Whether name names a history and nothing else.
The name goes straight into the file a history is stored in and into the images directory it owns, so one carrying a path component names something outside the history directory entirely: "../x" is a history file written wherever the traversal lands, and deleting that history sweeps every file of the directory it landed in. "." and ".." name the history directory itself. Names reach us from the bus, so this is a refusal, not an assertion.
Bounded too: a file name is 255 bytes at most on the filesystems a home directory is on, and a plain file history is saved through its own name with ".xml.tmp" after it. A name past that has a placeholder written and listed, while no save of it could ever land. The other flavours need more room after the name still -- the SQLite ones a "-journal" beside their database, the encrypted file one the original a re-key keeps aside -- which only the creation of a history asks of them (g_paste_storage_name_fits ()): a name a history was saved under stays one.
Installed and introspected, where the paths it guards are the daemon's alone: a client with nowhere to report a refusal -- the GNOME Shell extension's history chooser -- asks this before sending a name, and a copy of the rule there would drift from the one the daemon enforces.
| name |
the name of a history |
|
whether |