> So if we switch to the datetime + zone info + timezone we need to have a test suite that verifies this.
We need to have these test suites anyway. According to the issues and attempts to fix DST transitions we still have that problem even with use Pendulum, or maybe it is happen because we use pendulum rather than use fold argument in datetime.datetime On Fri, 1 Dec 2023 at 20:41, Bolke de Bruin <bdbr...@gmail.com> wrote: > The alternative is to vendor in or fork pendulum as part of the Airflow > project. Also a major undertaking but maybe less effort. > > We shouldn't just address this as a technical challenge. The question > becomes: does that combination fornally do DST transitions correctly, does > it do leap seconds correctly, does it add dates correctly, does it do > timezone traversal correctly and many many other things. > > So if we switch to the datetime + zone info + timezone we need to have a > test suite that verifies this. Otherwise we potentially expose our users to > different behavior IN their data. Not just a schedule by in their SQL or in > some scripting where logical date / execution date is used or where date > arithmetic is applied. > > So it is not changing one car for another. It is changing the means of > transportation and not knowing if you can get to all the places you want / > need to go as before. > > Honestly, I'm a bit scared (I do see the lack of responsiveness). I hope > someone can take that away. > > Bolke > > > Sent from my iPhone > > > On 1 Dec 2023, at 13:39, Jarek Potiuk <ja...@potiuk.com> wrote: > > > > Agree. I think this is the best way forward.- we have to bite the > bullet of > > potential backwards compatibility issues and separating out a compat > layer > > where we could optionally use pendulum instead would be the most "easy" > > approach for our users. Most of them would not even notice if we do a > good > > job, but those who somehow depend on pendulum objects would get a > > possibility to switch back. > > > >> On Fri, Dec 1, 2023 at 1:09 PM Andrey Anshin <andrey.ans...@taragol.is> > >> wrote: > >> > >> IMHO, if we would like to replace pendulum to alternatives, the better > >> alternative would be combination of datetime.datetime + zoneinfo + > >> datetime.timezone > >> > >> We still do hacks around pendulum functional, which makes hacks around > >> datetime.datetime, so in my perspective better remove redundant layers > and > >> use native objects directly. > >> > >> What confuses me the most is how we get around type changes in Tasks > >> Context. All dates return as pendulum.DateTime for a very long time > >> (literally for ages) and it is probably could be classified as breaking > >> changes, because it might break quite a bit users pipelines. It still > could > >> be resolved as optional pendulum dependency + config option for datetime > >> types in context, which might help to get familiar types for a while. > >> > >> > >>> On Wed, 29 Nov 2023 at 05:58, Jarek Potiuk <ja...@potiuk.com> wrote: > >>> > >>> "infamous fork of " should be "fork of infamous Akka" - just to be > clear > >> :) > >>> > >>>> On Wed, Nov 29, 2023 at 2:57 AM Jarek Potiuk <ja...@potiuk.com> > wrote: > >>> > >>>> It's a very rare occurrence that ASF accepts forks. Usually it has to > >> be > >>>> willingly donated by those who own the IP rights, - mainly because of > >>>> trademark issues. Licence is one thing, but carrying things like the > >> name > >>>> of the project is not possible without clearing IP and donating the > >>>> project. > >>>> > >>>> It happened last time last year - with Pekko project - infamous fork > or > >>>> Akka after Lightbend changed licensing for Akka - > >>>> > >> https://www.lightbend.com/blog/why-we-are-changing-the-license-for-akka > >>>> making it impossible to be used in ASF projects with the new licence. > >> But > >>>> it was a really "strong" case and the "we do not accept forks usually" > >>> was > >>>> quite a contentious issue (I followed the discussion - it was > >> fascinating > >>>> actually). It tool a LOT of time and effort for those who led the > >> effort > >>>> eventually - and also mostly because things had to be crystal-clear > and > >>> not > >>>> "litigable" because the other side had some business reasons to make > >> the > >>>> licensing changes and they did not understood why it is impossible to > >> use > >>>> Akka as-is even with all the exclusions they offered to ASF - and > >>>> effectively it was a "hostile" move. > >>>> > >>>> I don't think it's "worth" it in this case to be honest - especially > if > >>>> the maintainers are effectively "silent" and seem to ignore problems > of > >>> one > >>>> of the really serious users they have. If they were willing to > >> cooperate, > >>>> we could do a lot, but when we are left alone, I think it will be far > >>> more > >>>> convenient for us to look for alternatives - to be honest. > >>>> > >>>> J. > >>>> > >>>> > >>>> On Wed, Nov 29, 2023 at 2:46 AM Austin Bennett <aus...@apache.org> > >>> wrote: > >>>> > >>>>> An option would be to fork Pendulum? It is MIT Licensed, I don't > know > >>>>> whether that poses problems to get in ASF? > >>>>> > >>>>> If forking (?) which is somewhat non-ideal, would we want that 'in' > >>>>> airflow? > >>>>> > >>>>> If not 'in' airflow, I wonder if ASF incubator would accept a forked > >>>>> project? [ anecdotally Linux Foundation has incubated Starrocks > which > >>> is > >>>>> a > >>>>> fork of Apache Doris, so there's precedence for this to be OK in the > >>> wider > >>>>> OSS ecosystem ]. > >>>>> > >>>>> > >>>>> On Tue, Nov 28, 2023 at 6:41 PM Jarek Potiuk <ja...@potiuk.com> > >> wrote: > >>>>> > >>>>>> I have not heard back from maintainers, just a comment from someone > >>> else > >>>>>> who suggested donating a pendulum to the ASF (which is kinda > >>> interesting > >>>>>> idea). I followed up, let's see. I think if we do not hear back and > >>>>> there > >>>>>> will be another week or two where "Pendulum 3 supporting 3.12 > >> support > >>>>> "is > >>>>>> coming" but no-one knows when, effectively means that we should look > >>>>> for an > >>>>>> alternative like NOW. > >>>>>> > >>>>>> J. > >>>>>> > >>>>>> > >>>>>> > >>>>>> On Wed, Nov 22, 2023 at 3:27 AM Daniel Standish > >>>>>> <daniel.stand...@astronomer.io.invalid> wrote: > >>>>>> > >>>>>>> Thanks Jarek > >>>>>>> > >>>>>>> On Tue, Nov 21, 2023, 4:34 PM Jarek Potiuk <ja...@potiuk.com> > >>> wrote: > >>>>>>> > >>>>>>>> I think we miss important insight - straight from the source. > >>>>>>>> > >>>>>>>> I believe it's time to be candid and simply ask questions for > >> the > >>>>>> future > >>>>>>> of > >>>>>>>> Pendulum directly where we should - ie. we should just ask > >>>>>> maintainers. > >>>>>>>> > >>>>>>>> I've just started a very candid and open discussion there - > >>>>>>>> https://github.com/sdispater/pendulum/discussions/771 . > >>>>>>>> > >>>>>>>> I linked back to this discussion and explained where we are > >> coming > >>>>>> from. > >>>>>>> I > >>>>>>>> even offered an option, that maybe - if they accept some help > >> and > >>>>> would > >>>>>>> be > >>>>>>>> open to make a few of us maintainers of Pendulum, that I will > >> ask > >>> if > >>>>>> some > >>>>>>>> of the maintainers here in Airflow would like to step up (I > >> would > >>> be > >>>>>>>> willing to - for one) and help to at least move Pendulum through > >>>>> 3.12 > >>>>>> and > >>>>>>>> maybe keep it running for as long as it will be needed to > >> maintain > >>>>> it. > >>>>>>>> > >>>>>>>> I think if they would be open to some of us helping them - this > >>>>> might > >>>>>> be > >>>>>>>> actually the "cheapest" option for us to be honest at least in a > >>>>>> mid-term > >>>>>>>> if we could gain influence and even take part and speed up > >>> releases > >>>>> and > >>>>>>>> merging of issues that are blocking us (or would be blocking us > >> in > >>>>> the > >>>>>>>> future). > >>>>>>>> > >>>>>>>> I tried to be very friendly but candid and direct on what kind > >> of > >>>>>>> problems > >>>>>>>> it creates us and also expressed our thanks for supporting us > >> for > >>> so > >>>>>> long > >>>>>>>> (I do believe they deserve it). I think the sole (pretty much) > >>>>>>> maintainer > >>>>>>>> of Pendulum might really have a hard time maintaining and > >> keeping > >>>>> it up > >>>>>>> for > >>>>>>>> years - without being paid for and thanked. I hope - maybe all > >>> that > >>>>>> they > >>>>>>>> need are a few words of encouragement and thanks and realising > >>> that > >>>>>>> others > >>>>>>>> depend on their work - but also seeing that they are understood, > >>>>>>> empathised > >>>>>>>> with and that there are viable alternatives they might follow > >>> might > >>>>> be > >>>>>>>> helpful for them to be also candid and respond. I hope my > >>> intentions > >>>>>> will > >>>>>>>> be understood - that I am not complaining, but even trying to > >>> help. > >>>>>>>> > >>>>>>>> Let\s see where it gets us. > >>>>>>>> > >>>>>>>> J. > >>>>>>>> > >>>>>>>> On Fri, Nov 17, 2023 at 5:32 PM Andrey Anshin < > >>>>>> andrey.ans...@taragol.is> > >>>>>>>> wrote: > >>>>>>>> > >>>>>>>>> Just to clarify I'd like us to consider the possibility that > >> no > >>>>> new > >>>>>>>>> pendulum would be released or released at the end of 2024, > >> like > >>> a > >>>>>>>>> pessimistic scenario: > >>>>>>>>> - What should we do in this case? > >>>>>>>>> - Work out a backup plan. > >>>>>>>>> > >>>>>>>>> > >>>>>>>>> ---- > >>>>>>>>> Best Wishes > >>>>>>>>> *Andrey Anshin* > >>>>>>>>> > >>>>>>>>> > >>>>>>>>> > >>>>>>>>> On Fri, 17 Nov 2023 at 16:33, Jarek Potiuk <ja...@potiuk.com> > >>>>> wrote: > >>>>>>>>> > >>>>>>>>>> Also I think TP - had a document in the past (years ago) > >>>>>>> describing a > >>>>>>>>>> draft of a more complete alternative we can take to approach > >>>>>> datetime > >>>>>>>> vs. > >>>>>>>>>> pendulum dichotomy. I cannot easily find the document and > >>>>>> discussion > >>>>>>> - > >>>>>>>>> but > >>>>>>>>>> I do remember it was proposing some interesting changes in > >> the > >>>>>>> approach > >>>>>>>>> of > >>>>>>>>>> Airflow to have an abstraction layer on top of it (as far > >> as I > >>>>>>>> remember). > >>>>>>>>>> Maybe we can resurrect that idea if TP might find the > >>> proposal ? > >>>>>>>>>> > >>>>>>>>>> On Fri, Nov 17, 2023 at 1:06 PM Bolke de Bruin < > >>>>> bdbr...@gmail.com> > >>>>>>>>> wrote: > >>>>>>>>>> > >>>>>>>>>>> Hi, > >>>>>>>>>>> > >>>>>>>>>>> I agree that the current speed of development of Pendulum > >>>>> leaves > >>>>>>>>>> something > >>>>>>>>>>> to be desired. However, I think we should not > >> underestimate > >>>>> the > >>>>>>>> effort > >>>>>>>>> of > >>>>>>>>>>> replacing it. It is not just a matter of > >>>>> %s/pendulum/datetime/g > >>>>>> so > >>>>>>> to > >>>>>>>>>> say. > >>>>>>>>>>> If we are *truly* thinking about moving to native > >> datetime / > >>>>>>> zoneinfo > >>>>>>>>> etc > >>>>>>>>>>> we need *extensive* tests, basically copying what pendulum > >>>>> does > >>>>>> to > >>>>>>>>> check > >>>>>>>>>>> its behavior. The reason is that native implementations in > >>> the > >>>>>> past > >>>>>>>>> made > >>>>>>>>>>> serious mistakes and I do not put a lot of trust in them. > >>>>>>>>>>> > >>>>>>>>>>> An abstraction or vendoring in could be alternatives, but > >>> they > >>>>>>> bring > >>>>>>>>>> their > >>>>>>>>>>> own significant challenges. > >>>>>>>>>>> > >>>>>>>>>>> I have re-raised the issue here: > >>>>>>>>>>> > >>>>>>>>>>> https://github.com/sdispater/pendulum/issues/753 > >>>>>>>>>>> > >>>>>>>>>>> The upside is that it seems the amount of issues with the > >>>>> beta is > >>>>>>>>>> limited, > >>>>>>>>>>> so hopefully the maintainers can spend a couple of cycles > >> to > >>>>>>> address > >>>>>>>>>> them. > >>>>>>>>>>> > >>>>>>>>>>> Bolke > >>>>>>>>>>> > >>>>>>>>>>> On Fri, 17 Nov 2023 at 10:28, Andrey Anshin < > >>>>>>>> andrey.ans...@taragol.is> > >>>>>>>>>>> wrote: > >>>>>>>>>>> > >>>>>>>>>>>> There is no changes in stable pendulum so let's try to > >>>>> continue > >>>>>>>> this > >>>>>>>>>>>> discussion and start think about "Plan B" > >>>>>>>>>>>> > >>>>>>>>>>>> Just a reminder: > >>>>>>>>>>>> - pendulum 2.1.2 released 3 years ago (at the time > >> Airflow > >>>>>>> 1.10.x) > >>>>>>>>>>>> - pendulum 2 doesn't work well in Python 3.12, this is a > >>>>>>>> showstopper > >>>>>>>>>> for > >>>>>>>>>>>> the support Python 3.12 > >>>>>>>>>>>> - pendulum 2 have memory leaks ( > >>>>>>>>>>>> https://github.com/sdispater/pendulum/issues/720), and > >>>>> Airflow > >>>>>>> use > >>>>>>>>>>>> approach > >>>>>>>>>>>> to achieve this leaks, especially in K8S executor > >>>>>>>>>>>> - pendulum 2 doesn't use system tzdata by default, but > >> we > >>>>> have > >>>>>> a > >>>>>>>>>>>> workaround, thanks Bolke for the documentation > >>>>>>>>>>>> - pendulum it is a core dependency of Airflow, and I > >> guess > >>>>> we > >>>>>>> can't > >>>>>>>>>>>> remove/replace it without breaking changes. > >>>>>>>>>>>> > >>>>>>>>>>>> So my proposal if things won't change in the near future > >>>>> then > >>>>>> we > >>>>>>>> need > >>>>>>>>>> to > >>>>>>>>>>>> start removing the pendulum from the core and replace it > >>>>> with > >>>>>> the > >>>>>>>>>> native > >>>>>>>>>>>> datetime / zoneinfo / tzinfo. But maybe we have another > >>> way > >>>>> to > >>>>>>>>> resolve > >>>>>>>>>>> this > >>>>>>>>>>>> without breaking changes? Because for me it would be a > >>>>> little > >>>>>>> weird > >>>>>>>>> if > >>>>>>>>>>>> removal pendulum would be a main driver for release > >>> Airflow > >>>>> 3 > >>>>>>>>>>>> > >>>>>>>>>>>> > >>>>>>>>>>>> > >>>>>>>>>>>> > >>>>>>>>>>>> ---- > >>>>>>>>>>>> Best Wishes > >>>>>>>>>>>> *Andrey Anshin* > >>>>>>>>>>>> > >>>>>>>>>>>> > >>>>>>>>>>>> > >>>>>>>>>>>> On Thu, 28 Sept 2023 at 13:01, Andrey Anshin < > >>>>>>>>> andrey.ans...@taragol.is > >>>>>>>>>>> > >>>>>>>>>>>> wrote: > >>>>>>>>>>>> > >>>>>>>>>>>>> This discussion is more about the known problem of > >>>>> pendulum > >>>>>> and > >>>>>>>> how > >>>>>>>>>> we > >>>>>>>>>>>>> could deal with it and maybe how we (as Community) > >> might > >>>>> help > >>>>>>>>> autor. > >>>>>>>>>>>>> > >>>>>>>>>>>>> The library is mostly supported by a single author > >>>>> Sébastien > >>>>>>>>> Eustace > >>>>>>>>>> ( > >>>>>>>>>>>>> https://github.com/sdispater) and it seems like we > >> bump > >>>>> into > >>>>>>> the > >>>>>>>>>>>>> situation which is described in xkcd #2347 ( > >>>>>>>>>>>>> https://imgs.xkcd.com/comics/dependency.png). To be > >>>>> honest > >>>>>> it > >>>>>>> is > >>>>>>>>> not > >>>>>>>>>>>>> something new when library mainly supported by one > >>> author > >>>>> so > >>>>>>>> there > >>>>>>>>> is > >>>>>>>>>>>>> always a risk that the library will no longer be > >>>>> supported / > >>>>>>>>>> abandoned > >>>>>>>>>>>>> And if takes in account that pendulum provides core > >>>>>>> functionality > >>>>>>>>> in > >>>>>>>>>>>>> Airflow it could have dramatical impact in the future. > >>>>>>>>>>>>> > >>>>>>>>>>>>> Pendulum is a really nice library which helps a lot of > >>>>>>> developers > >>>>>>>>> to > >>>>>>>>>>> work > >>>>>>>>>>>>> with dates/datetimes. However there is one major > >>> problem, > >>>>> the > >>>>>>>> last > >>>>>>>>>>>> release > >>>>>>>>>>>>> of this library happened more than 3 years ago ( > >>>>>>>>>>>>> https://pypi.org/project/pendulum/#history) in the > >> time > >>>>> when > >>>>>>>>> Airflow > >>>>>>>>>>>>> 1.10.11 was released > >>>>>>>>>>>>> > >>>>>>>>>>>>> Fortunately, the project is not abandoned and on a > >>> regular > >>>>>>> basis > >>>>>>>>>>> commits > >>>>>>>>>>>>> add into the master branch. However these commits are > >>> not > >>>>>>>> included > >>>>>>>>>> into > >>>>>>>>>>>> any > >>>>>>>>>>>>> final release and that's why some things related to > >>>>> datetime > >>>>>>>> don't > >>>>>>>>>> work > >>>>>>>>>>>> as > >>>>>>>>>>>>> expected in Airflow. There are list of known (for me) > >>>>> issues > >>>>>>>> which > >>>>>>>>>> are > >>>>>>>>>>>>> affect Airflow > >>>>>>>>>>>>> > >>>>>>>>>>>>> *Memory Leak on parse*: > >>>>>>>>>>>>> - https://github.com/sdispater/pendulum/issues/720, > >>> this > >>>>> one > >>>>>>>> fixed > >>>>>>>>>> 2 > >>>>>>>>>>>>> years ago but not available yet ( > >>>>>>>>>>>>> https://github.com/sdispater/pendulum/pull/563). > >> Since > >>> we > >>>>>> use > >>>>>>>>> parse > >>>>>>>>>>>> dates > >>>>>>>>>>>>> in airflow codebase: datetime parameters and datetime > >> in > >>>>> logs > >>>>>>>> this > >>>>>>>>>> one > >>>>>>>>>>>>> could be a reason for memory leakage in Airflow: > >>>>>>>>>>>>> - https://github.com/apache/airflow/discussions/24694 > >>>>>>>>>>>>> - https://github.com/apache/airflow/discussions/28597 > >>>>>>>>>>>>> > >>>>>>>>>>>>> *Incorrect time zones*, known issues and should be > >>> already > >>>>>>> fixed > >>>>>>>> in > >>>>>>>>>>>>> master branch > >>>>>>>>>>>>> - https://github.com/sdispater/pendulum/issues/700, > >>>>> Mexico > >>>>>> do > >>>>>>>> not > >>>>>>>>>> use > >>>>>>>>>>>> DST > >>>>>>>>>>>>> anymore > >>>>>>>>>>>>> - https://github.com/sdispater/pendulum/issues/706, > >>> Egypt > >>>>>>>>> reinstate > >>>>>>>>>>> DST > >>>>>>>>>>>>> > >>>>>>>>>>>>> We add clarification in > >>>>>>>>> https://github.com/apache/airflow/pull/30467 > >>>>>>>>>> , > >>>>>>>>>>>>> however it seems like there is no other way rather > >> than > >>>>>>> patching > >>>>>>>>>>> Pendulum > >>>>>>>>>>>>> right now. > >>>>>>>>>>>>> > >>>>>>>>>>>>> All these issues should be solved as soon as pendulum > >> 3 > >>> is > >>>>>>>>> released. > >>>>>>>>>>> The > >>>>>>>>>>>>> current announced estimation is end of september/ > >>>>> beginning > >>>>>> of > >>>>>>>>>> October: > >>>>>>>>>>>>> > >>>>>>>>>>> > >>>>>>>>> > >>>>>>> > >>>>> > >>> > https://github.com/sdispater/pendulum/issues/600#issuecomment-1711299677 > >>>>>>>>>>>>> > >>>>>>>>>>>>> So in theory we would have a fixed version of pendulum > >>>>> soon, > >>>>>>> and > >>>>>>>> it > >>>>>>>>>>> might > >>>>>>>>>>>>> break something in Airflow but from my point of view > >> it > >>> is > >>>>>>> better > >>>>>>>>>> than > >>>>>>>>>>>>> current status. > >>>>>>>>>>>>> > >>>>>>>>>>>>> However there might be a situation where the release > >> of > >>>>> the > >>>>>>>>> pendulum > >>>>>>>>>>>> would > >>>>>>>>>>>>> be postponed, so maybe better to have a backup plan. > >>> What > >>>>>> could > >>>>>>>> we > >>>>>>>>> do > >>>>>>>>>>> in > >>>>>>>>>>>>> this case? > >>>>>>>>>>>>> > >>>>>>>>>>>>> Maybe we should start to use zoneinfo.ZoneInfo instead > >>> of > >>>>>>>> pendulum > >>>>>>>>>>>>> datetime? > >>> https://github.com/apache/airflow/issues/19450 > >>>>>>>>>>>>> Pros: > >>>>>>>>>>>>> - stdlib (python 3.9+) > >>>>>>>>>>>>> - In pendulum 3.0 Timezone based on zoneinfo.Zoneinfo > >>>>>>>>>>>>> > >>>>>>>>>>>>> Cons: > >>>>>>>>>>>>> - Current serialization model can't deal with backport > >>>>>>> packages. > >>>>>>>>> E.g. > >>>>>>>>>>>>> timezone which are serialized in backport_zoneinfo > >> can't > >>>>> be > >>>>>>>>>>> deserialized > >>>>>>>>>>>> in > >>>>>>>>>>>>> zoneinfo > >>>>>>>>>>>>> > >>>>>>>>>>>>> Maybe we should replace parse datetime with another > >>>>> solution. > >>>>>>>> Does > >>>>>>>>>>> anyone > >>>>>>>>>>>>> know a good replacement? > >>>>>>>>>>>>> > >>>>>>>>>>>>> Maybe someone from Airflow Community could propose > >> their > >>>>> help > >>>>>>>> with > >>>>>>>>>>>>> maintenance of library: > >>>>>>>>>>>>> - https://github.com/sdispater/pendulum/issues/590 > >>>>>>>>>>>>> > >>>>>>>>>>>>> Maybe we should get rid of the pendulum at all, as a > >>> last > >>>>>>> resort > >>>>>>>>>>>> solution. > >>>>>>>>>>>>> I can't imagine how we could do that, because a lot of > >>>>> stuff > >>>>>>>>> depends > >>>>>>>>>> on > >>>>>>>>>>>> the > >>>>>>>>>>>>> pendulum and removing it would be a breaking change. > >>>>>>>>>>>>> > >>>>>>>>>>>>> ---- > >>>>>>>>>>>>> Best Wishes > >>>>>>>>>>>>> *Andrey Anshin* > >>>>>>>>>>>>> > >>>>>>>>>>>>> > >>>>>>>>>>>> > >>>>>>>>>>> > >>>>>>>>>>> > >>>>>>>>>>> -- > >>>>>>>>>>> > >>>>>>>>>>> -- > >>>>>>>>>>> Bolke de Bruin > >>>>>>>>>>> bdbr...@gmail.com > >>>>>>>>>>> > >>>>>>>>>> > >>>>>>>>> > >>>>>>>> > >>>>>>> > >>>>>> > >>>>> > >>>> > >>> > >> > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@airflow.apache.org > For additional commands, e-mail: dev-h...@airflow.apache.org > >
