PR #4751 is now in the devel branch

[Edward K. Ream]

PR #4751 fixes an unintended consequence of a recent optimization involving Leo's colorizing code: Leo's  dump-last-colorizer-trace command became almost useless.

The PR fixes that problem by silently accumulating colorizing traces when --trace=verbose is in effect. You could call this an Easter Egg for my own use. See the first comment of the PR for details.

Edward

[Edward K. Ream]

On Tuesday, June 23, 2026 at 4:25:57 AM UTC-5 Edward K. Ream wrote:

PR #4751 fixes an unintended consequence of a recent optimization involving Leo's colorizing code: Leo's  dump-last-colorizer-trace command became almost useless.

The PR fixes that problem by silently accumulating colorizing traces when --trace=verbose is in effect. You could call this an Easter Egg for my own use. See the first comment of the PR for details.

On second thought, the trace=verbose hack is unbearably ugly. PR #4752 (in progress) always creates the data needed by the dump-last-colorizer-trace command. The performance gained by not creating the data isn't worth the loss of functionality.

Edward

[Edward K. Ream]

On Tuesday, June 23, 2026 at 8:03:14 AM UTC-5 Edward K. Ream wrote:

> PR #4752 (in progress) always creates the data needed by the dump-last-colorizer-trace command.

I've just merged this PR into the "devel" branch.  Onward!

Edward