[libcxx-commits] [libcxx] [libc++] Refactor basic_filebuf::overflow() (PR #144793)

[Louis Dionne via libcxx-commits]

https://github.com/ldionne created https://github.com/llvm/llvm-project/pull/144793

Refactor the function to streamline the logic so it matches the specification in [filebuf.virtuals] more closely. In particular, avoid modifying the put area pointers when we loop around after a partial codecvt conversion.

Note that we're technically not up-to-spec in this implementation, since the Standard says that we shouldn't try more than once after a partial codecvt conversion. However, this refactoring attempts not to change any functionality.

>From 73a7634f261d63e018ae338f53d7a308c2bc7006 Mon Sep 17 00:00:00 2001
From: Louis Dionne <ldionne.2 at gmail.com>
Date: Tue, 17 Jun 2025 12:23:18 -0400
Subject: [PATCH] [libc++] Refactor basic_filebuf::overflow()

Refactor the function to streamline the logic so it matches the specification
in [filebuf.virtuals] more closely. In particular, avoid modifying the put
area pointers when we loop around after a partial codecvt conversion.

Note that we're technically not up-to-spec in this implementation, since
the Standard says that we shouldn't try more than once after a partial
codecvt conversion. However, this refactoring attempts not to change any
functionality.
---
 libcxx/include/fstream | 57 ++++++++++++++++++++++++++----------------
 1 file changed, 36 insertions(+), 21 deletions(-)

diff --git a/libcxx/include/fstream b/libcxx/include/fstream
index 00aa00ff7e9cd..f9cbac2ab0578 100644
--- a/libcxx/include/fstream
+++ b/libcxx/include/fstream
@@ -835,35 +835,50 @@ typename basic_filebuf<_CharT, _Traits>::int_type basic_filebuf<_CharT, _Traits>
   }
   if (this->pptr() != this->pbase()) {
     if (__always_noconv_) {
-      size_t __nmemb = static_cast<size_t>(this->pptr() - this->pbase());
-      if (std::fwrite(this->pbase(), sizeof(char_type), __nmemb, __file_) != __nmemb)
+      size_t __n = static_cast<size_t>(this->pptr() - this->pbase());
+      if (std::fwrite(this->pbase(), sizeof(char_type), __n, __file_) != __n)
         return traits_type::eof();
     } else {
-      char* __extbe = __extbuf_;
-      codecvt_base::result __r;
+      if (!__cv_)
+        std::__throw_bad_cast();
+
+      // See [filebuf.virtuals]
+      char_type* __b = this->pbase();
+      char_type* __p = this->pptr();
+      const char_type* __end;
+      char* __extbuf_end = __extbuf_;
       do {
-        if (!__cv_)
-          std::__throw_bad_cast();
-
-        const char_type* __e;
-        __r = __cv_->out(__st_, this->pbase(), this->pptr(), __e, __extbuf_, __extbuf_ + __ebs_, __extbe);
-        if (__e == this->pbase())
+        codecvt_base::result __r = __cv_->out(__st_, __b, __p, __end, __extbuf_, __extbuf_ + __ebs_, __extbuf_end);
+        if (__end == __b)
           return traits_type::eof();
+
+        // No conversion needed: output characters directly to the file, done.
         if (__r == codecvt_base::noconv) {
-          size_t __nmemb = static_cast<size_t>(this->pptr() - this->pbase());
-          if (std::fwrite(this->pbase(), 1, __nmemb, __file_) != __nmemb)
+          size_t __n = static_cast<size_t>(__p - __b);
+          if (std::fwrite(__b, 1, __n, __file_) != __n)
             return traits_type::eof();
-        } else if (__r == codecvt_base::ok || __r == codecvt_base::partial) {
-          size_t __nmemb = static_cast<size_t>(__extbe - __extbuf_);
-          if (std::fwrite(__extbuf_, 1, __nmemb, __file_) != __nmemb)
+          break;
+
+          // Conversion successful: output the converted characters to the file, done.
+        } else if (__r == codecvt_base::ok) {
+          size_t __n = static_cast<size_t>(__extbuf_end - __extbuf_);
+          if (std::fwrite(__extbuf_, 1, __n, __file_) != __n)
+            return traits_type::eof();
+          break;
+
+          // Conversion partially successful: output converted characters to the file and repeat with the
+          // remaining characters.
+        } else if (__r == codecvt_base::partial) {
+          size_t __n = static_cast<size_t>(__extbuf_end - __extbuf_);
+          if (std::fwrite(__extbuf_, 1, __n, __file_) != __n)
             return traits_type::eof();
-          if (__r == codecvt_base::partial) {
-            this->setp(const_cast<char_type*>(__e), this->pptr());
-            this->__pbump(this->epptr() - this->pbase());
-          }
-        } else
+          __b = const_cast<char_type*>(__end);
+          continue;
+
+        } else {
           return traits_type::eof();
-      } while (__r == codecvt_base::partial);
+        }
+      } while (true);
     }
     this->setp(__pb_save, __epb_save);
   }