Turns out the cause is a wrong entry in the gpgsm.conf. Setting "dbug-level basic" without specifying an output file.
And encrypting a file is likewise affected.

Test on Windows by overwriting gpgtar from gpg4win-5.0.0-beta357 and also tested on Linux. Debian packages with patches are already available.

There is a new --keyserver-option update-before-send which is enabled by default.

the web client received some changes in the meantime and behaves more verbose regarding connections to a native client. we should check this again, especially if GpgOL/Web is supposed to be used for different outöook accounts at the same time.

in the meantime we decided it's not always desirable to have the process autostarted. therefore the current default (manual start) is considered ok, but if possible a configuration option should be implemented, so users can activate autostart if they find it useful.

Kleopatra from VSD installations should now use gpgv from the GnuPG installed by VSD when verifying the signature of the VERSION file. GPD and Gpg4win installations both use gpgv from the GnuPG installed at the value of the registry key "Software\\GnuPG:Install Directory".

Closing this ticket, the task is done. Further improvements should go into a new tickets after we decided what we want.

Ok, setting this to done, but not resolved so that timegrid will notice it and make a new ticket with more information for the "no data" message.

In T7642#203189, @gniibe wrote:

I wonder about GCC 15 preparation for the release. If it's good to have, three patches are needed to apply:

Ok, thanks. I pushed the powerpc patches to master.

If an error occurs then we emit a failure. But if gpg crashes then gpgme just sees a broken pipe.

But we emit a failure at the end of the gpg process; aren't we?

This might be related to the recent changes in now we spawn processes. Using gpgtar [...] --status-fd=3 3>a.log shows something different than directly using --status-fd=2. Do we handle the process termination correctly; i.e. wait for all status-fd output?

well the messages are improved, if we want to improve them further, this should go into a new ticket.

I assume you only forgot to move this to testing, as this is solved.
We now have separate windows for the notepad, several ones can be opened now.
Checked with Gpg4win-5.0.0-beta357

Some lines from the log:

2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_io_read: check: [GNUPG:] DECRYPTION_KEY 46ED5D7758C1BD71C27AF928
2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_io_read: check: FFA2FCCB2EC589F8 98111E67AE06F2BEFD2BDE10C5D6C91
2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_io_read: check: 9005F36A4 u<LF>
2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_io_read: check: [GNUPG:] BEGIN_DECRYPTION<LF>
2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_io_read: check: [GNUPG:] DECRYPTION_INFO 2 9 0 0<LF>
2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_io_read: check: [GNUPG:] PLAINTEXT 62 1753871344 <LF>
[...]
2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_io_read: check: [GNUPG:] NEWSIG<LF>
[...]
2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_io_read: check: [GNUPG:] KEY_CONSIDERED F8D51DE0EE16E9B57009B8DE
2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_io_read: check: 458612006D8E6F0D 0<LF>
[...]
2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_io_read: check: [GNUPG:] GPGTAR_EXTRACT 2 0 0 0 0 0 0<LF>
[...]
2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_io_read: check: [GNUPG:] SUCCESS<LF>
[...]
2025-07-30 12:35:29 gpgme[716.808] gpgme:reader: check: ctx->hdd=0x000001b90c8a6a60 got EOF (broken pipe)
[...]
2025-07-30 12:35:29 gpgme[716.23bc]   _gpgme_cancel_with_err: enter: ctx=0x000001b90c854430 ctx_err=117440570, op_err=0
[...]
2025-07-30 12:35:29 gpgme[716.1e08] gpgme:reader: check: ctx->hdd=0x000001b90c8a68c0 got EOF (closed by us)
[...]
2025-07-30 12:35:29 gpgme[716.23bc] gpgme_op_decrypt_ext:186: error: No data <GPGME>\n
2025-07-30 12:35:29 gpgme[716.23bc] gpgme_op_verify_result: enter: ctx=0x000001b90c854430 
2025-07-30 12:35:29 gpgme[716.23bc] gpgme_op_verify_result: leave: result=0x000001b90c7d1150
2025-07-30 12:35:29 gpgme[716.23bc] gpgme_get_protocol: call: ctx=0x000001b90c854430 ctx->protocol=0 (OpenPGP)
2025-07-30 12:35:29 gpgme[716.23bc] gpgme_op_decrypt_result: enter: ctx=0x000001b90c854430 
2025-07-30 12:35:29 gpgme[716.23bc] gpgme_op_decrypt_result: check: ctx=0x000001b90c854430 result: recipient: keyid=FFA2FCCB2EC589F8, pubkey_algo=302, status=Success
2025-07-30 12:35:29 gpgme[716.23bc] gpgme_op_decrypt_result: leave: result=0x000001b90c5e3290

-> there's a broken pipe; _gpgme_cancel_with_err has ctx_err 117440570 = (GPGME, No data)
-> gpgme_op_decrypt_ext reports an error: No data <GPGME>\n
-> There doesn't seem to be a verify result (although the archive is signed and encrypted)
-> There is a decrypt result with status Success

I had a quick look at the status messages that gpg emits. Unfortunately, gpg doesn't emit a status message after successfully signing a key. So, we may have to check whether the signed key really has a new certification to ensure that the operation succeeded. And we should make future version of gpg emit a success status.

tested with Gpg4win-5.0.0-beta357 (GnuPG 2.5.11):