On Monday, 9 March 2026 at 16:28, Jakub Zelenka <[email protected]> wrote:
> [...] > > We could probably relay on some developers noticing such changes and calling > for RFC which can be done for any feature. But my worry is that some of the > cases that could be problematic just slip in and they might cause problems > for users later. So my preference is to require RFC for all which should make > sure that there is a good visibility for any such change. Those changes are > usually trivial so grouping them to a single RFC should not be that much work. People have previously complained that the mass RFC deprecation has too many entries, and you want to propose something similar for ValueError? And I'm against dragging out this sort of simple maintenance work onto an RFC stage which has frankly only gotten more tedious in the last few years. Any RFC will delay the merging of those simple PRs by *at least* 4 weeks, and more likely multiple months. This is IMHO an unacceptable way to do regular maintenance that can prevent real bugs ending up in production. Best regards, Gina P. Banyard
