summaryrefslogtreecommitdiffstats
path: root/appearance/README.md
diff options
context:
space:
mode:
Diffstat (limited to 'appearance/README.md')
-rw-r--r--appearance/README.md9
1 files changed, 9 insertions, 0 deletions
diff --git a/appearance/README.md b/appearance/README.md
index 2c127a2..ef25428 100644
--- a/appearance/README.md
+++ b/appearance/README.md
@@ -97,6 +97,15 @@ how far it enlarges: not at all is green, up to 1.33x yellow, past that red.
Under the mock screens a note in the same colour says why, for the hovered
image or else the staged one. The staged image keeps the accent border.
+When the staged image is yellow or red, "Find a larger version" copies it to
+the clipboard and opens TinEye, where Ctrl+V searches for it and the matches
+can be sorted by size. A search engine takes no local file in a URL, and
+scripted uploads hit bot and consent walls (TinEye answered 403, Google Lens
+redirected to its consent page and dropped the image), so the paste in the
+browser is the upload. The copy is converted to PNG first: Firefox gives a
+page only `image/png` from the clipboard, and a JPEG copied as itself
+pasted nothing. Replacing the file with a larger find is done by hand.
+
`wallp` does the actual work, including running `udt-accent`, so the accent
follows the new wallpaper exactly as it does from a terminal. This panel
replaces its qarma file dialog, not the script.