tesseract-ocr: update to 4.0.0.

Closes: #5380 [via git-merge-pr] Signed-off-by: Jürgen Buchmüller <pullmoll@t-online.de>
2018-12-03 21:46:48 +01:00 · 2018-12-03 21:46:48 +01:00 · 7f7ff14e13
parent 21b98df42a
commit 7f7ff14e13
22 changed files with 194 additions and 34 deletions
--- a/common/shlibs
+++ b/common/shlibs
@ -2077,7 +2077,7 @@ libhttp_parser.so.2.8 http-parser-2.8.0_1
 libmaa.so.4 libmaa-1.4.2_1
 libcodeblocks.so.0 codeblocks-13.12_1
 liblept.so.5 leptonica-1.73_1
-libtesseract.so.3 tesseract-ocr-3.02.02_1
+libtesseract.so.4 tesseract-ocr-4.0.0_1
 libffmpegthumbnailer.so.4 ffmpegthumbnailer-2.0.10_1
 libopenraw.so.7 libopenraw-0.1.0_1
 libopenrawgnome.so.7 libopenraw-0.1.0_1
--- a/srcpkgs/tesseract-ocr-bre
+++ b/srcpkgs/tesseract-ocr-bre
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-cos
+++ b/srcpkgs/tesseract-ocr-cos
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-div
+++ b/srcpkgs/tesseract-ocr-div
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-fao
+++ b/srcpkgs/tesseract-ocr-fao
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-fil
+++ b/srcpkgs/tesseract-ocr-fil
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-fry
+++ b/srcpkgs/tesseract-ocr-fry
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-gla
+++ b/srcpkgs/tesseract-ocr-gla
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-hye
+++ b/srcpkgs/tesseract-ocr-hye
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-kur_ara
+++ b/srcpkgs/tesseract-ocr-kur_ara
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-ltz
+++ b/srcpkgs/tesseract-ocr-ltz
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-mon
+++ b/srcpkgs/tesseract-ocr-mon
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-mri
+++ b/srcpkgs/tesseract-ocr-mri
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-oci
+++ b/srcpkgs/tesseract-ocr-oci
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-que
+++ b/srcpkgs/tesseract-ocr-que
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-snd
+++ b/srcpkgs/tesseract-ocr-snd
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-sun
+++ b/srcpkgs/tesseract-ocr-sun
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-tat
+++ b/srcpkgs/tesseract-ocr-tat
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-ton
+++ b/srcpkgs/tesseract-ocr-ton
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-yor
+++ b/srcpkgs/tesseract-ocr-yor
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr/patches/svutil-sys_select_h.patch
+++ b/srcpkgs/tesseract-ocr/patches/svutil-sys_select_h.patch
@ -1,12 +0,0 @@
-Add include required to build with muslc libc
-
--- viewer/svutil.cpp	2012-03-03 12:53:33.000000000 +0100
-+++ viewer/svutil.cpp	2015-05-28 17:51:46.917525843 +0200
-@@ -39,6 +39,7 @@
- #include <string.h>
- #include <netdb.h>
- #include <sys/socket.h>
-+#include <sys/select.h>
- #ifdef __linux__
- #include <sys/prctl.h>
- #endif
--- a/srcpkgs/tesseract-ocr/template
+++ b/srcpkgs/tesseract-ocr/template
@ -1,12 +1,11 @@
 # Template file for 'tesseract-ocr'
 pkgname=tesseract-ocr
-version=3.05.02
-revision=3
-wrksrc=tesseract-${version}
-_tessdata_ver=074c37215b01ab8cc47a0e06ff7356383883d775
+version=4.0.0
+revision=1
+wrksrc="tesseract-${version}"
 build_style=gnu-configure
 configure_args="LIBLEPT_HEADERSDIR=${XBPS_CROSS_BASE}/usr/include"
-hostmakedepends="automake libtool pkg-config leptonica"
+hostmakedepends="automake libtool pkg-config leptonica libxslt asciidoc"
 makedepends="cairo-devel pango-devel leptonica-devel icu-devel"
 short_desc="Tesseract Open Source OCR engine"
 maintainer="Jürgen Buchmüller <pullmoll@t-online.de>"
@ -14,24 +13,25 @@ license="Apache-2.0"
 homepage="https://github.com/tesseract-ocr/tesseract"
 distfiles="
 https://github.com/tesseract-ocr/tesseract/archive/${version}.tar.gz>${pkgname}-${version}.tar.gz
- https://github.com/tesseract-ocr/tessdata/archive/${_tessdata_ver}.tar.gz>tessdata-${_tessdata_ver}.tar.gz"
-checksum="
- 494d64ffa7069498a97b909a0e65a35a213989e0184f1ea15332933a90d43445
- e33dea2118f447848a76e0fa5d50d45a2b8630cccc6adeb8d58221a1d09d6007"
+ https://github.com/tesseract-ocr/tessdata/archive/${version}.tar.gz>tessdata-${version}.tar.gz"
+checksum="a1f5422ca49a32e5f35c54dee5112b11b99928fc9f4ee6695cdc6768d69f61dd
+ 38c637d3a1763f6c3d32e8f1d979f045668676ec5feb8ee1869ee77cedd31b08"

 # Create a package for one specific language $1
 pkg_lang() {
 	local f lang=$1
 	vmkdir usr/share/tessdata
-	for f in $(find ${wrksrc}/tessdata -name "${lang}.*" -o -name "${lang}_frak.*"); do
+	for f in $(find ${wrksrc}/tessdata -name "${lang}.*" \
+	-o -name "${lang}_frak.*" \
+	-o -name "${lang}_vert.*" ); do
 		vinstall $f 644 usr/share/tessdata
 		rm $f
 	done
 }

 post_extract() {
-	mv ${XBPS_BUILDDIR}/tessdata-${_tessdata_ver}/* ${wrksrc}/tessdata
-	rmdir ${XBPS_BUILDDIR}/tessdata-${_tessdata_ver}
+	mv ${XBPS_BUILDDIR}/tessdata-${version}/* ${wrksrc}/tessdata
+	rmdir ${XBPS_BUILDDIR}/tessdata-${version}
 }
 pre_configure() {
 	NOCONFIGURE=1 ./autogen.sh
@ -39,6 +39,9 @@ pre_configure() {
 post_build() {
 	make ${makejobs} training
 }
+do_check() {
+	: # submodule not in tarball
+}
 post_install() {
 	local lang
 	# Rename binary to avoid conflict with tesseract package
@ -46,8 +49,6 @@ post_install() {
 	mv ${DESTDIR}/usr/share/man/man1/tesseract{,-ocr}.1
 	vdoc ChangeLog
 	vdoc README.md
-	vdoc testing/eurotext.tif
-	vdoc testing/phototest.tif
 	vlicense ${FILESDIR}/COPYING LICENSE-tessdata
 	# Move the pseudo languges "equ" (math / equation detection) and
 	# "osd" (orientation and script detection) to the main package
@ -113,13 +114,13 @@ tesseract-ocr-all_package() {
 	conflicts="tesseract-ocr-basic>=0"
 	short_desc+=" - all languages data"
 	# All available languages
-	for lang in afr amh ara asm aze aze_cyrl bel ben bod bos bul cat ceb \
-		ces chi_sim chi_tra chr cym dan deu dzo ell eng enm epo est eus \
-		fas fin fra frk frm gle glg grc guj hat heb hin hrv hun iku ind isl ita \
-		ita_old jav jpn kan kat kat_old kaz khm kir kor kur lao lat lav lit mal mar \
-		mkd mlt msa mya nep nld nor ori pan pol por pus ron rus san sin slk slv \
-		spa spa_old sqi srp srp_latn swa swe syr tam tel tgk tgl tha tir tur \
-		uig ukr urd uzb uzb_cyrl vie yid; do
+	for lang in afr amh ara asm aze aze_cyrl bel ben bod bos bre bul cat ceb \
+		ces chi_sim chi_tra chr cos cym dan deu div dzo ell eng enm epo est eus fao \
+		fas fil fin fra frk frm fry gla gle glg grc guj hat heb hin hrv hun hye iku ind isl ita \
+		ita_old jav jpn kan kat kat_old kaz khm kir kor kur kur_ara lao lat lav lit ltz mal mar \
+		mkd mlt mon mri msa mya nep nld nor oci ori pan pol por que pus ron rus san sin slk slv \
+		snd spa spa_old sqi srp srp_latn sun swa swe syr tam tat tel tgk tgl tha tir ton tur \
+		uig ukr urd uzb uzb_cyrl vie yor yid; do
 		depends+=" tesseract-ocr-${lang}>=${version}_${revision}"
 	done
 }
@ -203,6 +204,14 @@ tesseract-ocr-bos_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-bre_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Breton language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-bul_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -259,6 +268,14 @@ tesseract-ocr-chr_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-cos_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Corsican language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-cym_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -283,6 +300,14 @@ tesseract-ocr-deu_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-div_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Dhivehi language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-dzo_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -339,6 +364,14 @@ tesseract-ocr-eus_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-fao_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Faroese language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-fas_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -347,6 +380,14 @@ tesseract-ocr-fas_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-fil_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Filipino language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-fin_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -379,6 +420,22 @@ tesseract-ocr-frm_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-fry_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Frisian language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-gla_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Scottish Gaelic language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-gle_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -451,6 +508,14 @@ tesseract-ocr-hun_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-hye_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Armenian language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-iku_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -571,6 +636,14 @@ tesseract-ocr-kur_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-kur_ara_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Kurdish (Arabic) language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-lao_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -603,6 +676,14 @@ tesseract-ocr-lit_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-ltz_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Luxembourgish language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-mal_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -635,6 +716,22 @@ tesseract-ocr-mlt_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-mon_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Mongolian language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-mri_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Maori language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-msa_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -675,6 +772,14 @@ tesseract-ocr-nor_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-oci_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Occitan (post 1500) language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-ori_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -715,6 +820,14 @@ tesseract-ocr-pus_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-que_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Quechua language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-ron_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -763,6 +876,14 @@ tesseract-ocr-slv_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-snd_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Sindhi language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-spa_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -803,6 +924,14 @@ tesseract-ocr-srp_latn_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-sun_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Sundanese language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-swa_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -835,6 +964,14 @@ tesseract-ocr-tam_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-tat_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Tatar language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-tel_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -875,6 +1012,14 @@ tesseract-ocr-tir_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-ton_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Tonga language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
 tesseract-ocr-tur_package() {
 	noarch=yes
 	depends="${sourcepkg}>=${version}_${revision}"
@ -939,3 +1084,11 @@ tesseract-ocr-yid_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-yor_package() {
+	noarch=yes
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Yoruba language data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}