Prove: il commento dell'attesa sul database dice ciò che si è visto, non una causa

Il commento di TestLibrary.DeleteClosedFileAsync dava per certo che il file appena chiuso fosse
tenuto dall'antivirus. È solo una deduzione: su un PC il file restava in uso per pochi
millisecondi, il Restart Manager non elencava alcun processo e Defender era attivo, ma la causa
non è stata dimostrata (la revisione non ha riprodotto il caso). Il commento ora lo dice, così
nessuno smette di cercare né copia l'attesa nel prodotto per una causa mai vista. Il codice non
cambia.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Alby96andClaude Opus 5.5 committed 2026-10-20 13:39:18 +02:00
1 parent 815bd44a62
commit 198e48d4f8
1 file changed
+4 -4
+4 -4
View File
@@ -71,10 +71,10 @@ internal sealed class TestLibrary : IAsyncDisposable
}
/// <summary>
/// Deletes a file Febe has already closed. On Windows a file that another program holds open cannot be deleted, and a
/// database just written is sometimes held for a few milliseconds by a program that scans new files (the real-time
/// antivirus): that sharing violation is waited out (at most 5 s). On Linux a delete never meets one, so the first
/// attempt is the only one.
/// Deletes a file Febe has already closed. On Windows a file still held open elsewhere cannot be deleted, and on one PC a
/// database just closed was now and then held for a few milliseconds by something outside Febe (Restart Manager listed
/// no process; probably the real-time antivirus, which was active, but the cause was never shown): that sharing
/// violation is waited out (at most 5 s). On Linux a delete never meets one, so the first attempt is the only one.
/// </summary>
private static async Task DeleteClosedFileAsync(string path)
{