GeezerVison is showing github.com — Derby #8: Convert 28 sites to Argument Clinic across 2 files · Issue #64376 · python/cpython
Derby #8: Convert 28 sites to Argument Clinic across 2 files #64376Description
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 = None
closed_at = None
created_at = <Date 2014-01-07.23:49:34.309>
labels = ['extension-modules', 'type-feature', 'expert-argument-clinic', '3.9']
title = 'Derby #8: Convert 28 sites to Argument Clinic across 2 files'
updated_at = <Date 2020-05-19.06:28:09.667>
user = 'https://github.com/larryhastings' bugs.python.org fields:
activity = <Date 2020-05-19.06:28:09.667>
actor = 'hauntsaninja'
assignee = 'none'
closed = False
closed_date = None
closer = None
components = ['Extension Modules', 'Argument Clinic']
creation = <Date 2014-01-07.23:49:34.309>
creator = 'larry'
dependencies = []
files = ['33534', '33607', '33716', '33717']
hgrepos = []
issue_num = 20177
keywords = ['patch']
message_count = 21.0
messages = ['207622', '207652', '207654', '208383', '208391', '208424', '208425', '208426', '208429', '208439', '208441', '208746', '209153', '209268', '209270', '218647', '219051', '224758', '336195', '336556', '346297']
nosy_count = 6.0
nosy_names = ['larry', 'petr.viktorin', 'nikratio', 'serhiy.storchaka', 'ZackerySpytz', 'hauntsaninja']
pr_nums = ['8535', '14311', '14330', '20208']
priority = 'normal'
resolution = None
stage = 'patch review'
status = 'open'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue20177'
versions = ['Python 3.9'] Activitylarryhastings commented on Jan 7, 2014 ContributorAuthor More actions skrah commented on Jan 8, 2014 Please take out _decimal.c. It has a huge test suite that would break entirely. larryhastings commented on Jan 8, 2014 ContributorAuthor More actions Absolutely. decimal is a very special case.
Whoever takes over this issue: please ignore the part about _decimal.c. changed the title [-]Derby #8: Convert 52 sites to Argument Clinic across 3 files[/-] [+]Derby #8: Convert 28 sites to Argument Clinic across 2 files[/+] on Jan 8, 2014nikratio commented on Jan 18, 2014 I have started working on this (partial patch attached), but I am unsure how to handle cases where the docstring definition utilizes a macro. For example:
PyDoc_STRVAR(strftime_doc,
"strftime(format[, tuple]) -> string\n\
\n\
Convert a time tuple to a string according to a format specification.\n\
See the library reference manual for formatting codes. When the time tuple\n\
is not present, current time as returned by localtime() is used.\n\
\n" STRFTIME_FORMAT_CODES);
I don't think STRFTIME_FORMAT_CODES will be expanded if I put it into the [clinic] comment section.
Should I include the contents verbatim? Or do not convert these functions to argument clinic? larryhastings commented on Jan 18, 2014 ContributorAuthor More actions That macro, STRFTIME_FORMAT_CODES, is only used in two functions. If it were me I'd just copy and paste the text into both functions. nikratio commented on Jan 18, 2014 Will do. Another question: what should I do with return annotations, e.g. the "-> dict" part in
"get_clock_info(name: str) -> dict\n
Should I drop that part? In the clinic howto, I couldn't find anything about return annotations (only that parameter annotations are unsupported). nikratio commented on Jan 18, 2014 One more question: what's a good way to find sites that need changes? Searching for "PyArg_" finds both converted and unconverted functions... larryhastings commented on Jan 18, 2014 ContributorAuthor More actions Yes, you should drop things that look like return annotations in the original docstring.
I don't have a magic formula for finding unchanged functions, sorry. nikratio commented on Jan 19, 2014 Here is a patch for timemodule.c. I have converted everything except:
-
ctime() uses the gettmarg() function. Converting gettmarg() itself doesn't seem possible to me (because it's not Python callable), and duplicating the functionality in a single complicated clinic section just for ctime() did not seem reasonable either.
-
gmtime(), localtime(), ctime() use a the parse_time_t_args() utility function. I have not yet figured out how to replace this with argument clinic (sent a mail to python-dev about that).
nikratio commented on Jan 19, 2014 Hmm. After reading some of the threads on python-dev, I'm now confused if I did the conversion of optional arguments correctly.
Should something like "strftime(format[, tuple])" be converted using optional groups, or should tuple become a parameter with a default value? (so far, I have done the latter). nikratio commented on Jan 25, 2014 (I already sent this to python-dev, but maybe it makes more sense to have these thoughts together with the patch)
After looking at the conversion of parse_time_t_args again, I think I lost track of what we're actually gaining with this procedure. If all we need is a type that can distinguish between a valid time_t value and a default value, why don't we simply use PyObject?
In other words, what's the advantage of the extra custom strict, plus the custom C converter function, plus the custom converter python class over:
static int
PyObject_to_time_t(PyObject *obj, time_t *when)
{
if (obj == NULL || obj == Py_None)
*when = time(NULL);
else if (_PyTime_ObjectToTime_t(obj, when) == -1)
return 0;
return 1;
}/*[clinic input]
time.gmtime
[...]
static PyObject *
time_gmtime_impl(PyModuleDef *module, PyObject *seconds)
{
PyObject *return_value = NULL;
time_t when;
if(!PyObj_to_time_t(seconds, &when))
return NULL;[...]
To me this version looks shorter and clearer. nikratio commented on Jan 26, 2014 I've attached updated patches to reflect the recent changes in the derby policy.
part1 is the part of the patch that is suitable for 3.4.
part2 are the changes that have to wait for 3.5 (mostly because of default values). nikratio commented on Jan 26, 2014 I'll wait with working on _datetimemodule.c until someone had a chance to look over my _timemodule.c patch. (I want to make sure I'm following the right procedure). taleinat commented on May 16, 2014 Regarding the handling of time_t arguments which can be None, I agree that the second version (without custom convertors) is simpler and clearer while having no disadvantage that I can see.
I'd like to review the rest of the patches, but you mention changes to the AC policy regarding 3.4 and 3.5 and I can't find a reference to these changes or to the new policy. nikratio commented on May 24, 2014 larryhastings commented on Aug 4, 2014 ContributorAuthor More actions All the Derby patches should only go into trunk at this point. serhiy-storchaka commented on Feb 21, 2019 The modern workflow requires creating a PR on GitHub. Nikolaus, do you mind to convert your patch to a PR? nikratio commented on Feb 25, 2019 Sorry, no. I have long lost context and interest in this. ZackerySpytz commented on Jun 22, 2019
MetadataMetadataDevelopmentNo branches or pull requestsIssue actions
|