Fwd: Fixing zipfile.BadZipfile to zipfile.BadZipFile

Forwarded conversation Subject: Fixing zipfile.BadZipfile to zipfile.BadZipFile ------------------------ From: *Boštjan Mejak* <bostjan.mejak@gmail.com> Date: Fri, Oct 15, 2010 at 11:02 PM To: python-dev@python.org I am very glad you're reorganizing the Standard Library. Thumbs up! I hope everything will comply to PEP 8 after you're done. Since you're reorganizing, I have my own contribution. I have attached a patch. The issue7351 <http://bugs.python.org/issue7351> was not accepted at the time, so I hope you'll accept this fix now. My point is that every class name in the module zipfile is like this: - LargeZipFile - ZipFile - PyZipFile . . . So apply my patch to make the class name BadZipfile consistent to other class names in the zipfile module and let it be named BadZipFile. Thank you. Best regards, Boštjan Mejak ---------- From: *Éric Araujo* <merwok@netwok.org> Date: Sun, Oct 17, 2010 at 11:00 PM To: python-dev@python.org Cc: Boštjan Mejak <bostjan.mejak@gmail.com> Hello (A bit of context: The original message comes from bug #2775, “Implement PEP 3108”, a meta-bug tracking stdlib reorganization for py3k.) You may have missed the timeline: Most of the PEP 3108 changes have been done before the first 3.x release went out. Now that we have 3.1 out as a stable and supported, we cannot reorganize and break compatibility anymore. (A note about PEP 8 compliance: Module names have been mostly fixed, but not all function/method names, for example in logging and unittest. If I recall correctly, readability did not seem to make all the rewrites worth it.)
a patch. The issue7351 <http://bugs.python.org/issue7351> was not I’ve just re-read the answers there and they are still valid.
Ezio and me: “Your patch need to include an alias (BadZipfile = BadZipFile) to preserve compatibility with old pickles, as explains msg95477.” Antoine: “I don't think changing it for the sake of aesthetics is a good deal given that many existing programs will have to be converted to the new spelling.” Regards ---------- From: *Boštjan Mejak* <bostjan.mejak@gmail.com> Date: Mon, Oct 18, 2010 at 9:25 AM To: Éric Araujo <merwok@netwok.org> Shoot. Well, too bad. I thought Python is all about readability, but I think you developers don't take it very seriously. If I was a developer, I would certainly go to the trouble of all the rewrittes of package/module/class/method/function names that do not comply to PEP 8 and have them done by the time the first sub-version of Python 3 would be released. ---------- From: *Éric Araujo* <merwok@netwok.org> Date: Mon, Oct 18, 2010 at 2:14 PM To: Boštjan Mejak <bostjan.mejak@gmail.com> Hi Could you repost your message to the mailing list, so that other people can see it and eventually react? Thanks. (P.S. please make sure not to top-post on the mailing list) Regards ---------- From: *Éric Araujo* <merwok@netwok.org> Date: Mon, Oct 18, 2010 at 2:45 PM To: python-dev@python.org Hello [Sorry if this comes twice, connection errors here]
a patch. The issue7351 <http://bugs.python.org/issue7351> was not
Python-Dev mailing list Python-Dev@python.org http://mail.python.org/mailman/listinfo/python-dev Unsubscribe: http://mail.python.org/mailman/options/python-dev/bostjan.mejak%40gmail.com ---------- From: *Nick Coghlan* <ncoghlan@gmail.com> Date: Mon, Oct 18, 2010 at 3:13 PM To: Éric Araujo <merwok@netwok.org> Cc: python-dev@python.org Correct. We went through this for one module that I recall (threading) and that was annoying enough that we mostly left things alone after that unless they were truly obnoxious. For threading we were able to clean a lot of things up in the process (such as adding properties where appropriate), but even so, we still made sure all the old names continued to work. Cheers, Nick. -- Nick Coghlan | ncoghlan@gmail.com | Brisbane, Australia ---------- From: *Boštjan Mejak* <bostjan.mejak@gmail.com> Date: Mon, Oct 18, 2010 at 4:36 PM To: Nick Coghlan <ncoghlan@gmail.com> Then please make an alias for this custom BadZipfile exception class. Add BadZipfile = BadZipFile in the zipfile module. ---------- From: *Boštjan Mejak* <bostjan.mejak@gmail.com> Date: Tue, Oct 19, 2010 at 10:34 PM To: Nick Coghlan <ncoghlan@gmail.com> Sorry, let me correct myself: Add BadZipFile = BadZipfile in the zipfile module. Please add the above assignment statement in the right place to the zipfile module and create an alias. Thank you. ---------- From: *Boštjan Mejak* <bostjan.mejak@gmail.com> Date: Wed, Oct 20, 2010 at 11:13 PM To: Nick Coghlan <ncoghlan@gmail.com> Ah, leave it as is. If in the future you will be able to rename BadZipfile to BadZipFile, please do. ---------- From: *Éric Araujo* <merwok@netwok.org> Date: Wed, Oct 27, 2010 at 6:26 PM To: python-dev <python-dev@python.org> Cc: Boštjan Mejak <bostjan.mejak@gmail.com> [Reply received off-list quoted here] Le 18/10/2010 09:25, Boštjan Mejak a écrit : Readability has to be balanced with other important things: maintainability, discoverability, compatibility, rapidity and other things ending in -ity. I’m personally glad that readability gets a huge place in the language definition itself (no braces, standard indentation), I follow 98 % of PEP 8 where I can, and have learned to accept that sometimes, a piece of code in the standard lib won’t get more readable. Still beats some languages that I won’t name because dissing Java is too easy. <wink> You are free to break compat in your code but it python-dev has a duty to its users. Renames in threading have been deemed a good thing, renames in unittest not. You can read the python-3000 archives to get an idea of the years of work and thousands of messages that went into py3k. At some point, a release had to be done. Now that there is a stable release in the 3.x line, compatibility rules apply. Practicality beats purity; now is better than never. Regards ---------- From: *Boštjan Mejak* <bostjan.mejak@gmail.com> Date: Wed, Oct 27, 2010 at 7:13 PM To: Éric Araujo <merwok@netwok.org> Since Python 3.2 accepts feature requests, take this fix as a feature request. Please forget about preserving the compatibility with old pickles. ---------- From: *Éric Araujo* <merwok@netwok.org> Date: Wed, Oct 27, 2010 at 7:16 PM To: Boštjan Mejak <bostjan.mejak@gmail.com> Again, please send a copy to your messages to the mailing list (choose “reply to all” or “reply to list”), and don’t leave the whole message after your reply. Thanks.

From: *Boštjan Mejak* <bostjan.mejak@gmail.com> Since Python 3.2 accepts feature requests, take this fix as a feature request. Please forget about preserving the compatibility with old pickles.
You can reopen #7351 as a feature request for 3.2. A serious proposal has to take compatibility into account, note. It is simply not possible to break everyone’s pickles for a very minor inconsistency in one name. There is a wide Python world out there. I’d like to suggest again that the time you want to give to Python (thanks!) would be better employed on something else than BadZipfile. For example, if you don’t think you can contribute code now, you could suggest improvements for unclear sections in the documentation. Recent reports sent by beginners can be found on the archives of the docs@python.org mailing list. Regards

Can anybody summarize the outcome? Is it that renaming BadZipfile to BadZipFile with backward compatible alias and deprecation note breaks something? -- anatoly t.

Le 28/10/2010 22:52, anatoly techtonik a écrit :
Can anybody summarize the outcome? Is it that renaming BadZipfile to BadZipFile with backward compatible alias and deprecation note breaks something?
See #7351 or r85874. Regards

The tests prove that r85874 does not break the build. On Fri, Oct 29, 2010 at 12:49 AM, Éric Araujo <merwok@netwok.org> wrote:
Le 28/10/2010 22:52, anatoly techtonik a écrit :
Can anybody summarize the outcome? Is it that renaming BadZipfile to BadZipFile with backward compatible alias and deprecation note breaks something?
See #7351 or r85874.
Regards
_______________________________________________ Python-Dev mailing list Python-Dev@python.org http://mail.python.org/mailman/listinfo/python-dev Unsubscribe: http://mail.python.org/mailman/options/python-dev/bostjan.mejak%40gmail.com
participants (3)
-
anatoly techtonik
-
Boštjan Mejak
-
Éric Araujo