On 2025-07-31 19:27, Tomas Hlavaty <tom@logand.com> writes:

> On Thu 31 Jul 2025 at 12:27, Alan Schmitt <alan.schmitt@polytechnique.org> wrote:
>> (which is significant as all my messages where duplicated). Is there a
>> way to do this using notmuch (as it is already identifying duplicates),
>> or do you recommend another tool (maybe a file deduplication tool)?
>
> copy all messages to a directory, where each message filename is
> sha256sum of the file contents, then remove the originals
>
> notmuch does not seem to care much about the message filenames,
> except reindex might take a long time

Unfortunately not only mbsync cares about the file name, but it adds a
header when syncing, so the files are not fully identical (I tried
running fdupes before realizing this).

Alan