wtfpython (in which I rant about writexml)

python's writexml is just wrong. want proof? here it goes...

$ cat print.py
#!/usr/bin/python3
import sys
import xml.dom.minidom
test = xml.dom.minidom.parse(sys.argv[1])
test.writexml(sys.stdout, indent='', addindent='  ', newl='\n', encoding="UTF-8")

Let's run this over and over with the same XML? A simply one, like...

$ cat 0.xml
<out><one/><two/></out>

Let's run it 10 times:

for i in {0..9}; do ./print.py $i".xml" > $((i + 1)).xml; done

The result should be all more or less the same, right? I mean, the first one should be prettier than the input, because we're identing it properly, but after that nothing else needs changing...

Let's see how did the first run look:

$ cat 1.xml
<?xml version="1.0" encoding="UTF-8"?>
<out>
  <one/>
  <two/>
</out>

There you go, looks great! What about the others?

1.xml 2.xml 3.xml 4.xml 5.xml 6.xml 7.xml 8.xml 9.xml 10.xml

¯\_(ツ)_/¯

For what is worth, these scripts, in case copyright applies to them, are licensed as GPLv3.