jenkins-bot has submitted this change. ( https://gerrit.wikimedia.org/r/c/pywikibot/core/+/1225619?usp=email )

Change subject: get_delay: consider retry_after in delay calculation ......................................................................

get_delay: consider retry_after in delay calculation

Include any pending retry_after when computing the delay for read or write operations. The returned value still incorporates process concurrency via process_multiplicity.

Bug: T414354 Change-Id: I06e863e35cb8087b493b74bd7d7cdf958a8572f2 --- M pywikibot/throttle.py 1 file changed, 15 insertions(+), 7 deletions(-)

Approvals: jenkins-bot: Verified Xqt: Looks good to me, approved

diff --git a/pywikibot/throttle.py b/pywikibot/throttle.py index 0fe30ad..d0fd51f 100644 --- a/pywikibot/throttle.py +++ b/pywikibot/throttle.py @@ -10,7 +10,7 @@ on the number of concurrent bot instances, and optional lag-aware delays. """ # -# (C) Pywikibot team, 2008-2025 +# (C) Pywikibot team, 2008-2026 # # Distributed under the terms of the MIT license. # @@ -269,25 +269,33 @@ def get_delay(self, *, write: bool = False) -> float: """Return the current delay, adjusted for active processes.

- Compute the delay for a read or write operation, factoring in - process concurrency. This method does not account for how much - time has already passed since the last access — use + Computes the delay for a read or write operation, taking into + account process concurrency and any pending retry-after value. + The returned value already includes the + :attr:`process_multiplicity` factor. This method does not + consider how much time has passed since the last access — use :meth:`waittime` for that.

.. versionadded:: 10.3.0 Renamed from :meth:`getDelay`. + .. versionchanged:: 11.0 + The delay now takes any pending :attr:`retry_after` into + account.

:param write: Whether the operation is a write (uses writedelay). :return: The delay in seconds before the next operation should occur. """ - current_delay = self.writedelay if write else self.delay - # Refresh process count if the check interval has elapsed if time.time() > self.checktime + self.checkdelay: self.checkMultiplicity()

- current_delay = max(self.mindelay, min(current_delay, self.maxdelay)) + current_delay = max( + self.mindelay, + self.retry_after, + min(self.writedelay if write else self.delay, self.maxdelay) + ) + return current_delay * self.process_multiplicity

def waittime(self, write: bool = False):