Background
Some day, I found an error in RSS page generated by Hugo.
I tried to parse the source post and found an invisible character in it on GitHub.
I found there was a backspace character ^H in the file and it is not the only one in the only file.
Find All Posts and Execute On Them
Basically, This is our idea that using the command find to get all filenames and execute the command we want on them. But don’t forget to use -print0 to send them to xargs by the pipline.
1 | find ./content/posts -name "*.md" -print0 | xargs -0 wc -l |
As we see, it works.
Replace Invisible Character by Sed
Now we can use sed to replace all ^H and ^M in our posts. Pay attention to the character like ^H in command, they must be typed with such as ctrl+v+h.
1 | find ./content/posts -name "*.md" -print0 | xargs -0 sed -i '' -e 's/^H//g; s/^M//g; s/^\//g' |
Notice that we use an empty character after -i in the command instead of sed -i -e because of the character after the option -i of sed in Mac is used as the postfix of backup files. We need to use an empty character to tell it not to save a backup one.
Result
After refreshing the page I opened, everything done!
Reference
- This answer on StackOverflow provide the idea. But not work.
- This answer shows how to execute on all files.
- This one asked how to type control characters in command line.
- The comments of this answer explain how to use
sedfor replacing in Mac.