datetime.timestamp() takes daylight saving time into account without given timezone

[siegbert7]

Bug report

Bug description:

# Problem: datetime.timestamp() takes DST into account without any given timezone
# Expected: datetime.timestamp() does not calculate with DST if no timezone is available
# Python 3.14.6 (tags/v3.14.6:c63aec6, Jun 10 2026, 10:26:10) [MSC v.1944 64 bit (AMD64)] on win32

from datetime import datetime as dt

dt1 = dt(2026, 3, 29, 1, 59, 59)                    # Just before Daylight Saving Time
dt2 = dt(2026, 3, 29, 3,  0,  1)                      # Just after DST

print(dt1.tzinfo, dt2.tzinfo)                         # None None                (As expected)
print(dt1)                                                    # 2026-03-29 01:59:59  (As expected)
print(dt2)                                                    # 2026-03-29 03:00:01  (As expected)
print(dt2-dt1, dt1.tzinfo, dt2.tzinfo)            # 1:00:02 None None    (As expected, 3602 seconds)

# Problem with timestamp()
dt1_ts = dt1.timestamp()
dt2_ts = dt2.timestamp()

print()
print(dt1_ts)                                       # 1774745999.0         (As expected)
print(dt2_ts)                                       # 1774746001.0         (Surprising !!!)
print(dt2_ts - dt1_ts)                              # 2.0                  (Surprising, should be 3602 seconds like (dt2-dt1) !!!)

CPython versions tested on:

3.14

Operating systems tested on:

Windows

[picnixz]

I'm not entirely sure this is a bug. Timestamps are free of DST or anything. dt1_ts and dt2_ts are plain timestamps so there is no notion of timezones in this case. See https://docs.python.org/3/library/datetime.html#datetime.datetime.timestamp. In particular, since none of your datetimes have a tzinfo, you need to first put them with UTC TZ info as spelled in the docs.

cc @StanFromIreland

[StanFromIreland]

Naive datetime instances are assumed to represent local time, so timestamp() correctly reports that only 2 seconds of real time elapsed between local 01:59:59 and 03:00:01 DST (the hour in between never happened on your clock!). dt2 - dt1 differs because naive subtraction is pure wall-clock arithmetic with no time zone adjustment, this is also explicitly documented. Bénédikt's suggestion is correct.

[siegbert7]

I cannot agree to this opion, because in the middle of my example I have shown that the difference of the two datetimes is 3602 seconds: print(dt2-dt1, dt1.tzinfo, dt2.tzinfo) # 1:00:02 None None (As expected) Do we really want to argue, that calcutating the difference with "dt2-dt1" should deliver 3602 seconds, while the same difference calculated with two timestamps should be 2 seconds only? That is inconsistent.

…

________________________________ Von: Stan Ulbrych ***@***.***> Gesendet: Sonntag, 16. August 2026 12:07 An: python/cpython ***@***.***> Cc: siegbert7 ***@***.***>; Author ***@***.***> Betreff: Re: [python/cpython] datetime.timestamp() takes daylight saving time into account without given timezone (Issue #155883) [https://avatars.githubusercontent.com/u/89152624?s=20&v=4]StanFromIreland left a comment (python/cpython#155883)<#155883 (comment)> Naive datetime instances are assumed to represent local time, so timestamp()<https://docs.python.org/3/library/datetime.html#datetime.datetime.timestamp> correctly reports that only 2 seconds of real time elapsed between local 01:59:59 and 03:00:01 DST (the hour in between never happened on your clock!). dt2 - dt1 differs because naive subtraction is pure wall-clock arithmetic with no time zone adjustment, this is also explicitly documented<https://docs.python.org/3/library/datetime.html#datetime.datetime.timestamp:~:text=raised%2E-,If,case>. Bénédikt's suggestion is correct. — Reply to this email directly, view it on GitHub<#155883?email_source=notifications&email_token=ALTQUDJFCJCIYRSNQNWLCXT5KGBVDA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTKMZQGY4DQOBTGYY2M4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L#issuecomment-5306888361>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ALTQUDO56CIXPWVT3H2P72L5KGBVDAVCNFSNUABEKJSXA33TNF2G64TZHM4DCNJZHA4TMMJ3JFZXG5LFHM2TCNRTGMYDKNBZGWQXMAQ>. Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS<https://github.com/notifications/mobile/ios/ALTQUDNBLJSYOTU6EJOTJKD5KGBVDA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTKMZQGY4DQOBTGYY2M4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KUZTPN52GK4S7NFXXG> and Android<https://github.com/notifications/mobile/android/ALTQUDNKI5OS72GK7H6A3LL5KGBVDA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTKMZQGY4DQOBTGYY2M4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2K4ZTPN52GK4S7MFXGI4TPNFSA>. Download it today! You are receiving this because you authored the thread.Message ID: ***@***.***>

[StanFromIreland]

Please see my comment above:

dt2 - dt1 differs because naive subtraction is pure wall-clock arithmetic with no time zone adjustment, this is also explicitly documented.

[siegbert7]

That's the part I understood. But I would have excepted, that timestamp() without a timezone at the datetime-object is also completely pure wall-clock. How do you determine the date of DST without a timezone at the datetime-object ? There are a lot of DST's around the world with many different days. I would have expected the rule "no DST if there is no timezone at the object".

…

________________________________ Von: Stan Ulbrych ***@***.***> Gesendet: Sonntag, 16. August 2026 20:52 An: python/cpython ***@***.***> Cc: siegbert7 ***@***.***>; Author ***@***.***> Betreff: Re: [python/cpython] datetime.timestamp() takes daylight saving time into account without given timezone (Issue #155883) [https://avatars.githubusercontent.com/u/89152624?s=20&v=4]StanFromIreland left a comment (python/cpython#155883)<#155883 (comment)> Please see my comment above: dt2 - dt1 differs because naive subtraction is pure wall-clock arithmetic with no time zone adjustment, this is also explicitly documented<https://docs.python.org/3/library/datetime.html#datetime.datetime.timestamp:~:text=raised%2E-,If,case>. — Reply to this email directly, view it on GitHub<#155883?email_source=notifications&email_token=ALTQUDOZGFPAY6EIQCDV6WT5KH7H3A5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTKMZQHEYTANJSGIZKM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L#issuecomment-5309105222>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ALTQUDKX4W7HHBYKJPWYWA35KH7H3AVCNFSNUABEKJSXA33TNF2G64TZHM4DCNJZHA4TMMJ3JFZXG5LFHM2TCNRTGMYDKNBZGWQXMAQ>. Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS<https://github.com/notifications/mobile/ios/ALTQUDNJB4SDTHJIZXLFAX35KH7H3A5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTKMZQHEYTANJSGIZKM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KUZTPN52GK4S7NFXXG> and Android<https://github.com/notifications/mobile/android/ALTQUDPDBWWTKO7FVPCLLZD5KH7H3A5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTKMZQHEYTANJSGIZKM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2K4ZTPN52GK4S7MFXGI4TPNFSA>. Download it today! You are receiving this because you authored the thread.Message ID: ***@***.***>