# Buildsheet autogenerated by ravenadm tool -- Do not edit.

NAMEBASE=		tesseract
VERSION=		4.1.1
REVISION=		1
KEYWORDS=		graphics
VARIANTS=		standard
SDESC[standard]=	OCR (Optical Character Recognition) engine
HOMEPAGE=		https://github.com/tesseract-ocr/tesseract
CONTACT=		nobody

DOWNLOAD_GROUPS=	main
SITES[main]=		GITHUB/tesseract-ocr:tesseract:4.1.1
DISTFILE[1]=		generated:main
DF_INDEX=		1
SPKGS[standard]=	complete
			tools
			library

OPTIONS_AVAILABLE=	none
OPTIONS_STANDARD=	none

BUILDRUN_DEPENDS=	fontconfig:primary:standard
			icu:single:standard
			leptonica:single:standard
			llvm:single:standard
			pango:primary:standard
RUN_DEPENDS=		tesseract-data:latin:standard
EXRUN[tools]=		tesseract:library:standard

USES=			autoreconf libtool pkgconfig shebangfix
C_USES[freebsd]=	fortran:library

FPC_EQUIVALENT=		graphics/tesseract
SHEBANG_FILES=		{{WRKSRC}}/src/training/tesstrain.sh
			{{WRKSRC}}/src/training/language-specific.sh

MUST_CONFIGURE=		gnu
CONFIGURE_ENV=		LIBLEPT_HEADERSDIR="{{LOCALBASE}}/include/leptonica"

MAKE_ARGS=		datadir={{LOCALBASE}}/share/tesseract-data

INSTALL_TARGET=		install-strip training-install
INSTALL_REQ_TOOLCHAIN=	yes
SOVERSION=		4.0.1

LDFLAGS=		-lpthread

post-patch:
	${REINPLACE_CMD} -e 's|-O3 -DNDEBUG||' ${WRKSRC}/configure.ac

[FILE:54:descriptions/desc.tools]
This package contains the Tesseract OCR engine tools.


[FILE:291:descriptions/desc.library]
The Tesseract OCR engine was one of the top 3 engines in the 1995 UNLV
Accuracy test. Between 1995 and 2006 it had little work done on it, but
it is probably one of the most accurate open source OCR engines available.
The source code will read a binary, grey or color image and output text.


[FILE:115:distinfo]
2a66ff0d8595bff8f04032165e6c936389b1e5727c3ce5a27b3e059d218db1cb      1974988 tesseract-ocr-tesseract-4.1.1.tar.gz


[FILE:310:manifests/plist.tools]
bin/
 ambiguous_words
 classifier_tester
 cntraining
 combine_lang_model
 combine_tessdata
 dawg2wordlist
 language-specific.sh
 lstmeval
 lstmtraining
 merge_unicharsets
 mftraining
 set_unicharset_properties
 shapeclustering
 tesstrain.sh
 tesstrain_utils.sh
 text2image
 unicharset_extractor
 wordlist2dawg


[FILE:732:manifests/plist.library]
bin/tesseract
include/tesseract/
 apitypes.h
 baseapi.h
 capi.h
 genericvector.h
 helpers.h
 ltrresultiterator.h
 ocrclass.h
 osdetect.h
 pageiterator.h
 platform.h
 publictypes.h
 renderer.h
 resultiterator.h
 serialis.h
 strngs.h
 tess_version.h
 tesscallback.h
 thresholder.h
 unichar.h
lib/
 libtesseract.a
 libtesseract.so
 libtesseract.so.%%SOMAJOR%%
 libtesseract.so.%%SOVERSION%%
lib/pkgconfig/tesseract.pc
share/tesseract-data/
 alto
 ambigs.train
 api_config
 batch
 batch.nochop
 bigram
 box.train
 box.train.stderr
 digits
 get.images
 hocr
 inter
 kannada
 linebox
 logfile
 lstm.train
 lstmbox
 lstmdebug
 makebox
 matdemo
 msdemo
 nobatch
 pdf
 pdf.ttf
 quiet
 rebox
 segdemo
 strokewidth
 tsv
 txt
 unlv
 wordstrbox


[FILE:609:freebsd/patch-configure.ac]
sem_init(3) is part of libc on FreeBSD, so we ended up not linking against
libpthread. Look for a symbol that causes us to load it.
libtesseract.so is not linked against libomp if openmp is enabled,
which makes dependent ports fail. This adds a workaround by linking
directly against libomp.
--- configure.ac.orig	2019-07-07 12:34:08 UTC
+++ configure.ac
@@ -372,6 +372,9 @@ esac
 # ----------------------------------------
 
 AC_SEARCH_LIBS([pthread_create], [pthread])
+if test "$enable_openmp" != no; then
+  AC_SEARCH_LIBS([omp_get_thread_num ], [omp])
+fi
 
 
 # ----------------------------------------


[FILE:276:freebsd/patch-src_ccutil_ocrclass.h]
--- src/ccutil/ocrclass.h.orig	2019-07-07 12:34:08 UTC
+++ src/ccutil/ocrclass.h
@@ -31,6 +31,7 @@
 #ifdef _WIN32
 #include <winsock2.h> // for timeval
 #endif
+#include <sys/time.h>
 
 /**********************************************************************
  * EANYCODE_CHAR