Mailman 3 October 2018 - Python-Dev

AIX to stable, what does that take?
by Michael Felt 05 Oct '18

05 Oct '18

In the buildbots AIX is marked as "unstable"? What is needed to get it marked as a "stable" platform - that being one of my short-term goals. My assumption is that it needs to (at least) pass all tests - and that is why I keep asking for attention. All the PRs to fix individual tests mean less if they are not merged, for whatever reason. However, maybe there is another way, or even something additional needed. Maybe something I cannot provide and then I can adjust my expectations and goals. Regards, Michael

4 4

Summary of Python tracker Issues
by Python tracker 05 Oct '18

05 Oct '18

ACTIVITY SUMMARY (2018-09-28 - 2018-10-05) Python tracker at https://bugs.python.org/ To view or respond to any of the issues listed below, click on the issue. Do NOT respond to this message. Issues counts and deltas: open 6813 (+32) closed 39845 (+42) total 46658 (+74) Open issues with patches: 2722 Issues opened (53) ================== #18291: codecs.open interprets FS, RS, GS as line ends https://bugs.python.org/issue18291 reopened by serhiy.storchaka #31310: semaphore tracker isn't protected against crashes https://bugs.python.org/issue31310 reopened by serhiy.storchaka #32117: Tuple unpacking in return and yield statements https://bugs.python.org/issue32117 reopened by serhiy.storchaka #34811: test_gdb fails with latest gdb https://bugs.python.org/issue34811 reopened by vstinner #34834: test_ssl.test_options does not correctly account for built-in https://bugs.python.org/issue34834 opened by xnox #34836: test_ssl.test_default_ecdh_curve needs no tls1.3 flag in 2.7, https://bugs.python.org/issue34836 opened by xnox #34837: Multiprocessing.pool API Extension - Pass Data to Workers w/o https://bugs.python.org/issue34837 opened by seanharr11 #34838: Improve arg clinic code generation for cases with type checkin https://bugs.python.org/issue34838 opened by rhettinger #34839: doctest: Change example under warnings section https://bugs.python.org/issue34839 opened by cheryl.sabella #34840: dlopen() error with no error message from dlerror() https://bugs.python.org/issue34840 opened by shuoz #34841: Script’s directory not in sys.path with embeddable Windows d https://bugs.python.org/issue34841 opened by ssapin #34844: logging.Formatter enhancement - Checking on style and fmt fiel https://bugs.python.org/issue34844 opened by BNMetrics #34846: Runtime failure with Failed to import site module https://bugs.python.org/issue34846 opened by sat.tho(a)gmail.com #34847: asyncio: Add PHA for TLS 1.3 https://bugs.python.org/issue34847 opened by fantix #34848: range.index only takes one argument when it's documented as ta https://bugs.python.org/issue34848 opened by bup #34849: Drop logging when asyncio waits in selector.select() https://bugs.python.org/issue34849 opened by asvetlov #34850: Emit a syntax warning for "is" with a literal https://bugs.python.org/issue34850 opened by serhiy.storchaka #34852: Counter-intuitive behavior of Server.close() / wait_closed() https://bugs.python.org/issue34852 opened by aymeric.augustin #34855: batch file variables https://bugs.python.org/issue34855 opened by lindblad #34856: Make the repr of lambda containing the signature and body expr https://bugs.python.org/issue34856 opened by serhiy.storchaka #34857: IDLE: SyntaxWarning not handled properly https://bugs.python.org/issue34857 opened by terry.reedy #34858: MappingProxy objects should JSON serialize just like a diction https://bugs.python.org/issue34858 opened by Michael Smith2 #34861: Improve cProfile standard output https://bugs.python.org/issue34861 opened by Anders.Hovmöller #34864: In Idle, Mac tabs make editor status line disappear. https://bugs.python.org/issue34864 opened by andyharrington #34866: CGI DOS vulnerability via long post list https://bugs.python.org/issue34866 opened by Matthew Belisle #34867: Add mode to disable small integer and interned string caches https://bugs.python.org/issue34867 opened by steven.daprano #34870: Core dump when Python VSCode debugger is attached https://bugs.python.org/issue34870 opened by Per Lundberg #34872: investigate task/future cancellation in asynciomodule.c https://bugs.python.org/issue34872 opened by yselivanov #34876: Python3.8 changes how decorators are traced https://bugs.python.org/issue34876 opened by nedbat #34877: Inconsistent Behavior Of futures.ProcessPoolExecutor https://bugs.python.org/issue34877 opened by TensorTom #34880: About the "assert" bytecode https://bugs.python.org/issue34880 opened by vtheno athena #34881: unnecessary encoded-words usage breaks DKIM signatures https://bugs.python.org/issue34881 opened by bryced #34882: f(a=1, *args) should be a SyntaxError https://bugs.python.org/issue34882 opened by metaxm #34883: test_lzma: Multiple test failures when liblzma is built withou https://bugs.python.org/issue34883 opened by mgorny #34884: Python loads incorrect libraries https://bugs.python.org/issue34884 opened by Tim Hutt #34885: asyncio documention has lost its paragraph about cancellation https://bugs.python.org/issue34885 opened by abki #34886: subprocess.run throws exception when input and stdin are passe https://bugs.python.org/issue34886 opened by aecant #34888: Python3.8 optimizes away a "while" line https://bugs.python.org/issue34888 opened by nedbat #34890: Support functools.partial in inspect.is*function() checks https://bugs.python.org/issue34890 opened by asvetlov #34891: Multi-processing example inaccurate warning https://bugs.python.org/issue34891 opened by anthony-flury #34893: Add 2to3 fixer to change send and recv methods of socket objec https://bugs.python.org/issue34893 opened by devarakondapranav #34895: Mark optional stdlib modules in documentation https://bugs.python.org/issue34895 opened by of4tvziy #34897: distutils test errors when CXX is not set https://bugs.python.org/issue34897 opened by Michael.Felt #34898: add mtime argument to gzip.compress https://bugs.python.org/issue34898 opened by guoci #34899: Possible assertion failure due to int_from_bytes_impl() https://bugs.python.org/issue34899 opened by ZackerySpytz #34900: unittest subTests() fails when called from debug() https://bugs.python.org/issue34900 opened by Bruno Oliveira #34901: Missing isolated (-I) flag in sys.flags table https://bugs.python.org/issue34901 opened by danishprakash #34902: Azure pipelines PR build fails with "Unexpected vmImage 'vs201 https://bugs.python.org/issue34902 opened by xtreak #34903: strptime %d handling of single digit day of month https://bugs.python.org/issue34903 opened by Mike Gleen #34904: Crash in ZipFile.close() when writing zip file to /dev/null https://bugs.python.org/issue34904 opened by erik.bray #34905: Cannot assign memoryview values from array.array https://bugs.python.org/issue34905 opened by aparamon #34906: Fix typo in the documentation https://bugs.python.org/issue34906 opened by matrixise #34907: calculation not working properly https://bugs.python.org/issue34907 opened by hwk_un1te Most recent 15 issues with no replies (15) ========================================== #34907: calculation not working properly https://bugs.python.org/issue34907 #34905: Cannot assign memoryview values from array.array https://bugs.python.org/issue34905 #34902: Azure pipelines PR build fails with "Unexpected vmImage 'vs201 https://bugs.python.org/issue34902 #34899: Possible assertion failure due to int_from_bytes_impl() https://bugs.python.org/issue34899 #34898: add mtime argument to gzip.compress https://bugs.python.org/issue34898 #34897: distutils test errors when CXX is not set https://bugs.python.org/issue34897 #34895: Mark optional stdlib modules in documentation https://bugs.python.org/issue34895 #34884: Python loads incorrect libraries https://bugs.python.org/issue34884 #34883: test_lzma: Multiple test failures when liblzma is built withou https://bugs.python.org/issue34883 #34877: Inconsistent Behavior Of futures.ProcessPoolExecutor https://bugs.python.org/issue34877 #34870: Core dump when Python VSCode debugger is attached https://bugs.python.org/issue34870 #34866: CGI DOS vulnerability via long post list https://bugs.python.org/issue34866 #34861: Improve cProfile standard output https://bugs.python.org/issue34861 #34857: IDLE: SyntaxWarning not handled properly https://bugs.python.org/issue34857 #34855: batch file variables https://bugs.python.org/issue34855 Most recent 15 issues waiting for review (15) ============================================= #34906: Fix typo in the documentation https://bugs.python.org/issue34906 #34901: Missing isolated (-I) flag in sys.flags table https://bugs.python.org/issue34901 #34900: unittest subTests() fails when called from debug() https://bugs.python.org/issue34900 #34899: Possible assertion failure due to int_from_bytes_impl() https://bugs.python.org/issue34899 #34898: add mtime argument to gzip.compress https://bugs.python.org/issue34898 #34897: distutils test errors when CXX is not set https://bugs.python.org/issue34897 #34872: investigate task/future cancellation in asynciomodule.c https://bugs.python.org/issue34872 #34866: CGI DOS vulnerability via long post list https://bugs.python.org/issue34866 #34861: Improve cProfile standard output https://bugs.python.org/issue34861 #34856: Make the repr of lambda containing the signature and body expr https://bugs.python.org/issue34856 #34850: Emit a syntax warning for "is" with a literal https://bugs.python.org/issue34850 #34849: Drop logging when asyncio waits in selector.select() https://bugs.python.org/issue34849 #34847: asyncio: Add PHA for TLS 1.3 https://bugs.python.org/issue34847 #34844: logging.Formatter enhancement - Checking on style and fmt fiel https://bugs.python.org/issue34844 #34838: Improve arg clinic code generation for cases with type checkin https://bugs.python.org/issue34838 Top 10 most discussed issues (10) ================================= #34751: Hash collisions for tuples https://bugs.python.org/issue34751 39 msgs #34850: Emit a syntax warning for "is" with a literal https://bugs.python.org/issue34850 21 msgs #18291: codecs.open interprets FS, RS, GS as line ends https://bugs.python.org/issue18291 11 msgs #34711: Lib/http/server.py: Return HTTPStatus.NOT_FOUND if path.endswi https://bugs.python.org/issue34711 10 msgs #34880: About the "assert" bytecode https://bugs.python.org/issue34880 10 msgs #34313: IDLE crashes with Tk-related error on macOS with ActiveTcl 8.6 https://bugs.python.org/issue34313 9 msgs #34872: investigate task/future cancellation in asynciomodule.c https://bugs.python.org/issue34872 9 msgs #34812: [EASY] support.args_from_interpreter_flags() doesn't inherit - https://bugs.python.org/issue34812 8 msgs #34867: Add mode to disable small integer and interned string caches https://bugs.python.org/issue34867 8 msgs #34014: loop.run_in_executor should propagate current contextvars https://bugs.python.org/issue34014 7 msgs Issues closed (42) ================== #25300: Enable Intel MPX (Memory protection Extensions) feature https://bugs.python.org/issue25300 closed by benjamin.peterson #26005: Denial of Service in SimpleHTTPServer and BaseHTTPServer https://bugs.python.org/issue26005 closed by martin.panter #28441: Change sys.executable to include executable suffix https://bugs.python.org/issue28441 closed by inada.naoki #29551: unittest: TestSuite.debug() does not like subTest() https://bugs.python.org/issue29551 closed by berker.peksag #30156: PYTHONDUMPREFS segfaults on exit https://bugs.python.org/issue30156 closed by vstinner #31660: sys.executable different in os.execv'd python3.6 virtualenv se https://bugs.python.org/issue31660 closed by cheryl.sabella #31865: html.unescape does not work as per documentation https://bugs.python.org/issue31865 closed by ezio.melotti #32756: argparse: parse_known_args: raising exception on unknown arg f https://bugs.python.org/issue32756 closed by paul.j3 #32833: argparse doesn't recognise two option aliases as equal https://bugs.python.org/issue32833 closed by paul.j3 #33117: asyncio example uses non-existing/documented method https://bugs.python.org/issue33117 closed by asvetlov #34172: multiprocessing.Pool and ThreadPool leak resources after being https://bugs.python.org/issue34172 closed by pitrou #34476: asyncio.sleep(0) not documented https://bugs.python.org/issue34476 closed by asvetlov #34739: Get rid of tp_getattro in xml.etree.ElementTree.XMLParser https://bugs.python.org/issue34739 closed by serhiy.storchaka #34740: Get rid of tp_getattro in ossaudiodev.oss_audio_device https://bugs.python.org/issue34740 closed by serhiy.storchaka #34779: IDLE internals show up in tracebacks when returning objects th https://bugs.python.org/issue34779 closed by terry.reedy #34797: Convert heapq to the argument clinic https://bugs.python.org/issue34797 closed by rhettinger #34801: codecs.getreader() splits lines containing control characters https://bugs.python.org/issue34801 closed by serhiy.storchaka #34835: Multiprocessing module update fails with pip3 https://bugs.python.org/issue34835 closed by steven.daprano #34842: Incorrect error messages in bisect https://bugs.python.org/issue34842 closed by rhettinger #34843: logging cookbook docs: remove 'recent' when referring to multi https://bugs.python.org/issue34843 closed by rhettinger #34845: allow exprlist as the iterators of comprehensions to be consis https://bugs.python.org/issue34845 closed by gvanrossum #34851: re.error - for the search function in the re module https://bugs.python.org/issue34851 closed by ronaldoussoren #34853: Python django cors https://bugs.python.org/issue34853 closed by ammar2 #34854: Crash in string annotations in lambda with keyword-only argume https://bugs.python.org/issue34854 closed by serhiy.storchaka #34859: python core in string substring search https://bugs.python.org/issue34859 closed by ammar2 #34860: fix test_sqlite for AIX https://bugs.python.org/issue34860 closed by Michael.Felt #34862: No longer builds on OpenBSD due to missing definition of conve https://bugs.python.org/issue34862 closed by benjamin.peterson #34863: Idle Mac scrolling only down https://bugs.python.org/issue34863 closed by terry.reedy #34865: Incorrect assignment of optional argument when partial match w https://bugs.python.org/issue34865 closed by eric.smith #34868: bad error message when combining _ grouping specifier with inv https://bugs.python.org/issue34868 closed by benjamin.peterson #34869: remove LDLAST https://bugs.python.org/issue34869 closed by benjamin.peterson #34871: test_site fails if run after test_inspect https://bugs.python.org/issue34871 closed by yselivanov #34873: re.finditer behaviour in re.MULTILINE mode fails to match firs https://bugs.python.org/issue34873 closed by tdawes #34874: Python 3.6.3 command script wrapped in single quotes produces https://bugs.python.org/issue34874 closed by Tim McDonough #34875: Change .js mime to "text/javascript" https://bugs.python.org/issue34875 closed by asvetlov #34878: Lock Objects documentation bug https://bugs.python.org/issue34878 closed by benjamin.peterson #34879: bytesobject.c: Possible null pointer dereference due to format https://bugs.python.org/issue34879 closed by serhiy.storchaka #34887: bytes subclass __repr__ raise SystemError when set to bytes.de https://bugs.python.org/issue34887 closed by benjamin.peterson #34889: int.to_bytes and int.from_bytes should default to the system b https://bugs.python.org/issue34889 closed by benjamin.peterson #34892: persistence of attributes with new instance https://bugs.python.org/issue34892 closed by serhiy.storchaka #34894: Unexpected error while unpickling lxml.etree.Element object https://bugs.python.org/issue34894 closed by serhiy.storchaka #34896: Unable to install Python 3.5 https://bugs.python.org/issue34896 closed by zach.ware

1 0

Petr Viktorin as BDFL-Delegate for PEP 580
by Jeroen Demeyer 04 Oct '18

04 Oct '18

Hello, I would like to propose Petr Viktorin as BDFL-Delegate for PEP 580, titled "The C call protocol". He has co-authored several PEPs (PEP 394, PEP 489, PEP 534, PEP 547, PEP 573), several of which involve extension modules. Petr has agreed to become BDFL-Delegate for PEP 580 if asked. Also Antoine Pitrou, INADA Naoki and Nick Coghlan have approved Petr being BDFL-Delegate. I am well aware of the current governance issues, but several people have mentioned that the BDFL-Delegate process can still continue for now. I created a PR for the peps repository at https://github.com/python/peps/pull/797 Cheers, Jeroen Demeyer.

10 16

dear core-devs
by Michael Felt 04 Oct '18

04 Oct '18

Dear core-devs, I have some bad characteristics. I can be extremely enthusiastic - and write too much. I have been trying to not write - anything - worried that my enthusiasm is not matched by yours, or worse was a reason to ignore my work to get AIX passing all tests. FYI: since the end of July I have dedicated 16 to 24 hours of my free time to get this done. All for Python; all in my freetime. My employer does not care - I do, or did. I am grateful to Martin Panter - who helped me graciously when I knew absolutely nothing when I first got started; Victor was kind enough to answer some emails and help me along but also clear that he has zero interest in AIX and my questions were taking too much of his time. Regretfully for me. Again - Victor - thank you for your time. I appreciated the assistance and feedback. (Others have helped from time to time, my apologies for not naming you specifically.) I am, to put it lightly, extremely frustrated, at this point. I am sorry, for myself obviously - but also for Python. Obviously, I am doing it all wrong - as I see lots of other issues being picked up immediately. All volunteers need some level of recognition to keep moving on. And, while you may not give a damn about anything other than Windows, macos and/or Linux - there are other platforms that would like a stable Python. Sincerely, Michael

9 18

Re: [Python-Dev] Petr Viktorin as BDFL-Delegate for PEP 580
by Jeroen Demeyer 04 Oct '18

04 Oct '18

On 2018-10-03 23:27, Guido van Rossum wrote: > IMO changes to the C API should be taken just as seriously -- the > potential for breaking the world is just about the same (since most > serious Python applications use C extensions that risk breaking). Of course we are taking this seriously. I want this to be taken as seriously as any other PEP and any other BDFL-Delegate appointment in the past. To be clear: I'm not trying to rush my PEP though. It has been discussed and I have made changes to it based on comments. In fact, this is the second PEP with the same subject, I withdrew the first one, PEP 575. At some point in the past I asked one person to become BDFL-Delegate but he did not answer. And now recently Petr Viktorin made some insightful comments on it, so I asked him and he agreed.

1 0

Re: [Python-Dev] Petr Viktorin as BDFL-Delegate for PEP 580
by Jeroen Demeyer 04 Oct '18

04 Oct '18

On 2018-10-03 18:55, Barry Warsaw wrote: > Correct. It’s entirely possible that the different governance models will have different ways to pick delegates. And how does that affect *today*'s decision? The new governance model will only take effect 1 January (assuming that everything goes as planned). As long as there is no new governance model yet, can't we just continue the PEP 1 process which has worked for many years? I know that we cannot literally apply PEP 1 because there is no BDFL, but we can certainly continue the spirit of PEP 1 if the other core developers agree with the BDFL-Delegate.

1 0

Re: [Python-Dev] Petr Viktorin as BDFL-Delegate for PEP 580
by Jeroen Demeyer 03 Oct '18

03 Oct '18

On 2018-10-03 17:12, Wes Turner wrote: > AFAIU, there is not yet a documented process for BDFL-delegate assignment. PEP 1 says: """ However, whenever a new PEP is put forward, any core developer that believes they are suitably experienced to make the final decision on that PEP may offer to serve as the BDFL's delegate (or "PEP czar") for that PEP. If their self-nomination is accepted by the other core developers and the BDFL, then they will have the authority to approve (or reject) that PEP. """ I know that it says "core developers and the BDFL". However, if the core developers agree that Petr can become BDFL-Delegate, I don't see why that wouldn't be possible. Jeroen.

1 0

Change in Python 3's "round" behavior
by jab＠math.brown.edu 02 Oct '18

02 Oct '18

I recently found out about Python 3's round-to-even change (via https://github.com/cosmologicon/pywat!) and am having trouble finding where that change was discussed. I did find the revealingly-invalid bug report https://bugs.python.org/issue32956 ("python 3 round bug"), so I asked there, but wanted to invite anyone else on this list who might be interested to help. If interested, please see the comments there (copy/pasted below for convenience), and +nosy or comment on that issue. Thanks! Joshua Bronson added the comment: This was so surprising to me that I had to check some other languages that I had handy. It turns out that not one of JavaScript, Ruby, Perl, C++, Java, Go, or Rust agrees with Python. In fact they all agreed with one another that 2.5 should round to 3. Examples below. I understand from https://github.com/cosmologicon/pywat/pull/40#discussion_r219962259 that "to always round up... can theoretically skew the data" but it's not clear why that's a good enough reason to differ from the "round" function in all these other languages (as opposed to e.g. offering this alternative behavior in some additional "round_unskewed" function). I assume the rationale for having Python 3's "round" differ from that of so many other languages was written down when this decision was made, but I searched and couldn't find it. Could anyone link to it in a comment here? And would it be worth including rationale and a larger callout in the https://docs.python.org/3/library/functions.html#round docs? The documentation of this behavior is a bit buried among other things, and the rationale for it is missing entirely. $ node -e 'console.log(Math.round(2.5))' 3 $ ruby -e 'puts (2.5).round()' 3 $ perl -e 'use Math::Round; print round(2.5)' 3 $ cat test_round.cpp #include <stdio.h> #include <math.h> int main(void) { printf("%f\n", round(2.5)); } $ g++ test_round.cpp && ./a.out 3.000000 $ cat TestRound.java class TestRound { public static void main(String[] args) { System.out.println(Math.round(2.5)); } } $ javac TestRound.java && java TestRound 3 $ cat test_round.go package main import "fmt" import "math" func main() { fmt.Println(math.Round(2.5)) } $ go build test_round.go && ./test_round 3 $ cat test_round.rs fn main() { println!("{}", (2.5_f64).round()); } $ rustc test_round.rs && ./test_round 3 Serhiy Storchaka added the comment: See the discussion on the Python-Dev mailing list: https://mail.python.org/pipermail/python-dev/2008-January/075863.html. For C look at the rint() function. It is a common knowledge that rounding half-to-even is what users want in most cases, but it is a tiny bit more expensive in C. In Python the additional cost of such rounding is insignificant. Joshua Bronson added the comment: Thanks Serhiy, I read the Python-Dev thread you linked to, but that doesn't resolve the issues: - Its topic is Python 2.6 (where this behavior does not occur) rather than Python 3 (where it does). - A few messages into the thread Guido does address Python 3, but in fact says "I think the consensus is against round-to-even in 3.0" (see https://mail.python.org/pipermail/python-dev/2008-January/075897.html). - There is no discussion of the fact that this behavior differs from the function named "round" in all the other programming languages I mentioned, and whether it would therefore be better exposed as an additional function (e.g. "round_to_even" or "round_unbiased", and in the math or statistics package rather than builtins). Surprisingly, Excel is the only other programming environment I saw discussed in the thread. (And round(2.5) == 3 there.) So that all suggests there must be some other thread or issue where this change for Python 3 have been discussed, but I looked again and could not find it. The C "rint" example you gave just seems to prove the point that this behavior should have a distinct name from "round". Regarding: > It is a common knowledge that rounding half-to-even is what users want in most cases I don't think that's common knowledge; seems like citation needed? Based on all the other languages where this differs (not to mention Python 2), it's not clear users would want Python 3 to be the only different one. And this is definitely a surprise for the majority of programmers, whose experience with "round" is how it works everywhere else. (This is making it into pywat after all: https://github.com/cosmologicon/pywat/pull/40) I can submit a PR for at least updating the docs about this (as per my previous comment) if that would be welcomed.

11 20

LDLAST variable in configure.ac
by Michael Felt 02 Oct '18

02 Oct '18

Hi all, Before I submit a patch to increase the default MAXDATA setting for AIX when in 32-bit mode - I want to know if I can put this LDFLAG setting in LDLAST, or if I should introduce a new AC_SUBST() variable (e.g., LDMAXDATA). I have not looked yet, but I was thinking that MAYBE! LDLAST is intended as a last resort variable that can be modified in Makefile. Thanks! Michael

3 4

Heap-allocated StructSequences
by Eddie Elizondo 02 Oct '18

02 Oct '18

PEP-384 talks about defining a Stable ABI by making PyTypeObject opaque. Thus, banning the use of static PyTypeObjects. Specifically, I’d like to focus on those static PyTypeObjects that are initialized as StructSequences. As of today, there are 14 instances of these types (in timemodule.c, posixmodule.c, etc.) within cpython/Modules. These are all initialized through PyStructSequence_InitType2. This is very problematic when trying to make these types conform to PEP-384 as they are used through static PyTypeObjects. Problems: * PyStructSequence_InitType2 overrides the PyTypeObject: This C-API does a direct memcpy from a “prototype” structure. This effectively overrides anything set within the PyTypeObject. For example, if we were to initialize a heap allocated PyTypeObject and pass it on to this function, the C-API would just get rid of the Py_TPFLAG_HEAPTYPE flag, causing issues with the GC. * PyStructSequence_InitType2 does not work with heap allocated PyTypeObjects: Even if the function is fixed to preserve the state of the PyTypeObject and only overriding the specific slots (i.e. tp_new, tp_repr, etc.), it is expected that PyStructSequence_InitType2 will call PyType_Ready on the object. That means that the incoming object shouldn’t be initialized by a function such as PyType_FromSpec, as that would have already called PyType_Ready on it. Therefore, PyStructSequence_InitType2 will now have the responsibility of setting all the slots and properties of the PyHeapTypeObject, which is not feasible. * PyStructSequence_NewType is non-functional: This function was meant to be used as a way of creating a heap-allocated PyTypeObject that be passed to PyStructSequence_InitType2, effectively returning a heap allocated PyTypeObject. The current implementation doesn’t work in practice. Given that this struct is created in the heap, the GC has control over it. Thus, when the GC tries to traverse the type it complains with: “Error: type_traverse() called for non-heap type”, since it doesn’t have the Py_TPFLAG_HEAPTYPE flag. If we add the flag, we run into bullet point 1, if we are able to preserve the flag then we will still run into the problem of bullet point 2. Extra note: This C-API is not being used anywhere within CPython itself. Solution: * Fix the implementation of PyStructSequence_NewType: The best solution would be to fix the implementation of this function. This can easily be done by dynamically creating a PyType_Spec and calling PyType_FromSpec ``` PyObject* PyStructSequence_NewType(PyStructSequence_Desc *desc) { // … PyType_Spec* spec = PyMem_NEW(PyType_Spec, 1); spec->name = desc->name; spec->basicsize = sizeof(PyStructSequence) - sizeof(PyObject *); spec->itemsize = sizeof(PyObject *); spec->flags = Py_TPFLAGS_DEFAULT; spec->slots = PyMem_NEW(PyType_Slot, 6); spec->slots[0].slot = Py_tp_dealloc; spec->slots[0].pfunc = (destructor)structseq_dealloc; // … bases = PyTuple_Pack(1, &PyTuple_Type); type = PyType_FromSpecWithBases(spec, bases); // … ``` This will cleanly create a heap allocated PyStructSequence which can be used just like any stack allocated PyTypeObject initialized through PyStructSequence_InitType2. This means that any C-Extension should be using PyStructSequence_NewType and only built-in types should be calling PyStructSequence_InitType2. This will enable these types to comply with PEP-384 As an extra, I already have patches for this proposal. They can be found here: Branch: https://github.com/eduardo-elizondo/cpython/tree/heap-structseq Reimplement PyStructSequence_NewType: https://github.com/eduardo-elizondo/cpython/commit/413f8ca5bc008d84b3397ca1… Patch timemodule with NewType: https://github.com/eduardo-elizondo/cpython/commit/0a35ea263a531cb03c06be9e… Thoughts?

3 3