for your submission, we really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
1 out of 2 committers have signed the CLA.

✅ futz12
❌ nihui
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[nihui]

Thanks for your contribution !