[bug#60976] [PATCH 1/3] gnu: Add java-jericho-html

[Frank Pursel <frank.pursel@gmail.com>]

---
 gnu/packages/java-xml.scm | 79 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 79 insertions(+)

diff --git a/gnu/packages/java-xml.scm b/gnu/packages/java-xml.scm
index 5b16806e27..9cf256549b 100644
--- a/gnu/packages/java-xml.scm
+++ b/gnu/packages/java-xml.scm
@@ -6,6 +6,7 @@
 ;;; Copyright © 2020 Vincent Legoll <vincent.legoll@gmail.com>
 ;;; Copyright © 2021 Léo Le Bouter <lle-bout@zaclys.net>
 ;;; Copyright © 2022 Christopher Baines <mail@cbaines.net>
+;;; Copyright © 2023 Frank Pursel <frank.pursel@gmail.com>
 ;;;
 ;;; This file is part of GNU Guix.
 ;;;
@@ -37,6 +38,84 @@ (define-module (gnu packages java-xml)
   #:use-module (guix build-system ant)
   #:use-module (guix utils))
 
+(define-public java-jericho-html
+  (let ((ver "3.4")
+        (hash "1i7z3b8yamgkm7p3pv9qzv8l2kw15ccxy1aj9mpbf66kzkkds51i")
+        (internal-name "jericho-html"))
+    (package
+      (name "java-jericho-html")
+      (version ver)
+      (source (origin
+                (method url-fetch/zipbomb)
+                (uri (string-append
+                      "https://sourceforge.net/projects/jerichohtml"
+                      "/files/jericho-html/3.4/jericho-html-3.4.zip/download"))
+                (sha256
+                 (base32
+                  hash))
+                (modules '((guix build utils)))
+                (snippet '(begin
+                            (format #t "~%~a~%" "Removing sourced jar files.")
+                            (for-each (lambda (jarf)
+                                        (delete-file jarf)
+                                        (format #t "Deleted: ~a~%" jarf))
+                                      (find-files "." "\\.jar$"))))))
+      (build-system ant-build-system)
+      (arguments
+       `(#:jar-name (string-append ,name ".jar")
+         #:source-dir (string-append ,internal-name "-"
+                                     ,ver "/src/")
+         #:test-dir (string-append ,internal-name "-"
+                                   ,ver "/test/")
+         #:test-exclude (list "**/StAXTest.java"
+			      "**/NodeIteratorTest.java"
+			      "**/ScriptTest.java"
+			      "**/SegmentGetStyleURISegmentsTest.java"
+			      "**/SegmentTest.java"
+			      "**/StreamedSourceTest.java"
+			      "**/HTMLSanitiserTest.java")
+         #:phases (modify-phases %standard-phases
+                    (add-before 'build 'add-ant-env-options
+                      (lambda* _
+                        (setenv "ANT_OPTS" "-Dfile.encoding=iso-8859-1") #t))
+                    (add-after 'build 'check-prep
+                      (lambda* (#:key source #:allow-other-keys)
+                        (let* ((cwd (getcwd))
+			       (jericho-test (string-append cwd "/" ,internal-name
+						    "-" ,ver "/test"))
+			       (test (string-append jericho-test "/test"))
+			       (test-data (string-append jericho-test "/data"))
+			       (test-src (string-append jericho-test "/src"))
+			       (test-src-data (string-append test-src "/data")))
+                          (format #t "~%Check Prep dir: ~s\n" cwd)
+                          (substitute* "build.xml"
+                            (("\\$\\{test\\.home\\}/java")
+                             "${test.home}/src"))
+                          (mkdir-p (string-append cwd "/build/test-classes"))
+                          (copy-recursively (string-append cwd
+                                      "/"
+                                      ,internal-name
+                                      "-"
+                                      ,ver
+                                      "/samples/console/classes")
+                                     (string-append cwd
+                                      "/build/test-classes/"))
+			  (mkdir-p test-src-data)
+			  (copy-recursively test-data
+					    test-src-data))
+                        #t)))))
+      (native-inputs (list java-junit java-slf4j-api java-log4j-api
+                           java-commons-logging-minimal))
+      (home-page "http://jericho.htmlparser.net/docs/index.html")
+      (description
+       "A java library allowing analysis and manipulation of
+parts of an HTML document, including server-side tags, while
+reproducing verbatim any unrecognised or invalid HTML.  It also
+provides high-level HTML form manipulation functions.  
+")
+      (synopsis "Java HTML Parser library")
+      (license (list license:lgpl2.1+ license:asl2.0 license:epl1.0)))))
+
 (define-public java-simple-xml
   (package
     (name "java-simple-xml")

base-commit: 0e480ca7b60428a62fc4681d7aca6c7c067add42
-- 
2.38.1