pygettext doesn't work with f-strings

[RiccardoPolignieri]

BPO	33189
Nosy	@malemburg, @scoder, @ericvsmith, @merwok, @serhiy-storchaka, @miss-islington
PRs	bpo-33189: pygettext.py now accepts only literal strings #6364 [3.7] bpo-33189: pygettext.py now accepts only literal strings (GH-6364) #6533 [3.6] bpo-33189: pygettext.py now accepts only literal strings (GH-6364) #6534 [3.6] bpo-33189: pygettext.py now accepts only literal strings (GH-6364) #6535

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = 'https://github.com/serhiy-storchaka'
closed_at = <Date 2018-04-19.07:48:45.584>
created_at = <Date 2018-03-30.20:25:16.363>
labels = ['3.8', 'type-bug', '3.7']
title = "pygettext doesn't work with f-strings"
updated_at = <Date 2018-04-19.07:48:45.584>
user = 'https://bugs.python.org/RiccardoPolignieri'

bugs.python.org fields:

activity = <Date 2018-04-19.07:48:45.584>
actor = 'serhiy.storchaka'
assignee = 'serhiy.storchaka'
closed = True
closed_date = <Date 2018-04-19.07:48:45.584>
closer = 'serhiy.storchaka'
components = ['Demos and Tools']
creation = <Date 2018-03-30.20:25:16.363>
creator = 'Riccardo Polignieri'
dependencies = []
files = []
hgrepos = []
issue_num = 33189
keywords = ['patch']
message_count = 8.0
messages = ['314712', '314776', '314780', '314878', '315389', '315481', '315482', '315483']
nosy_count = 7.0
nosy_names = ['lemburg', 'scoder', 'eric.smith', 'eric.araujo', 'serhiy.storchaka', 'miss-islington', 'Riccardo Polignieri']
pr_nums = ['6364', '6533', '6534', '6535']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue33189'
versions = ['Python 3.6', 'Python 3.7', 'Python 3.8']

[RiccardoPolignieri]

Tested (on windows) with python 3.6, but I guess it's the same in py3.7:

# test.py
def hello(x):
  print(_(f'hello {x}'))

> py pygettext.py test.py
Traceback (most recent call last):
  File "C:\Program Files\Python36\Tools\i18n\pygettext.py", line 623, in <module>
    if __name__ == '__main__':
  File "C:\Program Files\Python36\Tools\i18n\pygettext.py", line 597, in main
    for _token in tokens:
  File "C:\Program Files\Python36\Tools\i18n\pygettext.py", line 328, in __call__
    ##              'tstring:', tstring
  File "C:\Program Files\Python36\Tools\i18n\pygettext.py", line 382, in __openseen
    elif ttype == tokenize.STRING:
  File "C:\Program Files\Python36\Tools\i18n\pygettext.py", line 236, in safe_eval
    # unwrap quotes, safely
  File "<string>", line 1, in <module>
NameError: name 'x' is not defined

[scoder]

Is there really a use case for this?

I would normally expect place holders to get replaced *after* translation, i.e. in the translated text, not before.

[serhiy-storchaka]

I agree that pygettext should not and can not support f-strings, but the error should be better. The problem is that an f-string expression is tokenized as literal strings, but eval() can't evaluate it out of context.

[RiccardoPolignieri]

the error should be better

Yes, that's what I maeant - sorry I should have phrased better.
I marked by mistake a couple of f-strings for translation and totally forgot about it for a while... it took me some time to figure out what was going on with this NameError popping out of nowhere.

Curiously, xgettext handles this thing better (...or, far worse actually): it just parses the f-string as a regular one, producing something like
msgid "hello {x}"
msgstr ""
But then of course Python will fail *silently* at runtime, never picking up the translated version of the string...

[serhiy-storchaka]

If there are no objections I'm going to merge PR 6364 in 24 hours.

[serhiy-storchaka]

New changeset 6952482 by Serhiy Storchaka in branch 'master':
bpo-33189: pygettext.py now accepts only literal strings (GH-6364)
6952482

[miss-islington]

New changeset a4fb580 by Miss Islington (bot) in branch '3.7':
bpo-33189: pygettext.py now accepts only literal strings (GH-6364)
a4fb580

[miss-islington]

New changeset 6f87093 by Miss Islington (bot) in branch '3.6':
bpo-33189: pygettext.py now accepts only literal strings (GH-6364)
6f87093