io: more hot-rodding for #376

* Change the interface of read-into to return a slice and an eof boolean separately so the compiler can optimize the slice.
* Add an each-stream-block-slice combinator that behaves like each-block but reuses a preallocated buffer for every iteration.
* Pull some strings in the stream-read-into implementation to further improve type propagation and bounds check elimination.

diff --git a/core/io/io-docs.factor b/core/io/io-docs.factor
index cd6422ed18e191aa3d0c5166091e4969c0bc55e4..5d0f0018facfcfcb3c89eb57e488bed4b8800587 100644 (file)
--- a/core/io/io-docs.factor
+++ b/core/io/io-docs.factor
@@ -51,13 +51,13 @@ HELP: stream-read-unsafe
 $io-error ;
 
 HELP: read-into
-{ $values { "buf" { $or byte-array specialized-array string } } { "buf-slice/f" { $or slice f } } }
-{ $contract "Reads from the current " { $link input-stream } " into the sequence " { $snippet "buf" } ", until either the length of " { $snippet "buf" } " is reached or the stream is exhausted. Returns a " { $link slice } " over the part of " { $snippet "buf" } " that was written to, or " { $link f } " if the stream was exhausted." }
+{ $values { "buf" { $or byte-array specialized-array string } } { "buf-slice" slice } { "more?" boolean } }
+{ $contract "Reads from the current " { $link input-stream } " into the sequence " { $snippet "buf" } ", until either the length of " { $snippet "buf" } " is reached or the stream is exhausted. Returns a " { $link slice } " over the part of " { $snippet "buf" } " that was written to, and a boolean value that will be " { $link f } " if the stream was exhausted." }
 $io-error ;
 
 HELP: stream-read-into
-{ $values { "buf" { $or byte-array specialized-array string } } { "stream" "an input stream" } { "buf-slice/f" { $or slice f } } }
-{ $contract "Reads from the stream into the sequence " { $snippet "buf" } ", until either the length of " { $snippet "buf" } " is reached or the stream is exhausted. Returns a " { $link slice } " over the part of " { $snippet "buf" } " that was written to, or " { $link f } " if the stream was exhausted." }
+{ $values { "buf" { $or byte-array specialized-array string } } { "stream" "an input stream" } { "buf-slice" slice } { "more?" boolean } }
+{ $contract "Reads from the stream into the sequence " { $snippet "buf" } ", until either the length of " { $snippet "buf" } " is reached or the stream is exhausted. Returns a " { $link slice } " over the part of " { $snippet "buf" } " that was written to, and a boolean value that will be " { $link f } " if the stream was exhausted." }
 { $notes "Most code only works on one stream at a time and should instead use " { $link read-into } "; see " { $link "stdio" } "." }
 $io-error ;
 
@@ -80,13 +80,13 @@ HELP: stream-read-partial-unsafe
 $io-error ;
 
 HELP: read-partial-into
-{ $values { "buf" { $or byte-array specialized-array string } } { "buf-slice/f" { $or slice f } } }
-{ $contract "Reads available data from the current " { $link input-stream } " into the sequence " { $snippet "buf" } " without blocking until all immediately available data is read or the length of " { $snippet "buf" } " is reached. If no data is immediately available, blocks until data is available. Returns a " { $link slice } " over the part of " { $snippet "buf" } " that was written to, or " { $link f } " if the stream was exhausted." }
+{ $values { "buf" { $or byte-array specialized-array string } } { "buf-slice" slice } { "more?" boolean } }
+{ $contract "Reads available data from the current " { $link input-stream } " into the sequence " { $snippet "buf" } " without blocking until all immediately available data is read or the length of " { $snippet "buf" } " is reached. If no data is immediately available, blocks until data is available. Returns a " { $link slice } " over the part of " { $snippet "buf" } " that was written to, and a boolean that will be " { $link f } " if the stream was exhausted." }
 $io-error ;
 
 HELP: stream-read-partial-into
-{ $values { "buf" { $or byte-array specialized-array string } } { "stream" "an input stream" } { "buf-slice/f" { $or slice f } } }
-{ $contract "Reads available data from the stream into the sequence " { $snippet "buf" } " without blocking until all immediately available data is read or the length of " { $snippet "buf" } " is reached. If no data is immediately available, blocks until data is available. Returns a " { $link slice } " over the part of " { $snippet "buf" } " that was written to, or " { $link f } " if the stream was exhausted." }
+{ $values { "buf" { $or byte-array specialized-array string } } { "stream" "an input stream" } { "buf-slice" slice } { "more?" boolean } }
+{ $contract "Reads available data from the stream into the sequence " { $snippet "buf" } " without blocking until all immediately available data is read or the length of " { $snippet "buf" } " is reached. If no data is immediately available, blocks until data is available. Returns a " { $link slice } " over the part of " { $snippet "buf" } " that was written to, and a boolean that will be " { $link f } " if the stream was exhausted." }
 { $notes "Most code only works on one stream at a time and should instead use " { $link read-partial-into } "; see " { $link "stdio" } "." }
 $io-error ;
 

diff --git a/core/io/io-tests.factor b/core/io/io-tests.factor
index da32cd00e3f5582eed884cb240d3702cea8ffef0..be79ff6f47e3f4751f58f61a518befbebd326fe9 100644 (file)
--- a/core/io/io-tests.factor
+++ b/core/io/io-tests.factor
@@ -27,9 +27,9 @@ M: up-to-13-reader stream-read1
 [ up-to-13-reader new [ 20 swap stream-read ] [ 20 swap stream-read ] bi ] unit-test
 
 {
-    T{ slice f 0 8 B{ 0 1 2 3 4 5 6 7 } }
-    T{ slice f 0 5 B{ 8 9 10 11 12 205 206 207 } }
-    f
+    T{ slice f 0 8 B{ 0 1 2 3 4 5 6 7 } } t
+    T{ slice f 0 5 B{ 8 9 10 11 12 205 206 207 } } t
+    T{ slice f 0 0 B{ 8 9 10 11 12 205 206 207 } } f
 } [
     up-to-13-reader new
     [ B{ 200 201 202 203 204 205 206 207 } swap stream-read-into ]

diff --git a/core/io/io.factor b/core/io/io.factor
index 8eb16b7653dd535fcb3b6c1ef5d9725445bae862..e2611fbe789ee59bfa2515794f29f7118ef3dc5a 100644 (file)
--- a/core/io/io.factor
+++ b/core/io/io.factor
@@ -129,7 +129,10 @@ SYMBOL: error-stream
 
 : (read-into) ( buf stream quot -- buf-slice/f )
     [ dup length over ] 2dip call
-    [ drop f ] [ head-slice ] if-zero ; inline
+    [ (head) <slice-unsafe> ] [ zero? not ] bi ; inline
+
+: fast>fixnum ( n -- n' )
+    dup fixnum? [ >fixnum ] unless ; inline
 
 PRIVATE>
 
@@ -141,20 +144,21 @@ PRIVATE>
 
 ERROR: invalid-read-buffer buf stream ;
 
-: stream-read-into ( buf stream -- buf-slice/f )
-    [ stream-read-unsafe ] (read-into) ; inline
+USE: kernel.private
+: stream-read-into ( buf stream -- buf-slice more? )
+    [ stream-read-unsafe { fixnum } declare ] (read-into) ; inline
 
-: stream-read-partial-into ( buf stream -- buf-slice/f )
-    [ stream-read-partial-unsafe ] (read-into) ; inline
+: stream-read-partial-into ( buf stream -- buf-slice more? )
+    [ stream-read-partial-unsafe { fixnum } declare ] (read-into) ; inline
 
 : read ( n -- seq ) input-stream get stream-read ; inline
 
 : read-partial ( n -- seq ) input-stream get stream-read-partial ; inline
 
-: read-into ( buf -- buf-slice/f )
+: read-into ( buf -- buf-slice more? )
     input-stream get stream-read-into ; inline
 
-: read-partial-into ( buf -- buf-slice/f )
+: read-partial-into ( buf -- buf-slice more? )
     input-stream get stream-read-partial-into ; inline
 
 : each-stream-line ( ... stream quot: ( ... line -- ... ) -- ... )
@@ -169,9 +173,17 @@ ERROR: invalid-read-buffer buf stream ;
 : lines ( -- seq )
     input-stream get stream-lines ; inline
 
+: each-stream-block-slice ( ... stream quot: ( ... block-slice -- ... ) -- ... )
+    [ drop ] prepose
+    swap [ 65536 swap (new-sequence-for-stream) ] keep
+    [ stream-read-partial-into ] 2curry each-morsel drop ; inline
+
 : each-stream-block ( ... stream quot: ( ... block -- ... ) -- ... )
     swap [ 65536 swap stream-read-partial ] curry each-morsel ; inline
 
+: each-block-slice ( ... quot: ( ... block -- ... ) -- ... )
+    input-stream get swap each-stream-block ; inline
+
 : each-block ( ... quot: ( ... block -- ... ) -- ... )
     input-stream get swap each-stream-block ; inline
 

diff --git a/core/sequences/sequences.factor b/core/sequences/sequences.factor
index 78fa81d0bcb970c89f256b81c64ea1b4c7b72dee..bd17c0ff7e396ec3e04a280597749aff859d5c24 100644 (file)
--- a/core/sequences/sequences.factor
+++ b/core/sequences/sequences.factor
@@ -224,6 +224,13 @@ TUPLE: slice-error from to seq reason ;
     [ drop > "start > end" slice-error ]
     3tri ; inline
 
+<PRIVATE
+
+: <slice-unsafe> ( from to seq -- slice )
+    slice boa ; inline
+
+PRIVATE>
+
 : <slice> ( from to seq -- slice )
     check-slice
     dup slice? [ collapse-slice ] when