1984
Comment:
|
2494
|
Deletions are marked like this. | Additions are marked like this. |
Line 3: | Line 3: |
Here are some possible tasks for the Py3k portion of the GoogleSprint. | Here are some Py3k tasks we could work on, are working on, have worked on at the GoogleSprint. |
Line 5: | Line 5: |
To claim a task, please add your name to it. | == Suggested Tasks == |
Line 7: | Line 7: |
* Fix various unit tests that are currently failing in the p3yk (sic) branch (all done). | To claim a task, please add your name to it and move it to the next section. |
Line 9: | Line 9: |
* Unify int/long (see recent email in python-3000 list; Martin von Löwis???) * Get rid of all references to file() in favor of open(). Anna has checked through Lib/g*.py and found no file() references that need to be changed. (only use of file in those are in docstrings or as variable names.) Anna will continue tomorrow on this task. (AMR) * Work on the new I/O library (I have much interest in this but need help -- Guido) |
* Make IDLE work -- there are relative import issues, perhaps others. |
Line 18: | Line 13: |
* Rewrite import in Python (Brett Cannon?, Alex Martelli?, Osvaldo Santana) * Rip out the last remains of coerce (Done, need to rename nb_coerce slot) |
|
Line 24: | Line 14: |
* Rip out classic classes (already done; Guido) * Make it so that __hash__ can be set to None to declare an unhashable type, and to automatically do this the first time __eq__ or __cmp__ are overridden but __hash__ is not (done) * Implement PEP PEP:3102 (keyword-only arguments) (Jiwon Seo?) |
|
Line 33: | Line 17: |
* Make map() and filter() iterators and make them stop at the end of the shortest input (like zip()) instead of at the end of the longest input (like itertools?) |
|
Line 35: | Line 21: |
* Make zip() an iterator (like itertools.izip()) | |
Line 37: | Line 22: |
* Make map() and filter() iterators and make them stop at the end of the shortest input (like zip()) instead of at the end of the longest input (like itertools?) | == Claimed Tasks == |
Line 39: | Line 24: |
* Rip out reduce() (done -- John Reese and Jacques Frechet) | * Work on the new I/O library (I have much interest in this but need help -- Guido); you can check http://sebulba.wikispaces.com/project+iostack+v2 for a reference on what was generally agreed in the py3k list. (Hasan, Charles) * Unify int/long (Martin von Löwis; see http://mail.python.org/pipermail/python-3000/2006-August/003046.html) * Implement PEP PEP:3102 (keyword-only arguments) (Jiwon Seo, Brett C.) * Rewrite import in Python (Brett Cannon?, Alex Martelli?, Osvaldo Santana) |
Line 43: | Line 34: |
* Kill uses of has_key() in idlelib (John R & Jacques F). | * Remove basestring.find and basestring.rfind, per [http://www.python.org/dev/peps/pep-3100/ PEP 3100] (done, Hasan -- needs to be reviewed and checked in) == Finished Tasks == * Make zip() an iterator (like itertools.izip()) (done; Guido) * Clean up comparisons. Remove the ability to order objects where it doesn't make sense. (Guido, Alex) * Fix various unit tests that are currently failing in the p3yk branch (done, various sprinters). * Rip out the last remains of coerce (done, Neal; need to rename nb_coerce slot) * Make it so that __hash__ can be set to None to declare an unhashable type, and to automatically do this the first time __eq__ or __cmp__ are overridden but __hash__ is not (done, Guido & Alex) * Rip out reduce() (done, John Reese and Jacques Frechet) * Kill uses of has_key() in idlelib (done, John R & Jacques F). * Get rid of all references to file() in favor of open(). (done, Anna) * Rip out classic classes (done pre-sprint, Guido) |
Python 3000 Goals for the Google Sprint
Here are some Py3k tasks we could work on, are working on, have worked on at the GoogleSprint.
Suggested Tasks
To claim a task, please add your name to it and move it to the next section.
- Make IDLE work -- there are relative import issues, perhaps others.
- When the I/O library is ready, start unifying str/unicode.
Put back nb_ <true division> slots (regular and augmented)
Py3kConversionTools: Work on tool to help convert (flag keywords like with/as, removed functions/methods: apply/{}.iter*/{}.has_key/etc) (Jeremy Hylton)
- Make map() and filter() iterators and make them stop at the end of the shortest input (like zip()) instead of at the end of the longest input (like itertools?)
See PEP 3100 for more ideas
Claimed Tasks
Work on the new I/O library (I have much interest in this but need help -- Guido); you can check http://sebulba.wikispaces.com/project+iostack+v2 for a reference on what was generally agreed in the py3k list. (Hasan, Charles)
Unify int/long (Martin von Löwis; see http://mail.python.org/pipermail/python-3000/2006-August/003046.html)
Implement PEP 3102 (keyword-only arguments) (Jiwon Seo, Brett C.)
- Rewrite import in Python (Brett Cannon?, Alex Martelli?, Osvaldo Santana)
Make xrange() support longs; then rename it to range() (see http://www.python.org/sf/1472639 for all but the supporting-long part) (Neal)
Remove basestring.find and basestring.rfind, per [http://www.python.org/dev/peps/pep-3100/ PEP 3100] (done, Hasan -- needs to be reviewed and checked in)
Finished Tasks
- Make zip() an iterator (like itertools.izip()) (done; Guido)
- Clean up comparisons. Remove the ability to order objects where it doesn't make sense. (Guido, Alex)
- Fix various unit tests that are currently failing in the p3yk branch (done, various sprinters).
- Rip out the last remains of coerce (done, Neal; need to rename nb_coerce slot)
Make it so that hash can be set to None to declare an unhashable type, and to automatically do this the first time eq or cmp are overridden but hash is not (done, Guido & Alex)
- Rip out reduce() (done, John Reese and Jacques Frechet)
Kill uses of has_key() in idlelib (done, John R & Jacques F).
- Get rid of all references to file() in favor of open(). (done, Anna)
- Rip out classic classes (done pre-sprint, Guido)