06/06: gnu: libextractor: Enable tidy-html support.

[guix-commits]

apteryx pushed a commit to branch master
in repository guix.

commit 43d6d5e6461d930dda4846279b3d56c97467bc06
Author: Maxim Cournoyer <maxim.cournoyer@gmail.com>
AuthorDate: Tue Mar 12 21:56:17 2024 -0400

    gnu: libextractor: Enable tidy-html support.
    
    * gnu/packages/gnunet.scm (libextractor)
    [source]: Apply patch.
    [phases] <force-reconfigure>: New phase.
    [inputs]: Add tidy-html.  Remove associated comment.
    * gnu/packages/patches/libextractor-tidy-support.patch: New file.
    * gnu/local.mk (dist_patch_DATA): Register it.
    
    Change-Id: Ic812e09504d522ec87410bbbb03ccd3d6e48dd71
---
 gnu/local.mk                                       |  1 +
 gnu/packages/gnunet.scm                            | 14 ++--
 .../patches/libextractor-tidy-support.patch        | 81 ++++++++++++++++++++++
 3 files changed, 89 insertions(+), 7 deletions(-)

diff --git a/gnu/local.mk b/gnu/local.mk
index e8b0a0e7be..574a67b49c 100644
--- a/gnu/local.mk
+++ b/gnu/local.mk
@@ -1514,6 +1514,7 @@ dist_patch_DATA =                                         
\
   %D%/packages/patches/julia-SOURCE_DATE_EPOCH-mtime.patch     \
   %D%/packages/patches/julia-Use-MPFR-4.2.patch                        \
   %D%/packages/patches/libcss-check-format.patch               \
+  %D%/packages/patches/libextractor-tidy-support.patch         \
   %D%/packages/patches/libftdi-fix-paths-when-FTDIPP-set.patch \
   %D%/packages/patches/libgeotiff-fix-tests-with-proj-9.1.1.patch      \
   %D%/packages/patches/libgeotiff-fix-tests-with-proj-9.3.0.patch      \
diff --git a/gnu/packages/gnunet.scm b/gnu/packages/gnunet.scm
index 828ddb99af..e23755302d 100644
--- a/gnu/packages/gnunet.scm
+++ b/gnu/packages/gnunet.scm
@@ -93,7 +93,9 @@
                                   version ".tar.gz"))
               (sha256
                (base32
-                "0mgprmwdhdwq9xhfxfhcncd304425nvcc4zi8ci5f0nja4n333xv"))))
+                "0mgprmwdhdwq9xhfxfhcncd304425nvcc4zi8ci5f0nja4n333xv"))
+              (patches
+               (search-patches "libextractor-tidy-support.patch"))))
     (build-system gnu-build-system)
     (outputs '("out"
                "static"))               ; 420 KiB .a files
@@ -103,6 +105,9 @@
                                   #$(this-package-input "libltdl")))
            #:phases
            #~(modify-phases %standard-phases
+               (add-after 'unpack 'force-reconfigure
+                 (lambda _
+                   (delete-file "configure")))
                (add-after 'install 'move-static-libraries
                  (lambda* (#:key outputs #:allow-other-keys)
                    ;; Move static libraries to the "static" output.
@@ -114,12 +119,6 @@
                                  (install-file file slib)
                                  (delete-file file))
                                (find-files lib "\\.a$"))))))))
-    ;; WARNING: Checks require /dev/shm to be in the build chroot, especially
-    ;; not to be a symbolic link to /run/shm.
-    ;; FIXME:
-    ;; The following dependency is optional, but should be
-    ;; available for maximum coverage:
-    ;; * libtidy-html (tidy-html) ; investigate failure
     (native-inputs
      (list autoconf-2.71
            automake
@@ -149,6 +148,7 @@
            libtiff
            libvorbis
            rpm
+           tidy-html
            zlib))
     (synopsis "Library to extract meta-data from media files")
     (description
diff --git a/gnu/packages/patches/libextractor-tidy-support.patch 
b/gnu/packages/patches/libextractor-tidy-support.patch
new file mode 100644
index 0000000000..8fe2f21b2b
--- /dev/null
+++ b/gnu/packages/patches/libextractor-tidy-support.patch
@@ -0,0 +1,81 @@
+Upstream status: submitted to bug-libextractor@gnu.org.
+
+From 1fc6daaeaf829fb941a176831c011888a73c43b9 Mon Sep 17 00:00:00 2001
+From: Maxim Cournoyer <maxim.cournoyer@gmail.com>
+Date: Mon, 11 Mar 2024 09:36:26 -0400
+Subject: [PATCH] html_extractor: Add support for modern tidy-html.
+
+* configure.ac: Use PKG_PROG_PKG_CONFIG to initialize pkg-config detection.
+<tidy>: Check for library via pkg-config.
+* src/plugins/html_extractor.c: Standardize tidy include file names.
+---
+ configure.ac                 | 28 +++++++++-------------------
+ src/plugins/html_extractor.c |  4 ++--
+ 2 files changed, 11 insertions(+), 21 deletions(-)
+
+diff --git a/configure.ac b/configure.ac
+index d17ff39..e89d70c 100644
+--- a/configure.ac
++++ b/configure.ac
+@@ -176,6 +176,8 @@ AS_CASE(["$target_os"],
+ 
+ AM_ICONV
+ 
++PKG_PROG_PKG_CONFIG()
++
+ # We define the paths here, because MinGW/GCC expands paths
+ # passed through the command line ("-DLOCALEDIR=..."). This would
+ # lead to hard-coded paths ("C:\mingw\mingw\bin...") that do
+@@ -424,25 +426,13 @@ AC_CHECK_LIB(magic, magic_open,
+    AM_CONDITIONAL(HAVE_MAGIC, false))],
+   AM_CONDITIONAL(HAVE_MAGIC, false))
+ 
+-AC_MSG_CHECKING(for tidyNodeGetValue -ltidy)
+-AC_LANG_PUSH(C++)
+-SAVED_LIBS=$LIBS
+-LIBS="$LIBS -ltidy"
+-AC_LINK_IFELSE(
+-  [AC_LANG_PROGRAM([[#include <tidy/tidy.h>]],
+-    [[ Bool b = tidyNodeGetValue (NULL, NULL, NULL); ]])],
+-  [AC_MSG_RESULT(yes)
+-   AM_CONDITIONAL(HAVE_TIDY, true)
+-   AC_DEFINE(HAVE_TIDY,1,[Have tidyNodeGetValue in libtidy])],
+-  [AC_MSG_RESULT(no)
+-   AM_CONDITIONAL(HAVE_TIDY, false)])
+-LIBS=$SAVED_LIBS
+-AC_LANG_POP(C++)
+-
+-# restore LIBS
+-LIBS=$LIBSOLD
+-
+-
++dnl tidyNodeGetValue was already available in 5.0.0, released in 2015.
++PKG_CHECK_MODULES([TIDY], [tidy >= 5.0.0],
++ [AC_DEFINE(HAVE_TIDY, 1, [Have tidy])
++  AM_CONDITIONAL(HAVE_TIDY, true)],
++ [AM_CONDITIONAL(HAVE_TIDY, false)])
++CFLAGS="$CFLAGS $TIDY_CFLAGS"
++LIBS="$LIBS $TIDY_LIBS"
+ 
+ # should 'make check' run tests?
+ AC_MSG_CHECKING(whether to run tests)
+diff --git a/src/plugins/html_extractor.c b/src/plugins/html_extractor.c
+index 5ebf97b..88100d3 100644
+--- a/src/plugins/html_extractor.c
++++ b/src/plugins/html_extractor.c
+@@ -26,8 +26,8 @@
+ #include "platform.h"
+ #include "extractor.h"
+ #include <magic.h>
+-#include <tidy/tidy.h>
+-#include <tidy/tidybuffio.h>
++#include <tidy.h>
++#include <tidybuffio.h>
+ 
+ /**
+  * Mapping of HTML META names to LE types.
+
+base-commit: a75f40b64b5868967c95ea214e8eaac4f7088b23
+-- 
+2.41.0
+