sys-apps/file: Fix character count heuristic

This restores behaviour of file-5.39, i.e. correctly identifies small
text files as text/plain rather than application/octet-stream again.

Patch taken from upstream and (trivially) backported to 5.40.

Closes: https://bugs.gentoo.org/799188
Bug: https://bugs.astron.com/view.php?id=261
Package-Manager: Portage-3.0.20, Repoman-3.0.3
Signed-off-by: Ulrich Müller <ulm@gentoo.org>
Signed-off-by: Lars Wendler <polynomial-c@gentoo.org>

2 files changed, 196 insertions, 0 deletions

diff --git a/sys-apps/file/file-5.40-r3.ebuild b/sys-apps/file/file-5.40-r3.ebuild
new file mode 100644
index 000000000000..7fd5ac3b71b9
--- /dev/null
+++ b/sys-apps/file/file-5.40-r3.ebuild
@@ -0,0 +1,147 @@
+# Copyright 1999-2021 Gentoo Authors
+# Distributed under the terms of the GNU General Public License v2
+
+EAPI=7
+
+PYTHON_COMPAT=( python3_{8..10} )
+DISTUTILS_OPTIONAL=1
+
+inherit distutils-r1 libtool toolchain-funcs multilib-minimal
+
+if [[ ${PV} == "9999" ]] ; then
+	EGIT_REPO_URI="https://github.com/glensc/file.git"
+	inherit autotools git-r3
+else
+	SRC_URI="ftp://ftp.astron.com/pub/file/${P}.tar.gz"
+	KEYWORDS="~alpha ~amd64 ~arm ~arm64 ~hppa ~ia64 ~m68k ~mips ~ppc ~ppc64 ~riscv ~s390 ~sparc ~x86 ~x64-cygwin ~amd64-linux ~x86-linux ~ppc-macos ~x64-macos ~sparc-solaris ~sparc64-solaris ~x64-solaris ~x86-solaris"
+fi
+
+DESCRIPTION="identify a file's format by scanning binary data for patterns"
+HOMEPAGE="https://www.darwinsys.com/file/"
+
+LICENSE="BSD-2"
+SLOT="0"
+IUSE="bzip2 lzma python seccomp static-libs zlib"
+REQUIRED_USE="python? ( ${PYTHON_REQUIRED_USE} )"
+
+DEPEND="
+	bzip2? ( app-arch/bzip2[${MULTILIB_USEDEP}] )
+	lzma? ( app-arch/xz-utils[${MULTILIB_USEDEP}] )
+	python? (
+		${PYTHON_DEPS}
+		dev-python/setuptools[${PYTHON_USEDEP}]
+	)
+	zlib? ( >=sys-libs/zlib-1.2.8-r1[${MULTILIB_USEDEP}] )"
+RDEPEND="${DEPEND}
+	python? ( !dev-python/python-magic )
+	seccomp? ( sys-libs/libseccomp[${MULTILIB_USEDEP}] )"
+
+PATCHES=(
+	"${FILESDIR}/file-5.39-portage-sandbox.patch" #713710 #728978
+	"${FILESDIR}/file-5.40-xz_magic.patch" #784773
+	"${FILESDIR}/file-5.40-seccomp-faccessat.patch"
+	"${FILESDIR}/file-5.40-seccomp-fstatat64.patch" #784857
+	"${FILESDIR}/file-5.40-revert-char-count.patch" #799188
+)
+
+src_prepare() {
+	default
+
+	if [[ ${PV} == 9999 ]] ; then
+		eautoreconf
+	fi
+
+	elibtoolize
+
+	# don't let python README kill main README #60043
+	mv python/README.md python/README.python.md || die
+	sed 's@README.md@README.python.md@' -i python/setup.py || die #662090
+}
+
+multilib_src_configure() {
+	local myeconfargs=(
+		--enable-fsect-man5
+		$(use_enable bzip2 bzlib)
+		$(use_enable lzma xzlib)
+		$(use_enable seccomp libseccomp)
+		$(use_enable static-libs static)
+		$(use_enable zlib)
+	)
+	econf "${myeconfargs[@]}"
+}
+
+build_src_configure() {
+	local myeconfargs=(
+		--disable-shared
+		--disable-libseccomp
+		--disable-bzlib
+		--disable-xzlib
+		--disable-zlib
+	)
+	tc-env_build econf "${myeconfargs[@]}"
+}
+
+need_build_file() {
+	# when cross-compiling, we need to build up our own file
+	# because people often don't keep matching host/target
+	# file versions #362941
+	tc-is-cross-compiler && ! has_version -b "~${CATEGORY}/${P}"
+}
+
+src_configure() {
+	local ECONF_SOURCE="${S}"
+
+	if need_build_file ; then
+		mkdir -p "${WORKDIR}"/build || die
+		cd "${WORKDIR}"/build || die
+		build_src_configure
+	fi
+
+	multilib-minimal_src_configure
+}
+
+multilib_src_compile() {
+	if multilib_is_native_abi ; then
+		emake
+	else
+		cd src || die
+		emake magic.h #586444
+		emake libmagic.la
+	fi
+}
+
+src_compile() {
+	if need_build_file ; then
+		emake -C "${WORKDIR}"/build/src magic.h #586444
+		emake -C "${WORKDIR}"/build/src file
+		local -x PATH="${WORKDIR}/build/src:${PATH}"
+	fi
+	multilib-minimal_src_compile
+
+	if use python ; then
+		cd python || die
+		distutils-r1_src_compile
+	fi
+}
+
+multilib_src_install() {
+	if multilib_is_native_abi ; then
+		default
+	else
+		emake -C src install-{nodist_includeHEADERS,libLTLIBRARIES} DESTDIR="${D}"
+	fi
+}
+
+multilib_src_install_all() {
+	dodoc ChangeLog MAINT README
+
+	# Required for `file -C`
+	insinto /usr/share/misc/magic
+	doins -r magic/Magdir/*
+
+	if use python ; then
+		cd python || die
+		distutils-r1_src_install
+	fi
+	find "${ED}" -type f -name "*.la" -delete || die
+}
diff --git a/sys-apps/file/files/file-5.40-revert-char-count.patch b/sys-apps/file/files/file-5.40-revert-char-count.patch
new file mode 100644
index 000000000000..9d6f5be60fac
--- /dev/null
+++ b/sys-apps/file/files/file-5.40-revert-char-count.patch
@@ -0,0 +1,49 @@
+From c07e242e766242a44ff720c149b1bdd4924ec247 Mon Sep 17 00:00:00 2001
+From: Christos Zoulas <christos@zoulas.com>
+Date: Tue, 27 Apr 2021 19:37:14 +0000
+Subject: [PATCH] Revert the fix for PR/180. It lead to PR/261. Using character
+ count heuristics ends up with confusing behavior, the following should not be
+ producing different results:     echo -n xx | ./file -     echo -n xy |
+ ./file -
+
+---
+[patch backported to 5.40 release -- ulm]
+
+ src/encoding.c | 15 ++-------------
+ 1 file changed, 2 insertions(+), 13 deletions(-)
+
+diff --git a/src/encoding.c b/src/encoding.c
+index 31d4d125..3647a481 100644
+--- a/src/encoding.c
++++ b/src/encoding.c
+@@ -265,9 +265,7 @@ private int \
+ looks_ ## NAME(const unsigned char *buf, size_t nbytes, file_unichar_t *ubuf, \
+     size_t *ulen) \
+ { \
+-	size_t i, u; \
+-	unsigned char dist[256]; \
+-	memset(dist, 0, sizeof(dist)); \
++	size_t i; \
+ \
+ 	*ulen = 0; \
+ \
+@@ -278,16 +276,7 @@ looks_ ## NAME(const unsigned char *buf, size_t nbytes, file_unichar_t *ubuf, \
+ 			return 0; \
+ \
+ 		ubuf[(*ulen)++] = buf[i]; \
+-		dist[buf[i]]++; \
+ 	} \
+-	u = 0; \
+-	for (i = 0; i < __arraycount(dist); i++) { \
+-		if (dist[i]) \
+-			u++; \
+-	} \
+-	if (u < 3) \
+-		return 0; \
+-\
+ 	return 1; \
+ }
+ 
+-- 
+2.32.0
+