Enum: modify __repr__, __str__; update docs #84247
Comments
|
Serhiy had the idea of having Enum._convert also modify the __str__ and __repr__ of newly created enumerations to display the module name instead of the enumeration name (https://bugs.python.org/msg325007): --> socket.AF_UNIX --> print(socket.AF_UNIX) Thoughts? |
ethanfurman
added
type-feature
Yes. And repr should be inverse of eval, but it's probably too late for that. :-/ |
|
Looks like the >>> import re
>>> re.I
re.IGNORECASE
>>> print(re.I)
# should also be re.IGNORECASE
>>> re.I|re.S|re.X
re.IGNORECASE|re.DOTALL|re.VERBOSEFor stdlib Enum conversions are we happy with that? Or should __str__ just print the numeric value? |
|
If it's considered to be not too backwards-incompatible, I think it would be nice to have str different from repr. That way we can finetune what exactly we need. But we can already do almost exactly that with *int* instead of *str*, so it's not too compelling. Much more important thing is the "repr as inverse of eval". Is there any way we can have that for our own enums (as a mixin or a decorator)? @module_global(re)
class RegexFlag(Enum):
...It would be fantastic. :-) |
|
"repr as inverse of eval" is nice to have, but it is not a requirement. |
|
Noone said it is a requirement, I just said it would be nice to have it factored out as a decorator or something instead of having to write __repr__ over and over again. |
|
At this point, the PR has made the following changes:
|
|
Python-Dev thread [0], summary below:
After discussions with Guido I made a (largely done) PR [3] which: for stdlib global constants (such as RE)
for stdlib non-global constants, and enums in general
The questions I would most appreciate an answer to at this point:
As a reminder, the underlying issue is trying to keep at least the stdlib Enum representations the same for those that are replacing preexisting constants. [1] I'm working on making their creation faster. If anyone wanted to convert EnumMeta to C I would be grateful. [2] https://bugs.python.org/issue36548 [3] #22392 |
Absolutely.
Well, standard backward compatibility stuff. :-) |
ethanfurman
changed the title
It breaks doctests, and probably some other unit tests, too, e.g. for output formatting. What should we suggest users to do? Replace >>> get_flag(…)
<app_enums.TrickyFlag: 1>by >>> get_flag(…) == app_enums.TrickyFlag or get_flag(…) # (also show result on failure)
Trueand assertEqual(
"You caught the %s flag!" % result,
"You caught the app_enums.TrickyFlag flag!")by ? Note that using "%r" does not help, since it's also backwards incompatible. For their own enums, users can probably backport (or forward-port) "__str__()" and "__repr__()" themselves in order to work around this difference. But it's something they would have to do. I certainly understand the reasoning, and it also makes new Py3.10-only doctests nicer, actually, but IMHO it counts as deliberate breakage. |
|
I'm probably a bit late to the party as well, but as some additional datapoints:
|
|
Thank you for the feedback. The new str() and repr() make more sense for Flag-based enumerations, and I'm hesitant to have different formats for Enum- vs Flag-based enums. Would it be helpful to have another base Enum class to derive from that maintained the original str() and repr() formats? |
|
It may be helpful for the enum module to come with transitional functions like "pre310_str()" "pre310_repr()" "pre310_flag_str()" etc. so that people who are writing doctests that need to function on both < 3.10 and 3.10+ can temporarily do a "Enum.__str__ = pre310_str" in their test suites (and of course restore it later) until <=3.9 is no longer supported. Otherwise everyone with doctest suites will be doing this ourselves. |
|
#74767 broke doc tests in 3.11 branch: File "library/enum.rst", line ?, in default
Failed example:
Color(0)
Exception raised:
Traceback (most recent call last):
File "/home/runner/work/cpython/cpython/Lib/doctest.py", line 1346, in __run
exec(compile(example.source, filename, "single",
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "<doctest default[0]>", line 1, in <module>
Color(0)
^^^^^
NameError: name 'Color' is not defined |
|
This also broke our Solaris build with the following error: ====================================================================== Traceback (most recent call last):
File "/..../cpython-main/Lib/test/test_socket.py", line 1523, in testGetaddrinfo
self.assertEqual(repr(type), '<SocketKind.SOCK_STREAM: 1>')
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: '<SocketKind.SOCK_STREAM: 2>' != '<SocketKind.SOCK_STREAM: 1>'
- <SocketKind.SOCK_STREAM: 2>
? ^
+ <SocketKind.SOCK_STREAM: 1>
? ^(test.test_socket.GeneralModuleTests fails with the same error). The issue is almost certainly that on Solaris, SOCK_STREAM is defined as 2 rather than 1; the following simple program confirms that: #include <stdio.h>
#include <sys/socket.h>
void main() {
printf("%d\n", SOCK_STREAM);
}I'm just not sure whether to fix this with |
|
Sorry, I had to revert the change since it broke the CI and it prevented to merge new PRs. Tell me if I can help to get this test fixed and to get this change merged again. By the way, the PR 30582 was merged even if the Docs CI failed. |
|
After merging in doc fix by kumaraditya303, I'll update tests so Solaris passes. |
For this one, I suggest to replace the value with "..." doctest pattern. |
|
vstinner wrote:
That bit of code is from the unittest suite, not the doctest suite. I went with: |
|
I created python/core-workflow#424 "Should we make the Docs CI mandatory on the Python main branch?". |
Python 3.11 has changed behavior of __str__ for IntEnum. Before it would
return the name, like "Color.RED", now it's equivalent to
str(self.value), for example "1".__repr__ is unchanged [1].
Python 3.11 also introduced StrEnum, which we can use instead of mixing
in str type for our StrEnumDefinition, and that introcudes the same
change of __str__ behavior: it returns the value as string.
The corresponds, in spirit, to what I often did in Python:
class Color(Enum):
RED="red"
BLUE="blue"
def __str__(self):
return self.value
To ensure consistent behavior for different Python versions, and
because we can determine our own rules for the enum types we define, we
mimic the 3.11 __str__ behavior on IntEnumDefinition and
StrEnumDefinition.
We use the same trick, by setting the __str__ property to the version of
the underlying type: str.__str__ and int.__repr__. int.__str__ won't do
because int directly inherits object.__str__ that basically repr(self),
and thus back to Enum.__repr__. So we need int.__repr__ to return the
integer as string.
[1] python/cpython#84247
support python3.11 python/cpython#100458 python/cpython#84247
repr()andstr()#22392Note: 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:
bugs.python.org fields:
The text was updated successfully, but these errors were encountered: