[issue37475] What is urllib.request.urlcleanup() function?

[Karthikeyan Singaravelan]

Karthikeyan Singaravelan <tir.karthi at gmail.com> added the comment:

I have added Senthil for thoughts. The temporary files is also stored in a module level variable. Looking into the git history there were some changes in e24f96a05973ddbb59d88c03570aef8545c5ef10 . The function is also marked as legacy in __all__ with compat code in 2to3. urlcleanup also sets the global variable _opener to None so it does the extra work of uninstalling the global variable opener set by install_opener which is also not documented.

urlretrieve enables retrieving and storing the content in a temporary file to return (tempfilename, headers) to be read later. In case the user doesn't give a filename it implicitly creates a temporary file. There is similar code in urllib.request.URLopener().retrieve [0] too where the temporary files are created implicitly but __del__ is overridden where the temp files are deleted as the program exits. I think it's better to ask the user to give filename and so that the user is responsible for the file but since the behavior is present for a long time there is backwards compatibility in changing this and there might be some code depending on the implicit temporary file created as a feature. 

One possible way would be to have a wrapper class that creates the temporary file when not given and then deletes it or calls urlcleanup on __del__ to clean it up as the GC is called like URLopener.retrieve? This would be done only when user doesn't give a file and for the temporary files generated by urlretrieve.


[0] https://github.com/python/cpython/blob/67310023f299b5a2fad71fca449b46d280036690/Lib/urllib/request.py#L1702

----------

_______________________________________
Python tracker <report at bugs.python.org>
<https://bugs.python.org/issue37475>
_______________________________________