From a22ea327fb80ffd5a490ae858be93f1dd9d1acf4 Mon Sep 17 00:00:00 2001
From: Jeremy Evans <code@jeremyevans.net>
Date: Tue, 26 Nov 2019 17:31:47 -0800
Subject: [PATCH] Add Zlib::GzipReader.each_file for handling multiple files in
 gz file

Most gzip tools support concatenated gz streams in a gz file. This
offers an easy way to handle such gz files in Ruby.

Fixes [Bug #9790]
Fixes [Bug #11180]
Fixes [Bug #14804]
---
 ext/zlib/zlib.c        | 44 ++++++++++++++++++++++++++++++++++++++++++
 test/zlib/test_zlib.rb | 20 +++++++++++++++++++
 2 files changed, 64 insertions(+)

diff --git a/ext/zlib/zlib.c b/ext/zlib/zlib.c
index 23466d1c94..b6d4a249b5 100644
--- a/ext/zlib/zlib.c
+++ b/ext/zlib/zlib.c
@@ -3723,6 +3723,49 @@ rb_gzreader_s_open(int argc, VALUE *argv, VALUE klass)
     return gzfile_s_open(argc, argv, klass, "rb");
 }
 
+/*
+ * Document-method: Zlib::GzipReader.each_file
+ *
+ * call-seq:
+ *   Zlib::GzipReader.each_file(io, options = {}, &block) => nil
+ *
+ * Many gzip handling utilities support multiple concatenated files
+ * in a single .gz file. Zlib::GzipReader instances only support
+ * a single compressed file in .gz file, starting at the current
+ * position in the .gz file.  Zlib::GzipReader.each_file yields
+ * Zlib::GzipReader instances, one per file in the .gz file.
+ * Example:
+ *
+ *   Zlib::GzipReader.each_file(io) do |gz|
+ *     puts gz.read
+ *   end
+ */
+static VALUE
+rb_gzreader_s_each_file(int argc, VALUE *argv, VALUE klass)
+{
+    VALUE io, unused, obj;
+    long pos;
+
+    rb_check_arity(argc, 1, 2);
+    io = argv[0];
+
+    do {
+        obj = rb_funcallv(klass, rb_intern("new"), argc, argv);
+        rb_yield(obj);
+
+        rb_gzreader_read(0, 0, obj);
+        pos = NUM2LONG(rb_funcall(io, rb_intern("pos"), 0));
+        unused = rb_gzreader_unused(obj);
+        rb_gzfile_finish(obj);
+        if (!NIL_P(unused)) {
+            pos -= NUM2LONG(rb_funcall(unused, rb_intern("length"), 0));
+            rb_funcall(io, rb_intern("pos="), 1, LONG2NUM(pos));
+        }
+    } while (pos < NUM2LONG(rb_funcall(io, rb_intern("size"), 0)));
+
+    return Qnil;
+}
+
 /*
  * Document-method: Zlib::GzipReader.new
  *
@@ -4696,6 +4739,7 @@ Init_zlib(void)
     rb_define_method(cGzipWriter, "puts", rb_gzwriter_puts, -1);
 
     rb_define_singleton_method(cGzipReader, "open", rb_gzreader_s_open,-1);
+    rb_define_singleton_method(cGzipReader, "each_file", rb_gzreader_s_each_file, -1);
     rb_define_alloc_func(cGzipReader, rb_gzreader_s_allocate);
     rb_define_method(cGzipReader, "initialize", rb_gzreader_initialize, -1);
     rb_define_method(cGzipReader, "rewind", rb_gzreader_rewind, 0);
diff --git a/test/zlib/test_zlib.rb b/test/zlib/test_zlib.rb
index 7d703d15e4..9b90342460 100644
--- a/test/zlib/test_zlib.rb
+++ b/test/zlib/test_zlib.rb
@@ -446,6 +446,26 @@ def test_set_dictionary
   end
 
   class TestZlibGzipFile < Test::Unit::TestCase
+    def test_gzip_reader_each_file
+      Tempfile.create("test_zlib_gzip_file_to_io") {|t|
+        gz = Zlib::GzipWriter.new(t)
+        gz.print("foo")
+        gz.close
+        t = File.open(t.path, 'ab')
+        gz = Zlib::GzipWriter.new(t)
+        gz.print("bar")
+        gz.close
+
+        results = []
+        t = File.open(t.path)
+        Zlib::GzipReader.each_file(t) do |f|
+          results << f.read
+        end
+        assert_equal(["foo", "bar"], results)
+        t.close
+      }
+    end
+
     def test_to_io
       Tempfile.create("test_zlib_gzip_file_to_io") {|t|
         t.close
-- 
2.23.0