IO#peek now returns an empty slice on EOF. Fixes #4240

crystal-lang · Apr 5, 2017 · eae4193 · eae4193
1 parent c01a6c7
commit eae4193
Show file tree

Hide file tree

Showing 9 changed files with 53 additions and 28 deletions.
diff --git a/spec/std/io/buffered_spec.cr b/spec/std/io/buffered_spec.cr
@@ -315,8 +315,8 @@ describe "IO::Buffered" do
     # Peek doesn't advance
     io.gets_to_end.should eq("foo")
 
-    # Returns nil if no more data
-    io.peek.should be_nil
+    # Returns EOF if no more data
+    io.peek.should eq(Bytes.empty)
   end
 
   it "skips" do

diff --git a/spec/std/io/memory_spec.cr b/spec/std/io/memory_spec.cr
@@ -342,7 +342,7 @@ describe IO::Memory do
     io.peek.should eq("lo world".to_slice)
 
     io.skip_to_end
-    io.peek.should be_nil
+    io.peek.should eq(Bytes.empty)
   end
 
   it "skips" do

diff --git a/spec/std/io/sized_spec.cr b/spec/std/io/sized_spec.cr
@@ -114,7 +114,7 @@ describe "IO::Sized" do
     sized = IO::Sized.new(io, read_size: 6)
     sized.peek.should eq("123456".to_slice)
     sized.gets_to_end.should eq("123456")
-    sized.peek.should be_nil
+    sized.peek.should eq(Bytes.empty)
   end
 
   it "skips" do

diff --git a/src/http/content.cr b/src/http/content.cr
@@ -102,7 +102,7 @@ module HTTP
             peek = peek[0, @chunk_remaining]
           end
 
-          return peek.empty? ? nil : peek
+          return peek
         elsif @read_chunk_start
           read_chunk_start
           next

diff --git a/src/io.cr b/src/io.cr
@@ -392,35 +392,36 @@ module IO
   private def read_char_with_bytesize
     # For UTF-8 encoding, try to see if we can peek 4 bytes.
     # If so, this will be faster than reading byte per byte.
-    if !decoder && (peek = self.peek) && peek.size == 4
-      read_char_with_bytesize_peek(peek)
+    if !decoder && (peek = self.peek)
+      if peek.empty?
+        return nil
+      else
+        return read_char_with_bytesize_peek(peek)
+      end
     else
       read_char_with_bytesize_slow
     end
   end
 
   private def read_char_with_bytesize_peek(peek)
     first = peek[0].to_u32
+    skip(1)
     if first < 0x80
-      skip(1)
       return first.unsafe_chr, 1
     end
 
-    second = (peek[1] & 0x3f).to_u32
+    second = peek_or_read_masked(peek, 1)
     if first < 0xe0
-      skip(2)
       return ((first & 0x1f) << 6 | second).unsafe_chr, 2
     end
 
-    third = (peek[2] & 0x3f).to_u32
+    third = peek_or_read_masked(peek, 2)
     if first < 0xf0
-      skip(3)
       return ((first & 0x0f) << 12 | (second << 6) | third).unsafe_chr, 3
     end
 
-    fourth = (peek[3] & 0x3f).to_u32
+    fourth = peek_or_read_masked(peek, 3)
     if first < 0xf8
-      skip(4)
       return ((first & 0x07) << 18 | (second << 12) | (third << 6) | fourth).unsafe_chr, 4
     end
 
@@ -451,6 +452,15 @@ module IO
     (byte & 0x3f).to_u32
   end
 
+  private def peek_or_read_masked(peek, index)
+    if byte = peek[index]?
+      skip(1)
+      (byte & 0x3f).to_u32
+    else
+      read_utf8_masked_byte
+    end
+  end
+
   # Reads a single decoded UTF-8 byte from this `IO`.
   # Returns `nil` if there is no more data to read.
   #
@@ -527,9 +537,10 @@ module IO
 
   # Peeks into this IO, if possible.
   #
-  # If this IO can peek into some data, it returns a slice
-  # with that data. Returns `nil` if this IO isn't peekable,
-  # or if there's no data to peek.
+  # It returns:
+  # - `nil` if this IO isn't peekable
+  # - an empty slice if it is, but EOF was reached
+  # - a non-empty slice if some data can be peeked
   #
   # The returned bytes are only valid data until a next call
   # to any method that reads from this IO is invoked.
@@ -698,7 +709,11 @@ module IO
     # If there's no encoding, the delimiter is ASCII and we can peek,
     # use a faster algorithm
     if ascii && !decoder && (peek = self.peek)
-      gets_peek(delimiter, limit, chomp, peek)
+      if peek.empty?
+        nil
+      else
+        gets_peek(delimiter, limit, chomp, peek)
+      end
     else
       gets_slow(delimiter, limit, chomp)
     end
@@ -756,7 +771,7 @@ module IO
           peek = self.peek
         end
 
-        unless peek
+        if !peek || peek.empty?
           if buffer.bytesize == 0
             return nil
           else
@@ -786,7 +801,7 @@ module IO
     buffer = String::Builder.new
     total = 0
     while true
-      info = read_char_with_bytesize
+      info = read_char_with_bytesize_slow
       unless info
         return buffer.empty? ? nil : buffer.to_s
       end
@@ -795,7 +810,7 @@ module IO
 
       # Consider the case of \r\n when the delimiter is \n and chomp = true
       if chomp_rn && char == '\r'
-        info2 = read_char_with_bytesize
+        info2 = read_char_with_bytesize_slow
         unless info2
           buffer << char
           break

diff --git a/src/io/argf.cr b/src/io/argf.cr
@@ -36,10 +36,21 @@ class IO::ARGF
     first_initialize unless @initialized
 
     if current_io = @current_io
-      current_io.peek
-    elsif !@read_from_stdin && !@argv.empty?
+      peek = current_io.peek
+      if peek && peek.empty? # EOF
+        peek_next
+      else
+        peek
+      end
+    else
+      peek_next
+    end
+  end
+
+  private def peek_next
+    if !@read_from_stdin && !@argv.empty?
       read_next_argv
-      peek
+      self.peek
     else
       nil
     end

diff --git a/src/io/buffered.cr b/src/io/buffered.cr
@@ -82,7 +82,7 @@ module IO::Buffered
     if @in_buffer_rem.empty?
       fill_buffer
       if @in_buffer_rem.empty?
-        return nil
+        return Bytes.empty # EOF
       end
     end
 

diff --git a/src/io/memory.cr b/src/io/memory.cr
@@ -192,8 +192,7 @@ class IO::Memory
   def peek
     check_open
 
-    peek = Slice.new(@buffer + @pos, @bytesize - @pos)
-    peek.empty? ? nil : peek
+    Slice.new(@buffer + @pos, @bytesize - @pos)
   end
 
   # :nodoc:

diff --git a/src/io/sized.cr b/src/io/sized.cr
@@ -59,7 +59,7 @@ module IO
         peek = peek[0, @read_remaining]
       end
 
-      peek.empty? ? nil : peek
+      peek
     end
 
     def skip(bytes_count) : Nil