From 245eebdf293ac19f1fb85c36e51daaaa1b5e5a3e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zdenko=20Podobn=C3=BD?= <zdenop@gmail.com>
Date: Wed, 7 Dec 2016 21:40:51 +0100
Subject: [PATCH] Multi-page TIFF buffering is broken - fix #233

---
 api/baseapi.cpp            | 6 ++++--
 classify/mastertrainer.cpp | 6 +++++-
 2 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/api/baseapi.cpp b/api/baseapi.cpp
index 582fbfbb..4104b93d 100644
--- a/api/baseapi.cpp
+++ b/api/baseapi.cpp
@@ -1028,6 +1028,7 @@ bool TessBaseAPI::ProcessPagesMultipageTiff(const l_uint8 *data,
   OpenclDevice od;
 #endif  // USE_OPENCL
   int page = (tessedit_page_number >= 0) ? tessedit_page_number : 0;
+  size_t offset = 0;
   for (; ; ++page) {
     if (tessedit_page_number >= 0)
       page = tessedit_page_number;
@@ -1039,8 +1040,8 @@ bool TessBaseAPI::ProcessPagesMultipageTiff(const l_uint8 *data,
     } else {
 #endif  // USE_OPENCL
       pix = (data) ?
-          pixReadMemTiff(data, size, page) :
-          pixReadTiff(filename, page);
+          pixReadMemFromMultipageTiff(data, size, &offset) :
+          pixReadFromMultipageTiff(filename, &offset);
 #ifdef USE_OPENCL
     }
 #endif  // USE_OPENCL
@@ -1054,6 +1055,7 @@ bool TessBaseAPI::ProcessPagesMultipageTiff(const l_uint8 *data,
     pixDestroy(&pix);
     if (!r) return false;
     if (tessedit_page_number >= 0) break;
+    if (!offset) break;
   }
   return true;
 #else
diff --git a/classify/mastertrainer.cpp b/classify/mastertrainer.cpp
index cd7e93b9..849fb060 100644
--- a/classify/mastertrainer.cpp
+++ b/classify/mastertrainer.cpp
@@ -214,10 +214,14 @@ void MasterTrainer::AddSample(bool verification, const char* unichar,
 // Must be called after ReadTrainingSamples, as the current number of images
 // is used as an offset for page numbers in the samples.
 void MasterTrainer::LoadPageImages(const char* filename) {
+  size_t offset = 0;
   int page;
   Pix* pix;
-  for (page = 0; (pix = pixReadTiff(filename, page)) != NULL; ++page) {
+  for (page = 0; ; page++) {
+    pix = pixReadFromMultipageTiff(filename, &offset);
+    if (!pix) break;
     page_images_.push_back(pix);
+    if (!offset) break;
   }
   tprintf("Loaded %d page images from %s\n", page, filename);
 }