💬 fanquake commented on issue "Option to prevent sleep":
(https://github.com/bitcoin/bitcoin/issues/27692#issuecomment-1552909089)
Thanks. We've had similar proposals in the past, i.e #16145, and it's unlikely we are going to implement this kind of functionality. You can use your own scripts / external tools etc to acheive what you need.

✅ fanquake closed an issue: "Option to prevent sleep"
(https://github.com/bitcoin/bitcoin/issues/27692)

👍 hebasto approved a pull request: "build: Bump minimum supported GCC to g++-9"
(https://github.com/bitcoin/bitcoin/pull/27662#pullrequestreview-1432518265)
ACK fa953f15bfcf95df9aa9c91e1c4b56a205f4d1ae

💬 fanquake commented on issue "Mac osx 12.6.5 ":
(https://github.com/bitcoin/bitcoin/issues/27681#issuecomment-1552912525)
Thanks for the additional info. This is clearly a macOS, GUI-specific, issue, so I think this bug report can be moved to the GUI repo: https://github.com/bitcoin-core/gui, and someone there might be able to follow up?

💬 fanquake commented on issue "Can't start bitcoin-qt by double-click on Debian 11":
(https://github.com/bitcoin/bitcoin/issues/27655#issuecomment-1552913425)
Not sure what else we can do here? I think this should be moved over to the GUI repo: https://github.com/bitcoin-core/gui, in any case, and maybe someone can follow up there?

💬 MarcoFalke commented on issue "Mac osx 12.6.5 ":
(https://github.com/bitcoin/bitcoin/issues/27681#issuecomment-1552916702)
Is it confirmed that the daemon `bitcoind` is running successfully?

⚠️ hebasto opened an issue: "Can't start bitcoin-qt by double-click on Debian 11"
(https://github.com/bitcoin-core/gui/issues/730)
Moved from https://github.com/bitcoin/bitcoin/issues/27655.

### Is there an existing issue for this?

- [X] I have searched the existing issues

### Current behaviour

When I double-click the bitcoin-qt executable file in the graphical file manager, Bitcoin Core doesn't start and this message pops up:


I think this issue is the same as described at https://
...

⚠️ kallerosenbaum opened an issue: "bitcoin-qt generates bad psbt"
(https://github.com/bitcoin/bitcoin/issues/27697)
### Is there an existing issue for this?

- [X] I have searched the existing issues

### Current behaviour

I run bitcoin-qt in regtest mode. The wallet is created blank and I've imported descriptors to it. I have a total of 30 BTC in two outputs. I create a new transaction and save it as a PSBT for offline signing. When I try to read this file into bitcoin-qt again, it shows an error dialog with the message:

```
Unable to decode PSBT
CDataStream::read(): end of data: iostream error
```

...

💬 hebasto commented on issue "Can't start bitcoin-qt by double-click on Debian 11":
(https://github.com/bitcoin/bitcoin/issues/27655#issuecomment-1552919155)
Moved to https://github.com/bitcoin-core/gui/issues/730.

✅ hebasto closed an issue: "Can't start bitcoin-qt by double-click on Debian 11"
(https://github.com/bitcoin/bitcoin/issues/27655)

💬 fanquake commented on issue "Mac osx 12.6.5 ":
(https://github.com/bitcoin/bitcoin/issues/27681#issuecomment-1552920830)
I can't see a reason that it wouldn't be (absent a new report), and the crash here is coming from objective C code that isn't part of bitcoind.

💬 ajtowns commented on pull request "p2p: Stop relaying non-mempool txs":
(https://github.com/bitcoin/bitcoin/pull/27625#issuecomment-1552923162)
> > I think it would also be good to continue serving txs from `vExtraTxnForCompact`
>
> I wasn't sure how to do this, given the planned removal of `m_recently_announced_invs` in #27675 . Otherwise we may relay invalid/non-standard transactions that were injected into us by a spy node.

I guess a straightforward approach would be to have the mempool separately keep track of the last ~700 txs removed, as well as their entry time and removal time. Then they could be relayed to a node (under #
...

💬 MarcoFalke commented on pull request "p2p: Stop relaying non-mempool txs":
(https://github.com/bitcoin/bitcoin/pull/27625#issuecomment-1552934253)
>I guess a straightforward approach would be to have the mempool separately keep track of the last ~700 txs removed, as well as their entry time and removal time. Then they could be relayed to a node (under https://github.com/bitcoin/bitcoin/pull/27675's proposed logic)

That would still leave the fingerprint vector if the spy node sent a transaction and its replacement immediately after. That is, before we'd have a chance to announce or send the original to anyone. Leaving us with a unique fi
...

💬 kallerosenbaum commented on issue "bitcoin-qt generates bad psbt":
(https://github.com/bitcoin/bitcoin/issues/27697#issuecomment-1552966704)
Tested on version 25.0rc2. It works there.

💬 fanquake commented on issue "bitcoin-qt generates bad psbt":
(https://github.com/bitcoin/bitcoin/issues/27697#issuecomment-1552967514)
> Tested on version 25.0rc2. It works there.

Ok. We can close this then.

Maybe fixed with https://github.com/bitcoin-core/gui/pull/469 ?

✅ fanquake closed an issue: "bitcoin-qt generates bad psbt"
(https://github.com/bitcoin/bitcoin/issues/27697)

💬 kallerosenbaum commented on issue "bitcoin-qt generates bad psbt":
(https://github.com/bitcoin/bitcoin/issues/27697#issuecomment-1552978120)
A psbt file that's saved in 24.0.1 and that failed to open in 24.0.1, could successfully be opened in 25.0rc2. I guess this means that the saved psbt file is valid and that it's the parsing of the psbt file that is buggy.

💬 kallerosenbaum commented on issue "bitcoin-qt generates bad psbt":
(https://github.com/bitcoin/bitcoin/issues/27697#issuecomment-1552982794)
However, I'm not sure https://github.com/bitcoin-core/gui/pull/469 fixed it, because the saved psbt is binary. Attaching my psbt file (renamed to *.txt to circumvent upload restrictions)
[test3.psbt.txt](https://github.com/bitcoin/bitcoin/files/11507730/test3.psbt.txt)

👍 stickies-v approved a pull request: "test: Return dict in MiniWallet::send_to"
(https://github.com/bitcoin/bitcoin/pull/27640#pullrequestreview-1432623162)
Code review ACK faf4315c88d8c81c2ff84870bc81aef3cf719816

I'd have preferred doing things a bit differently but I think as per the PRs rationale, it's better to move away from the tuple return type quickly to avoid larger refactorings in the future. My comments can easily be implemented in follow-ups, if people want them.

---

> The idea is to use the same return type as the other MiniWallet methods.

Fair enough, I hadn't considered that.

> If you want compile-time type checks <...>
...

💬 kallerosenbaum commented on issue "bitcoin-qt generates bad psbt":
(https://github.com/bitcoin/bitcoin/issues/27697#issuecomment-1553009843)
If I convert the test3.psbt.txt (binary) file to a base64 file, I can then read the base64 into bitcoin core.
```
$ cat test3.psbt.txt|base64 -w0 > test3.psbt
```