Closed hachacha closed 5 years ago
Can you confirm that the sandbox has the latest version of pan doc? This error occurs a lot with the older 1.x version...
Sent from my iPhone
On 4 Mar 2019, at 16:47, jon kirichacha notifications@github.com wrote:
I was consistently having a problem with bs4 trying to read the file that was opened at this line as ascii on the sandbox server of for CMU's janeway instance.
Adding .encoding("utf-8") to the prettify() function was injecting \n through the document. Unfortunately, I was unable to recreate the same bug on my local but this line is just being more explicit.
You can view, comment on, or merge this pull request online at:
https://github.com/ajrbyers/pandoc_plugin/pull/11
Commit Summary
explicit encoding utf8 with open File Changes
M views.py (2) Patch Links:
https://github.com/ajrbyers/pandoc_plugin/pull/11.patch https://github.com/ajrbyers/pandoc_plugin/pull/11.diff — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.
Yes, I did an exhaustive check. When the file is converted it's all fine, but when the stdout is being opened by bs4 it's reading as ascii for some reason. I get the error at bs4 not pandoc.... I checked the versions of bs4, python, and pandoc and they're all up-to-date. the sys.getdefaultencoding() also prints utf8 for python on the server.
$ pandoc -v
pandoc 2.5
$ sudo apt-get purge pandoc
...Removing pandoc (2.5-1) ......
$ pandoc -v
The program 'pandoc' is currently not installed....
$ wget 'https://github.com/jgm/pandoc/releases/download/2.5/pandoc-2.5-1-amd64.deb'
$ sudo dpkg -i pandoc-2.5-1-amd64.deb
.......
$ pandoc -v
pandoc 2.5
After this I still get the error.
Thanks for being so thorough. I wonder if we should put together a testing package of simple docs with different encoding and expected outcomes for this plugin?
I was consistently having a problem with bs4 trying to read the file that was opened at this line as ascii on the sandbox server of for CMU's janeway instance. Adding .encoding("utf-8") to the
prettify()
function was injecting\n
through the document. Unfortunately, I was unable to recreate the same bug on my local but this line is just being more explicit.