Using the following test tree folder for example:
- test1
- folder2
- test1 # This is the file rubyzip will break on.
- test2
And copied this code from here:
path = File.expand_path path
archive = File.join(__dir__, File.basename(path)) + '.zip'
FileUtils.rm archive, force: true
Zip::File.open(archive, Zip::File::CREATE) do | zipfile |
Dir["#{path}/**/**"].reject{|f|f==archive}.each do | item |
basename = File.basename(item)
zipfile.add(basename, item)
end
end
It fails because there is two files having the same name even if their are not in the same directory (test1
in my example).
Is there something I am missing ?
Thanks to @simonoff (here), I shouldn't use the basename but the full relative path so Rubyzip could make the difference between test1
and folder2/test1
.
Here is a code fixing it:
basename = File.basename path
dirname = File.dirname path
internal_path = path.sub %r[^#{__dir__}/], ''
archive = File.join(dirname, basename) + '.zip'
FileUtils.rm archive, force: true
Zip::File.open(archive, Zip::File::CREATE) do | zipfile |
Dir["#{internal_path}/**/**"].map{|e|e.sub %r[^#{internal_path}/],''}.reject{|f|f==archive}.each do | item |
zipfile.add(item, File.join(internal_path, item))
end
end
There is certainly a much cleaner way to do it.