raw_input breaking with multiple futurize passes #427
Labels
Comments
|
👍 |
guyer
added a commit
to guyer/fipy
that referenced
this issue
May 9, 2019
Undoes changes made in 98af1a6. If not careful, `futurize` converts `raw_input()` into `input()`, and then into `eval(input())`, which is VeryBad\TM. See PythonCharmers/python-future#427
guyer
added a commit
to usnistgov/fipy
that referenced
this issue
May 10, 2019
* futurize --fix lib2to3.fixes.fix_except * futurize_doctests lib2to3.fixes.fix_except * futurize --fix lib2to3.fixes.fix_exec * futurize --fix lib2to3.fixes.fix_funcattrs * futurize --fix lib2to3.fixes.fix_has_key * futurize --fix lib2to3.fixes.fix_idioms * futurize_doctests lib2to3.fixes.fix_idioms * futurize_doctests lib2to3.fixes.fix_numliterals * futurize --fix lib2to3.fixes.fix_reduce * futurize --fix lib2to3.fixes.fix_repr * futurize --fix lib2to3.fixes.fix_types * futurize --fix lib2to3.fixes.fix_ws_comma * futurize_doctests lib2to3.fixes.fix_ws_comma * futurize --fix libfuturize.fixes.fix_absolute_import * futurize --fix libfuturize.fixes.fix_print_with_import * futurize --fix libfuturize.fixes.fix_raise * 2to3 --doctests_only --fix import * 2to3 --doctests_only --fix print * Import print_function where needed for doctests * Scrub usage of 2to3 * Install libglu in test apt-get installations not shared between different job environments * futurize --fix lib2to3.fixes.fix_dict * futurize_doctests lib2to3.fixes.fix_dict * futurize --fix lib2to3.fixes.fix_map * futurize --fix lib2to3.fixes.fix_nonzero * futurize --fix lib2to3.fixes.fix_raw_input * futurize_doctests lib2to3.fixes.fix_raw_input * futurize --fix lib2to3.fixes.fix_zip * Alias __nonzero__ to __bool__ Recommended by https://portingguide.readthedocs.io/en/latest/core-obj-misc.html#customizing-truthiness-bool and get weird errors without it * futurize --fix libfuturize.fixes.fix_division_safe * futurize --fix libfuturize.fixes.fix_future_builtins * futurize_doctests libfuturize.fixes.fix_future_builtins * futurize --fix libfuturize.fixes.fix_future_standard_library * futurize --fix libfuturize.fixes.fix_object * futurize_doctests libfuturize.fixes.fix_xrange_with_import * futurize --fix libpasteurize.fixes.fix_newstyle * Specify what to futurizerize * Purge past.utils.old_div We always(?) know whether we're doing float or integer division * Import future * Fix fixes to future_builtins Fix native string typechecking Fix range as list * Pervasively switch to unicode_literals Resolves many inconsistent appearances of strings as sometimes unicode and sometimes not. Requires conversion of `__all__` entries because ["Python 2 names are strings, not unicode values"](https://stackoverflow.com/a/19913680/2019542) Force unit to render as native string Force .gz filenames to be native * Convert raw latex docstrings to double-backslashed latex docstrings Python 2 has [no way to represent `\u` in a raw unicode string literal](https://matplotlib.org/2.1.1/devel/portable_code.html#the-dreaded-u-escapes) * Reformat numpy output before run Reformatting was done too late to be effective * Discard futurization of versioneer * Ping the build * Fix typo * Strip whitespace * Undo bad integer division * Nativize command strings * Define _nativize_all locally in test command * Add future package to Nix build * Fix latex backslashing * Document need for `numpydoc` * Remove redundant type * Convert leading tabs to spaces * Add future package to Nix build * Remove sphinxext file that was never used(?) * Credit @pya and @woodscn * Comment on plans for Py3k development * Remove conversion scripts No longer needed now that conversion is done * Remove gratuitous (and dangerous) `eval` Undoes changes made in 98af1a6. If not careful, `futurize` converts `raw_input()` into `input()`, and then into `eval(input())`, which is VeryBad\TM. See PythonCharmers/python-future#427 * Remove only_directive We weren't using it and it's been removed in Py3k * Remove backlash line continuation that suddenly stopped working * Fix more line continuations
guyer
added a commit
to usnistgov/steppyngstounes
that referenced
this issue
Jan 26, 2021
* futurize --fix lib2to3.fixes.fix_except * futurize_doctests lib2to3.fixes.fix_except * futurize --fix lib2to3.fixes.fix_exec * futurize --fix lib2to3.fixes.fix_funcattrs * futurize --fix lib2to3.fixes.fix_has_key * futurize --fix lib2to3.fixes.fix_idioms * futurize_doctests lib2to3.fixes.fix_idioms * futurize_doctests lib2to3.fixes.fix_numliterals * futurize --fix lib2to3.fixes.fix_reduce * futurize --fix lib2to3.fixes.fix_repr * futurize --fix lib2to3.fixes.fix_types * futurize --fix lib2to3.fixes.fix_ws_comma * futurize_doctests lib2to3.fixes.fix_ws_comma * futurize --fix libfuturize.fixes.fix_absolute_import * futurize --fix libfuturize.fixes.fix_print_with_import * futurize --fix libfuturize.fixes.fix_raise * 2to3 --doctests_only --fix import * 2to3 --doctests_only --fix print * Import print_function where needed for doctests * Scrub usage of 2to3 * Install libglu in test apt-get installations not shared between different job environments * futurize --fix lib2to3.fixes.fix_dict * futurize_doctests lib2to3.fixes.fix_dict * futurize --fix lib2to3.fixes.fix_map * futurize --fix lib2to3.fixes.fix_nonzero * futurize --fix lib2to3.fixes.fix_raw_input * futurize_doctests lib2to3.fixes.fix_raw_input * futurize --fix lib2to3.fixes.fix_zip * Alias __nonzero__ to __bool__ Recommended by https://portingguide.readthedocs.io/en/latest/core-obj-misc.html#customizing-truthiness-bool and get weird errors without it * futurize --fix libfuturize.fixes.fix_division_safe * futurize --fix libfuturize.fixes.fix_future_builtins * futurize_doctests libfuturize.fixes.fix_future_builtins * futurize --fix libfuturize.fixes.fix_future_standard_library * futurize --fix libfuturize.fixes.fix_object * futurize_doctests libfuturize.fixes.fix_xrange_with_import * futurize --fix libpasteurize.fixes.fix_newstyle * Specify what to futurizerize * Purge past.utils.old_div We always(?) know whether we're doing float or integer division * Import future * Fix fixes to future_builtins Fix native string typechecking Fix range as list * Pervasively switch to unicode_literals Resolves many inconsistent appearances of strings as sometimes unicode and sometimes not. Requires conversion of `__all__` entries because ["Python 2 names are strings, not unicode values"](https://stackoverflow.com/a/19913680/2019542) Force unit to render as native string Force .gz filenames to be native * Convert raw latex docstrings to double-backslashed latex docstrings Python 2 has [no way to represent `\u` in a raw unicode string literal](https://matplotlib.org/2.1.1/devel/portable_code.html#the-dreaded-u-escapes) * Reformat numpy output before run Reformatting was done too late to be effective * Discard futurization of versioneer * Ping the build * Fix typo * Strip whitespace * Undo bad integer division * Nativize command strings * Define _nativize_all locally in test command * Add future package to Nix build * Fix latex backslashing * Document need for `numpydoc` * Remove redundant type * Convert leading tabs to spaces * Add future package to Nix build * Remove sphinxext file that was never used(?) * Credit @pya and @woodscn * Comment on plans for Py3k development * Remove conversion scripts No longer needed now that conversion is done * Remove gratuitous (and dangerous) `eval` Undoes changes made in 98af1a61. If not careful, `futurize` converts `raw_input()` into `input()`, and then into `eval(input())`, which is VeryBad\TM. See PythonCharmers/python-future#427 * Remove only_directive We weren't using it and it's been removed in Py3k * Remove backlash line continuation that suddenly stopped working * Fix more line continuations
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Let's start with a python file (test.py) containing a single line:
test = raw_input("Test: ")Pass 1: Running futurize.exe test.py --both-stages -w converts this to:
Pass 2: Running futurize.exe test.py --both-stages -w again converts this to:
The second pass is incorrect. The presence of "from builtins import input" should indicate that the bare input has already been corrected.
Note that --both-stages can be replaced with --fix lib2to3.fixes.fix_input --fix lib2to3.fixes.fix_raw_input --fix libfuturize.fixes.fix_future_builtins.
The text was updated successfully, but these errors were encountered: