From fa9cf8c758d9dc8b3841edd813be68b8734642ac Mon Sep 17 00:00:00 2001
From: darioleanbit <62099044+darioleanbit@users.noreply.github.com>
Date: Wed, 15 Mar 2023 14:33:08 +0100
Subject: [PATCH] SQLite: check for base64 for text content in blob

If the SQLite driver returns the value of a blob column a string, don't make the assumption that the value is base64-encoded. Instead, check for actual base64 content by trying to decode the string as base64. See https://github.com/dimitri/pgloader/issues/415 .
---
 src/sources/sqlite/sqlite.lisp | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)

diff --git a/src/sources/sqlite/sqlite.lisp b/src/sources/sqlite/sqlite.lisp
index 345dafb3..0798a174 100644
--- a/src/sources/sqlite/sqlite.lisp
+++ b/src/sources/sqlite/sqlite.lisp
@@ -4,6 +4,17 @@
 
 (in-package :pgloader.source.sqlite)
 
+(in-package :pgloader.source.sqlite)
+
+(declaim (inline has-base64-sequence-p))
+
+;;; Return the decoded value of a base64 string, ignoring errors.
+;;; Returns nil if the string doesn't contain a valid base64 string.
+;;;
+(defun has-base64-value-p (string)
+  (ignore-errors
+    (base64:base64-string-to-string string)))
+
 ;;; Map a function to each row extracted from SQLite
 ;;;
 (declaim (inline parse-value))
@@ -20,10 +31,16 @@
          (babel:octets-to-string value :encoding encoding))
 
         ((and (string-equal "bytea" pgsql-type)
+              (has-base64-value-p value)
               (stringp value))
-         ;; we expected bytes and got a string instead, must be base64 encoded
+         ;; we expected bytes and got a be base64 encoded, string instead
          (base64:base64-string-to-usb8-array value))
 
+        ((and (string-equal "bytea" pgsql-type)
+              (stringp value))
+         ;; we expected bytes and got a string instead, convert it to octets
+         (babel:string-to-octets value :encoding :utf-8))
+
         ;; default case, just use what's been given to us
         (t value)))