I am however reluctant to make a release with this as the bug would still wreak breakage on downstreams that use a non-default bitmap maker. Which would be unusual, but legit API use.
Also the issue seems to be very circumstance-specific and the cause in the control flow is not yet known to me (although we seem to have narrowed down the commit and relation to stride), so I am not certain if this would truly work around all cases.
I am however reluctant to make a release with this as the bug would still wreak breakage on downstreams that use a non-default bitmap maker. Which would be unusual, but legit API use.
Also the issue seems to be very circumstance-specific and the cause in the control flow is not yet known to me (although we seem to have narrowed down the commit and relation to stride), so I am not certain if this would truly work around all cases.