Dear Tesseract developers,
The source code in mainblk.c reads:
* TESSDATA_PREFIX Environment variable overrules everything.
* Compiled in -DTESSDATA_PREFIX is next.
* An actual value of argv0 is used if not NULL, otherwise current directory.
But actually it is:
* TESSDATA_PREFIX Environment variable overrules everything.
* Compiled in -DTESSDATA_PREFIX overrides argv0.
* argv0 is used if not NULL, otherwise current directory.
I think that things should be all way around: more specific
(application-specific) values should override system-wide settings. In
particular:
* If argv0 is not NULL, it is used.
* Otherwise if TESSDATA_PREFIX environment variable is defined, it is used.
* Otherwise (if defined) compiled-in TESSDATA_PREFIX is used.
* Otherwise current directory is used.
I have problem when trying to use system-wide library
(/usr/lib/libtesseract.so.3) with my trained datasets in multi-threading
environment: obviously setting TESSDATA_PREFIX env does not work
reliably and the only correct way is to pass directory to
TessBaseAPI::Init(), but it is ignored.
The attached patch solves this for me, but it also makes sure that
despite of source (argument, environment, pre-compiled) directory is
post-processed the same way.
--
With best regards,
Dmitry
--
--
You received this message because you are subscribed to the Google
Groups "tesseract-ocr" group.
To post to this group, send email to tesseract-ocr@googlegroups.com
To unsubscribe from this group, send email to
tesseract-ocr+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/tesseract-ocr?hl=en
---
You received this message because you are subscribed to the Google Groups
"tesseract-ocr" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to tesseract-ocr+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
--- ./ccutil/mainblk.cpp.orig 2012-10-26 19:03:06.000000000 +0200
+++ ./ccutil/mainblk.cpp 2013-06-06 20:46:27.000000000 +0200
@@ -51,30 +51,34 @@
// TESSDATA_PREFIX Environment variable overrules everything.
// Compiled in -DTESSDATA_PREFIX is next.
// An actual value of argv0 is used if not NULL, otherwise current directory.
- if (!getenv("TESSDATA_PREFIX")) {
+ if (argv0 != NULL) {
+ datadir = argv0;
+ } else {
+ if (getenv("TESSDATA_PREFIX")) {
+ datadir = getenv("TESSDATA_PREFIX");
+ } else {
#ifdef TESSDATA_PREFIX
#define _STR(a) #a
#define _XSTR(a) _STR(a)
- datadir = _XSTR(TESSDATA_PREFIX);
+ datadir = _XSTR(TESSDATA_PREFIX);
#undef _XSTR
#undef _STR
#else
- if (argv0 != NULL) {
- datadir = argv0;
- // Remove tessdata from the end if present, as we will add it back!
- int length = datadir.length();
- if (length >= 8 && strcmp(&datadir[length - 8], "tessdata") == 0)
- datadir.truncate_at(length - 8);
- else if (length >= 9 && strcmp(&datadir[length - 9], "tessdata/") == 0)
- datadir.truncate_at(length - 9);
- if (datadir.length() == 0)
- datadir = "./";
- } else {
datadir = "./";
- }
#endif
+ }
+ }
+
+ // datadir may still be empty:
+ if (datadir.length() == 0) {
+ datadir = "./";
} else {
- datadir = getenv("TESSDATA_PREFIX");
+ // Remove tessdata from the end if present, as we will add it back!
+ int length = datadir.length();
+ if (length >= 8 && strcmp(&datadir[length - 8], "tessdata") == 0)
+ datadir.truncate_at(length - 8);
+ else if (length >= 9 && strcmp(&datadir[length - 9], "tessdata/") == 0)
+ datadir.truncate_at(length - 9);
}
// check for missing directory separator
