New upstream version 2.9.0.upstream/2.9.0

8 files changed, 43 insertions, 28 deletions

diff --git a/lib/roo/base.rb b/lib/roo/base.rb
index 19eb844..f4ac9a3 100644
--- a/lib/roo/base.rb
+++ b/lib/roo/base.rb
@@ -544,7 +544,7 @@ class Roo::Base
     tempfilename = File.join(tmpdir, find_basename(uri))
     begin
       File.open(tempfilename, "wb") do |file|
-        open(uri, "User-Agent" => "Ruby/#{RUBY_VERSION}") do |net|
+        URI.open(uri, "User-Agent" => "Ruby/#{RUBY_VERSION}") do |net|
           file.write(net.read)
         end
       end
diff --git a/lib/roo/csv.rb b/lib/roo/csv.rb
index 516def6..4431bc2 100644
--- a/lib/roo/csv.rb
+++ b/lib/roo/csv.rb
@@ -90,17 +90,23 @@ module Roo
     def each_row(options, &block)
       if uri?(filename)
         each_row_using_tempdir(options, &block)
-      elsif is_stream?(filename_or_stream)
-        ::CSV.new(filename_or_stream, options).each(&block)
       else
-        ::CSV.foreach(filename, options, &block)
+        csv_foreach(filename_or_stream, options, &block)
       end
     end
 
     def each_row_using_tempdir(options, &block)
       ::Dir.mktmpdir(Roo::TEMP_PREFIX, ENV["ROO_TMP"]) do |tmpdir|
         tmp_filename = download_uri(filename, tmpdir)
-        ::CSV.foreach(tmp_filename, options, &block)
+        csv_foreach(tmp_filename, options, &block)
+      end
+    end
+
+    def csv_foreach(path_or_io, options, &block)
+      if is_stream?(path_or_io)
+        ::CSV.new(path_or_io, **options).each(&block)
+      else
+        ::CSV.foreach(path_or_io, **options, &block)
       end
     end
 
diff --git a/lib/roo/excelx.rb b/lib/roo/excelx.rb
index f9f0ee2..91ebc1e 100755
--- a/lib/roo/excelx.rb
+++ b/lib/roo/excelx.rb
@@ -60,15 +60,16 @@ module Roo
       @filename = local_filename(filename_or_stream, @tmpdir, packed)
       process_zipfile(@filename || filename_or_stream)
 
-      @sheet_names = workbook.sheets.map do |sheet|
-        unless options[:only_visible_sheets] && sheet['state'] == 'hidden'
-          sheet['name']
-        end
-      end.compact
+      @sheet_names = []
       @sheets = []
       @sheets_by_name = {}
-      @sheet_names.each_with_index do |sheet_name, n|
-        @sheets_by_name[sheet_name] = @sheets[n] = Sheet.new(sheet_name, @shared, n, sheet_options)
+
+      workbook.sheets.each_with_index do |sheet, index|
+        next if options[:only_visible_sheets] && sheet['state'] == 'hidden'
+
+        sheet_name = sheet['name']
+        @sheet_names << sheet_name
+        @sheets_by_name[sheet_name] = @sheets[index] = Sheet.new(sheet_name, @shared, index, sheet_options)
       end
 
       if cell_max
@@ -428,6 +429,11 @@ module Roo
       entries.each do |entry|
         path =
         case entry.name.downcase
+        when /richdata/
+          # FIXME: Ignore richData as parsing is not implemented yet and can cause
+          # Zip::DestinationFileExistsError when including a second "styles.xml" entry
+          # see http://schemas.microsoft.com/office/spreadsheetml/2017/richdata2
+          nil
         when /sharedstrings.xml$/
           "#{@tmpdir}/roo_sharedStrings.xml"
         when /styles.xml$/
diff --git a/lib/roo/excelx/cell/number.rb b/lib/roo/excelx/cell/number.rb
index 9f23c4f..7ea48b5 100644
--- a/lib/roo/excelx/cell/number.rb
+++ b/lib/roo/excelx/cell/number.rb
@@ -48,7 +48,7 @@ module Roo
           when /^(0+)$/ then "%0#{$1.size}d"
           when /^0\.(0+)$/ then "%.#{$1.size}f"
           when '#,##0' then number_format('%.0f')
-          when '#,##0.00' then number_format('%.2f')
+          when /^#,##0.(0+)$/ then number_format("%.#{$1.size}f")
           when '0%'
             proc do |number|
               Kernel.format('%d%%', number.to_f * 100)
@@ -64,6 +64,7 @@ module Roo
           when '#,##0.00;[Red](#,##0.00)' then number_format('%.2f', '[Red](%.2f)')
             # FIXME: not quite sure what the format should look like in this case.
           when '##0.0E+0' then '%.1E'
+          when "_-* #,##0.00\\ _€_-;\\-* #,##0.00\\ _€_-;_-* \"-\"??\\ _€_-;_-@_-" then number_format('%.2f', '-%.2f')
           when '@' then proc { |number| number }
           else
             raise "Unknown format: #{format.inspect}"
diff --git a/lib/roo/excelx/cell/time.rb b/lib/roo/excelx/cell/time.rb
index a1f0864..5fed1e2 100644
--- a/lib/roo/excelx/cell/time.rb
+++ b/lib/roo/excelx/cell/time.rb
@@ -13,7 +13,7 @@ module Roo
           super
           @format = excelx_type.last
           @datetime = create_datetime(base_date, value)
-          @value = link ? Roo::Link.new(link, value) : (value.to_f * 86_400).to_i
+          @value = link ? Roo::Link.new(link, value) : (value.to_f * 86_400).round.to_i
         end
 
         def formatted_value
diff --git a/lib/roo/excelx/sheet_doc.rb b/lib/roo/excelx/sheet_doc.rb
index 6da4c26..adbb77a 100755
--- a/lib/roo/excelx/sheet_doc.rb
+++ b/lib/roo/excelx/sheet_doc.rb
@@ -101,12 +101,7 @@ module Roo
         cell_xml_children.each do |cell|
           case cell.name
           when 'is'
-            content = +""
-            cell.children.each do |inline_str|
-              if inline_str.name == 't'
-                content << inline_str.content
-              end
-            end
+            content = cell.search('t').map(&:content).join
             unless content.empty?
               return Excelx::Cell.cell_class(:string).new(content, formula, style, hyperlink, coordinate)
             end
@@ -197,11 +192,12 @@ module Roo
         # Extract merged ranges from xml
         merges = {}
         doc.xpath('/worksheet/mergeCells/mergeCell').each do |mergecell_xml|
-          tl, br = mergecell_xml["ref"].split(/:/).map { |ref| ::Roo::Utils.ref_to_key(ref) }
-          for row in tl[0]..br[0] do
-            for col in tl[1]..br[1] do
-              next if row == tl[0] && col == tl[1]
-              merges[[row, col]] = tl
+          src, dst = mergecell_xml["ref"].split(/:/).map { |ref| ::Roo::Utils.ref_to_key(ref) }
+          next unless cells[src]
+          for row in src[0]..dst[0] do
+            for col in src[1]..dst[1] do
+              next if row == src[0] && col == src[1]
+              merges[[row, col]] = src
             end
           end
         end
diff --git a/lib/roo/spreadsheet.rb b/lib/roo/spreadsheet.rb
index cdc93f0..54063b6 100644
--- a/lib/roo/spreadsheet.rb
+++ b/lib/roo/spreadsheet.rb
@@ -24,8 +24,14 @@ module Roo
           options[:file_warning] = :ignore
           extension.tr('.', '').downcase.to_sym
         else
-          res = ::File.extname((path =~ /\A#{::URI::DEFAULT_PARSER.make_regexp}\z/) ? ::URI.parse(::URI.encode(path)).path : path)
-          res.tr('.', '').downcase.to_sym
+          parsed_path =
+            if path =~ /\A#{::URI::DEFAULT_PARSER.make_regexp}\z/
+              # path is 7th match
+              Regexp.last_match[7]
+            else
+              path
+            end
+          ::File.extname(parsed_path).tr('.', '').downcase.to_sym
         end
       end
     end
diff --git a/lib/roo/version.rb b/lib/roo/version.rb
index 262847c..d2bb2bf 100644
--- a/lib/roo/version.rb
+++ b/lib/roo/version.rb
@@ -1,3 +1,3 @@
 module Roo
-  VERSION = "2.8.3"
+  VERSION = "2.9.0"
 end