From 78b42cc4ae45ccd1409f90420082c9db646b56ff Mon Sep 17 00:00:00 2001
From: Dimitri Fontaine <dim@tapoueh.org>
Date: Fri, 22 Nov 2013 11:01:20 +0100
Subject: [PATCH] Properly handle blob formats, converting then to PostgreSQL
 bytea input.

---
 src/sources/mysql-cast-rules.lisp | 24 +++++++++++++++---------
 src/transforms.lisp               | 28 +++++++++++++++++++++++++++-
 2 files changed, 42 insertions(+), 10 deletions(-)

diff --git a/src/sources/mysql-cast-rules.lisp b/src/sources/mysql-cast-rules.lisp
index 0ba4d66..f7373f1 100644
--- a/src/sources/mysql-cast-rules.lisp
+++ b/src/sources/mysql-cast-rules.lisp
@@ -93,16 +93,22 @@
      :target (:type "varchar" :drop-typemod nil))
 
     ;;
-    ;; cl-mysql and postmodern are adapting binary values as a simple-array
-    ;; (or vector) of ‘(UNSIGNED-BYTE 8), so there should be no other
-    ;; explicit conversion to do here.
+    ;; cl-mysql returns binary values as a simple-array of bytes (as in
+    ;; ‘(UNSIGNED-BYTE 8)), that we then need to represent as proper
+    ;; PostgreSQL bytea input.
     ;;
-    (:source (:type "binary")     :target (:type "bytea"))
-    (:source (:type "varbinary")  :target (:type "bytea"))
-    (:source (:type "tinyblob")   :target (:type "bytea"))
-    (:source (:type "blob")       :target (:type "bytea"))
-    (:source (:type "mediumblob") :target (:type "bytea"))
-    (:source (:type "longblob")   :target (:type "bytea"))
+    (:source (:type "binary") :target (:type "bytea")
+	     :using pgloader.transforms::byte-vector-to-bytea)
+    (:source (:type "varbinary")  :target (:type "bytea")
+	     :using pgloader.transforms::byte-vector-to-bytea)
+    (:source (:type "tinyblob")   :target (:type "bytea")
+	     :using pgloader.transforms::byte-vector-to-bytea)
+    (:source (:type "blob")       :target (:type "bytea")
+	     :using pgloader.transforms::byte-vector-to-bytea)
+    (:source (:type "mediumblob") :target (:type "bytea")
+	     :using pgloader.transforms::byte-vector-to-bytea)
+    (:source (:type "longblob")   :target (:type "bytea")
+	     :using pgloader.transforms::byte-vector-to-bytea)
 
     (:source (:type "datetime" :default "0000-00-00 00:00:00" :not-null t)
      :target (:type "timestamptz" :drop-default t :drop-not-null t)
diff --git a/src/transforms.lisp b/src/transforms.lisp
index c1cec4d..a1e6e8b 100644
--- a/src/transforms.lisp
+++ b/src/transforms.lisp
@@ -21,7 +21,8 @@
 		 convert-mysql-point
 		 float-to-string
 		 set-to-enum-array
-		 right-trim))
+		 right-trim
+		 byte-vector-to-bytea))
 
 
 ;;;
@@ -143,3 +144,28 @@
   "Remove whitespaces at end of STRING."
   (declare (type simple-string string))
   (string-right-trim '(#\Space) string))
+
+(defun byte-vector-to-bytea (vector)
+  "Transform a simple array of unsigned bytes to the PostgreSQL bytea
+  representation as documented at
+  http://www.postgresql.org/docs/9.3/interactive/datatype-binary.html
+
+  Note that we choose here the bytea Hex Format."
+  (declare (type simple-array vector))
+  (let ((hex-digits "0123456789abcdef")
+	(bytea (make-array (+ 2 (* 2 (length vector)))
+			   :initial-element #\0
+			   :element-type 'standard-char)))
+
+    ;; The entire string is preceded by the sequence \x (to distinguish it
+    ;; from the escape format).
+    (setf (aref bytea 0) #\\)
+    (setf (aref bytea 1) #\x)
+
+    (loop for pos from 2 by 2
+       for byte across vector
+       do (let ((high (ldb (byte 4 4) byte))
+		(low  (ldb (byte 4 0) byte)))
+	    (setf (aref bytea pos)       (aref hex-digits high))
+	    (setf (aref bytea (+ pos 1)) (aref hex-digits low)))
+       finally (return bytea))))