Fixed an issue where a server error was raised when downloading a PDF for an invalid article identifier

[gamboz]

URLs in the form /article/<identifier_type>/<identifier>/download/pdf/ call a view that tries to get the article using the identifier.

If the article does not exist (maybe because there was a typo in the identifier), or if the article has not yet been published, the view should return a 404.

This MR tries to fix a small bug that makes the clause above to fail badly when the article does not exists, because, ATM, the code checks if the non-existing article has been published 🙂

Unfortunately I wasn't able to replicate the issue on any online system where I tried. For instance, both the following URLs return a neat 404, but AFAICT the first should work and the second should give 500:

• https://olh.openlibhums.org/article/doi/10.16995%2Folh.16937/download/pdf
• https://olh.openlibhums.org/article/doi/10.16995%2Folh.16937xxxxxxxxxxxxxxxx/download/pdf

So, it's very possible I'm missing something here (i.e. please feel free to trash this MR).

[ajrbyers]

@gamboz in this instance if article_object is none the second part of the if statement is not evaluated so it immediately raises the 404. You can test this by doing something like this in a python shell:

In [1]: article = None

In [2]: if not article or test:
   ...:     print('Article is None so even though test is never set it wont break.')
   ...: 
Article is None so even though test is never set it wont break.

In [3]: 

[gamboz]

@ajrbyers I agree, but the original code has if not article and... (with and, not or).

[ajrbyers]

@ajrbyers I agree, but the original code has if not article and... (with and, not or).

Haha, of course. Stupid me, I was looking at your new version XD

[mauromsl]

URLs in the form /article/<identifier_type>/<identifier>/download/pdf/ call a view that tries to get the article using the identifier.

If the article does not exist (maybe because there was a typo in the identifier), or if the article has not yet been published, the view should return a 404.

This MR tries to fix a small bug that makes the clause above to fail badly when the article does not exists, because, ATM, the code checks if the non-existing article has been published 🙂

Unfortunately I wasn't able to replicate the issue on any online system where I tried. For instance, both the following URLs return a neat 404, but AFAICT the first should work and the second should give 500:

* https://olh.openlibhums.org/article/doi/10.16995%2Folh.16937/download/pdf

* https://olh.openlibhums.org/article/doi/10.16995%2Folh.16937xxxxxxxxxxxxxxxx/download/pdf

So, it's very possible I'm missing something here (i.e. please feel free to trash this MR).

Provided examples do not reproduce it because they are missing a trailing slash, I can reproduce it like this: https://olh.openlibhums.org/article/id/999999999/download/pdf/