feat(core,fuse,gui): rebuild Trash and shared pages on one file list

The list and grid view from My files moves into widgets/file_list.rs as
FileList, so Trash, Shared with me and Shared by me share its columns,
selection, context menus, keyboard handling and list/grid switch instead of
each drawing its own rows.

Trash shows where each item came from and when it was deleted. Proton does
not expose the trash time through the SDK yet, so the daemon records when it
first sees a node in the trash (migration V35, trash_seen) and reports that;
the date becomes exact once proton-sdk-rs exposes Node::trash_time. Restoring
offers a Show action, and opening a trashed item explains that it needs to be
restored first.

Shared with me splits files and saved links into two views. Files show who
shared them, with a warning pill for unverified senders, the access they give
and when they were shared. Pending invitations sit in a banner that opens a
dialog to accept or decline each one. Several items can be left at once.
Saved links get their own list with open, copy and remove actions.

Shared by me gains a filter (all, links, people, expired links) and columns
for location, people and link state, with a copy button for live links.
Several items can stop being shared at once, the page refreshes after the
share dialog closes, and a reload rebinds only the rows that changed so the
selection survives.

All catalogs are updated and complete.

crates/pdfs-core/src/db/migrations.rs

@@ -9,7 +9,7 @@ use super::Db;
 use crate::Result;
 
 /// Current schema version. Bump on every forward migration added below.
-pub(super) const SCHEMA_VERSION: i64 = 34;
+pub(super) const SCHEMA_VERSION: i64 = 35;
 
 impl Db {
     pub(super) fn migrate(&self) -> Result<()> {
@@ -316,6 +316,22 @@ impl Db {
                 tx.execute_batch(MIGRATION_V34)?;
             }
         }
+        if current < 35 {
+            // Same guards as V26-V34.
+            let has_column: bool = tx.query_row(
+                "SELECT COUNT(*) FROM pragma_table_info('trash') WHERE name = 'trashed_at'",
+                [],
+                |row| row.get::<_, i64>(0),
+            )? > 0;
+            let has_trash: bool = tx.query_row(
+                "SELECT COUNT(*) FROM sqlite_master WHERE type = 'table' AND name = 'trash'",
+                [],
+                |row| row.get::<_, i64>(0),
+            )? > 0;
+            if has_trash && !has_column {
+                tx.execute_batch(MIGRATION_V35)?;
+            }
+        }
         tx.execute(
             "INSERT INTO sync_state (key, value) VALUES ('schema_version', ?1)
              ON CONFLICT(key) DO UPDATE SET value = excluded.value",
@@ -1033,3 +1049,13 @@ UPDATE photos SET resolved_at = NULL;
 const MIGRATION_V34: &str = "
 ALTER TABLE photos ADD COLUMN similar_hash INTEGER;
 ";
+
+/// Schema v35: when each trashed node went to the trash.
+///
+/// The API does not tell this client, so `trashed_at` is when the daemon first
+/// saw the node in the trash, in epoch seconds. `NULL` when it cannot say: rows
+/// from before this version, and nodes found by a refresh that did not follow a
+/// complete one, which may have been in the trash for a long time already.
+const MIGRATION_V35: &str = "
+ALTER TABLE trash ADD COLUMN trashed_at INTEGER;
+";