app-text/tesseract: add 5.3.2

Note removing GCC 13 patch as per tesseract-ocr/tesseract@e3fb0c6 Closes: https://bugs.gentoo.org/910403 Signed-off-by: Alexander Kurakin <[email protected]> Closes: gentoo#31899 Signed-off-by: Bernard Cafarelli <[email protected]>
AndCycle · Jul 19, 2023 · b7162c7 · b7162c7
1 parent e2407ea
commit b7162c7
Show file tree

Hide file tree

Showing 2 changed files with 93 additions and 0 deletions.
diff --git a/app-text/tesseract/Manifest b/app-text/tesseract/Manifest
@@ -1,3 +1,4 @@
 DIST tesseract-4.1.1.tar.gz 1974988 BLAKE2B f729beb017ec81704b84f5e6539706ff0cd6a51a4d1b7b730c394af58ce2ea2a659203bf239417e8d500cfbde845a73d6e6e181ca946257c2f22525820899602 SHA512 017723a2268be789fe98978eed02fd294968cc8050dde376dee026f56f2b99df42db935049ae5e72c4519a920e263b40af1a6a40d9942e66608145b3131a71a2
 DIST tesseract-5.2.0.tar.gz 1913462 BLAKE2B 6b80e5bb71543cf971683f66b87fd45a971a6f337ca53567d134c3bdc227f1338a3340f5bb5c2f183288135acd6521031b0f6de6e96e333fed4dc10e93b7319e SHA512 23ed931495e4ea793904f99a06a95a47873b923c4e5d8f3fc8f9203cac86b0e0013c416ab4d8d4a863a081eea00a196a0198c6fb17a39917a1baea2280d84794
 DIST tesseract-5.3.0.tar.gz 1913678 BLAKE2B 975b3a50dc2c9ec7ded82caaa068d2d9d362ee0bc6d51b41a52c692ac7f51231d4c64a9dda9e7b850297d4c1f13eb40a709e59eebe143df4f5e2a060cfbbfd3a SHA512 fc0cfda980059390b7d6d3d8e475b4ecc9dcfee1abc3b16fc9ffc8904db0a0847a5f6f1b32bd2e1f6f5a714a8c4df0fb3dc7c6aa25ece6162bf52fe86b4d0f34
+DIST tesseract-5.3.2.tar.gz 1916850 BLAKE2B d900f7e21a53a4ca77ab820359cf8b528ac0114816029c257dffa141ae9fdecfe4ddb97160b598d1f144c139f3fbceea080fb050802d6d3fa7ac8bc8e1353d72 SHA512 92db2a513f00b931a0207cc705020bee9fea4ff7fa93c57a6f446b5777713864cce19121b316ef8b78b6bd625ee7e6fd473352515f98c4a2187569cbd0a9713a
diff --git a/app-text/tesseract/tesseract-5.3.2.ebuild b/app-text/tesseract/tesseract-5.3.2.ebuild
@@ -0,0 +1,92 @@
+# Copyright 1999-2023 Gentoo Authors
+# Distributed under the terms of the GNU General Public License v2
+
+EAPI=8
+
+inherit autotools multilib-minimal toolchain-funcs
+
+DESCRIPTION="An OCR Engine, originally developed at HP, now open source"
+HOMEPAGE="https://github.com/tesseract-ocr"
+SRC_URI="https://github.com/tesseract-ocr/${PN}/archive/${PV}.tar.gz -> ${P}.tar.gz"
+
+LICENSE="Apache-2.0"
+SLOT="0/5"
+KEYWORDS="~alpha ~amd64 ~arm ~arm64 ~mips ~ppc ~ppc64 ~riscv ~sparc ~x86"
+IUSE="doc float32 jpeg opencl openmp png static-libs tiff training webp"
+
+COMMON_DEPEND=">=media-libs/leptonica-1.74:=[${MULTILIB_USEDEP},zlib,tiff?,jpeg?,png?,webp?]
+	opencl? (
+		virtual/opencl[${MULTILIB_USEDEP}]
+		media-libs/tiff:=[${MULTILIB_USEDEP}]
+		media-libs/leptonica:=[tiff]
+	)
+	training? (
+		dev-libs/icu:=
+		x11-libs/pango:=
+		x11-libs/cairo:=
+	)"
+RDEPEND="${COMMON_DEPEND}
+	|| (
+		>=app-text/tessdata_fast-4.0.0
+		>=app-text/tessdata_best-4.0.0
+		>=app-text/tessdata_legacy-4.0.0
+	)"
+DEPEND="${COMMON_DEPEND}
+	app-text/asciidoc
+	app-text/docbook-xsl-stylesheets
+	dev-libs/libxslt
+	doc? ( app-doc/doxygen )"
+
+pkg_pretend() {
+	[[ ${MERGE_TYPE} != binary ]] && use openmp && tc-check-openmp
+}
+
+pkg_setup() {
+	[[ ${MERGE_TYPE} != binary ]] && use openmp && tc-check-openmp
+}
+
+src_prepare() {
+	default
+	eautoreconf
+}
+
+multilib_src_configure() {
+	# scrollview disabled for now, see bug #686944
+	local myeconfargs=(
+		--enable-shared
+		--disable-graphics
+		$(use_enable float32)
+		$(use_enable opencl)
+		$(use_enable openmp)
+		$(use_enable static-libs static)
+	)
+
+	ECONF_SOURCE="${S}" econf "${myeconfargs[@]}"
+}
+
+multilib_src_compile() {
+	default
+	if multilib_is_native_abi; then
+		use doc && emake doc
+		use training && emake training
+	fi
+}
+
+multilib_src_install() {
+	if multilib_is_native_abi; then
+		DOCS=( "${S}"/AUTHORS "${S}"/ChangeLog "${S}"/README.md )
+		if use doc; then
+			HTML_DOCS=( doc/html/. )
+		fi
+		einstalldocs
+
+		if use training; then
+			emake DESTDIR="${D}" training-install
+		fi
+	fi
+	emake DESTDIR="${D}" install
+}
+
+multilib_src_install_all() {
+	find "${D}" -name '*.la' -type f -delete || die
+}