Closed jmercouris closed 4 months ago
Thanks @jmercouris. I made a slight change to the changelog.
And so the wind blew with an unbeknownst coordinated grace. As if the mouths of all beings of the Earth were summoned to blow in unison.
In other words, it seems that we're roaming in a direction that actually matters.
For completeness, a note follows.
Commit 83f12c0b63af282d03930a20435d127f7e9ef498 correctly acknowledged that relying on the renderer for reloads is a superior solution. However, it broke reopen-dead-buffer
(as in the case of #3329). The aforementioned commit was eventually reverted (in fe825393db6f764ddfbd7a31c06e4880011b2832) to fix the reopen-dead-buffer
bug.
What strategy is superior - 83f12c0b63af282d03930a20435d127f7e9ef498 or #3329? I'm convinced that the latter is superior. The renderer will always know best. In the extreme case that the renderer isn't able to reload
, then ffi-buffer-reload
can always fallback to ffi-buffer-load
. Being explicit is better than fitting logic into one of the renderer's method specializations.
CC @jmercouris.
Description
Fixes #3325
Checklist:
(asdf:test-system :nyxt/gi-gtk)
)