busted in v4: process from file list

[jbreiden]

v3.0.4 works, v4.0.0 fails

$ echo phototest.tif > manifest
$ tesseract manifest - -
This is a lot of 12 point text to test the
ocr code and see if it works on all types
of file format.

The quick brown dog jumped over the
lazy fox. The quick brown dog jumped
over the lazy fox. The quick brown dog
jumped over the lazy fox. The quick
brown dog jumped over the lazy fox.

$ echo phototest.tif > manifest
$ tesseract manifest - -
Error during processing.
ObjectCache(0xb408678)::~ObjectCache(): WARNING! LEAK! object 0x7ff6a69316a0 still has count 1 (id third_party/tesseract/tessdata/eng.traineddatalstm-punc-dawg)
ObjectCache(0xb408678)::~ObjectCache(): WARNING! LEAK! object 0x7ff6a694a140 still has count 1 (id third_party/tesseract/tessdata/eng.traineddatalstm-word-dawg)
ObjectCache(0xb408678)::~ObjectCache(): WARNING! LEAK! object 0x7ff6a6dc2650 still has count 1 (id third_party/tesseract/tessdata/eng.traineddatalstm-number-dawg)
ObjectCache(0xb408678)::~ObjectCache(): WARNING! LEAK! object 0x7ff6a6dc25c0 still has count 1 (id third_party/tesseract/tessdata/eng.traineddatapunc-dawg)
ObjectCache(0xb408678)::~ObjectCache(): WARNING! LEAK! object 0x7ff6a6de0260 still has count 1 (id third_party/tesseract/tessdata/eng.traineddataword-dawg)
ObjectCache(0xb408678)::~ObjectCache(): WARNING! LEAK! object 0x7ff6a69314f0 still has count 1 (id third_party/tesseract/tessdata/eng.traineddatanumber-dawg)
ObjectCache(0xb408678)::~ObjectCache(): WARNING! LEAK! object 0x7ff6a6de1e80 still has count 1 (id third_party/tesseract/tessdata/eng.traineddatabigram-dawg)
ObjectCache(0xb408678)::~ObjectCache(): WARNING! LEAK! object 0x7ff6a6de1d60 still has count 1 (id third_party/tesseract/tessdata/eng.traineddatafreq-dawg)

[jbreiden]

Here's a fix. I don't know if 3.0.5 is busted or not (if I had to guess, yes)

--- api/baseapi.cpp	2016-12-15 15:28:37.000000000 -0800
+++ api/baseapi.cpp	2017-01-18 15:03:52.000000000 -0800
@@ -1083,7 +1083,15 @@
 
   // Maybe we have a filelist
   if (r != 0 || format == IFF_UNKNOWN) {
-    STRING s(buf.c_str());
+    STRING s;
+    if (stdInput) {
+      s = buf.c_str();
+    } else {
+      std::ifstream t(filename);
+      std::string u((std::istreambuf_iterator<char>(t)),
+                    std::istreambuf_iterator<char>());
+      s = u.c_str();
+    }
     return ProcessPagesFileList(NULL, &s, retry_config,
                                 timeout_millisec, renderer,
                                 tesseract_->tessedit_page_number);

[Shreeshrii]

Same error (WARNING! LEAK! ) reported in issue

#529 (comment)

[zdenop]

thanks!