Blog post output paths are wrong

[candlerb]

Using the default my-awesome-site jekyll site, with

pandoc:
  skip: false
  output: ./tmp
  outputs:
    pdf:
    epub:

Jekyll creates the required parent directory, e.g. ./tmp/pdf/jekyll/update/2013/11/20, but then tries to write to ./tmp/pdf/jekyll/update/2013/11/20/jekyll/update/2013/11/20/welcome-to-jekyll.pdf instead of ./tmp/pdf/jekyll/update/2013/11/20/welcome-to-jekyll.pdf, and hence fails.

$ jekyll build
Configuration file: /Users/brian/tmp/my-awesome-site/_config.yml
            Source: /Users/brian/tmp/my-awesome-site
       Destination: /Users/brian/tmp/my-awesome-site/_site
      Generating... Creating ./tmp/pdf/jekyll/update/2013/11/20
pandoc  -o ./tmp/pdf/jekyll/update/2013/11/20/jekyll/update/2013/11/20/welcome-to-jekyll.pdf
pandoc: ./tmp/pdf/jekyll/update/2013/11/20/jekyll/update/2013/11/20/welcome-to-jekyll.pdf: openBinaryFile: does not exist (No such file or directory)
Creating ./tmp/epub/jekyll/update/2013/11/20
pandoc  -o ./tmp/epub/jekyll/update/2013/11/20/jekyll/update/2013/11/20/welcome-to-jekyll.epub
pandoc: ./tmp/epub/jekyll/update/2013/11/20/jekyll/update/2013/11/20/welcome-to-jekyll.epub: openBinaryFile: does not exist (No such file or directory)

done.

Tested with jekyll 1.3.0 under ruby 2.0.0p247

[candlerb]

Here is the result of adding some STDERR.puts lines

base_dir="./tmp"
output="pdf"
File.dirname(post.url)="/jekyll/update/2013/11/20"
post_path="./tmp/pdf/jekyll/update/2013/11/20"
Creating ./tmp/pdf/jekyll/update/2013/11/20
filename="/jekyll/update/2013/11/20/welcome-to-jekyll.pdf"
filename_with_path="./tmp/pdf/jekyll/update/2013/11/20/jekyll/update/2013/11/20/welcome-to-jekyll.pdf"
pandoc  -o ./tmp/pdf/jekyll/update/2013/11/20/jekyll/update/2013/11/20/welcome-to-jekyll.pdf
pandoc: ./tmp/pdf/jekyll/update/2013/11/20/jekyll/update/2013/11/20/welcome-to-jekyll.pdf: openBinaryFile: does not exist (No such file or directory)

This simple fix makes it work:

--- a/lib/jekyll-pandoc-multiple-formats.rb
+++ b/lib/jekyll-pandoc-multiple-formats.rb
@@ -27,7 +27,7 @@ class PandocGenerator < Generator
         # Have a filename!
         filename = "#{post.url.gsub(/\//, "-").gsub(/-$/, "")}.#{output}" if filename =~ /\/$/

-        filename_with_path = File.join(post_path, filename)
+        filename_with_path = File.join(base_dir, output, filename)

         # Special cases, stdout is disabled for these
         if ['pdf', 'epub', 'odt', 'docx'].include?(output)

but I can't see how it could have worked before without this.

[fauno]

thanks! if you do the pull request(s) i'll merge your changes :)

[vhuber]

Is there any pull request done @fauno @candlerb ?

[candlerb]

This is someone else's code which, as I said, I don't understand why it was written the way it was. I think the author needs to make a decision as to whether the frig patch I made is in fact the correct solution, or there is something else underlying which needs to be changed.

[fauno]

the patch is already applied

https://github.com/fauno/jekyll-pandoc-multiple-formats/blob/master/lib/jekyll-pandoc-multiple-formats.rb#L21

i guess since it was so simple i just went ahead and applied it

http://librevpn.org.ar