[Swan-dev] regressions since v3.22-1715-gc22eea8cd-master

[D. Hugh Redelmeier]

| From: Antony Antony <antony at phenome.org>

| I stumbled on something similar and from a quick look pointed me to 
| the commit 1dbc99118f .  The test passed in v3.25 

Thanks.  You were quite right.

A merge silently moved a chunk of code inside a conditional.  So
important stuff was only done in the AH case (i.e. rarely).  This was
very hard to see.

The perils of leaving something in the stash, awaiting the code freeze
to end.