Guillaume Maudoux

But then, we cannot reasonably keep a write lock on the database for the whole duration of the `nix search` command. Would you consider a patch that performs only reads,...

It is really difficult to say, given the codebase changed quite a lot. > * Output several more locations during --show-trace This one seems similar to my changes, although more...

Hi, having a look right now. There are plenty of potential fixes. Finding failing tests would be tremendously useful indeed. Any help welcome. For now I am focusing on reproducing...

So first there is indeed a change where I introduced "frame" traces (pardon the naming here). Think of them as more important trace points, like function calls and builtins.derivation calls....

There are no particular worries here. I wanted to report a fact that may come as a surprise to other users. The database's size has the same order of magnitude...

Hi, I just got bitten by this one, and it hurts pretty bad given I knew I added the flags, I just missed that damn "build" in front of them....

So, I have #20895 addressing this issue. It is a bit bold, but anything more involved requires the client to know the list of valid commands, which introduces redundancy. I...

It seems that there is a difference between execution within the shell and from the command line: ``` (zsh)$ ./fish -c "read n | grep nothing" read> lol (zsh)$ ./fish...

> This change looks great! I did notice it still doesn't diagnose this invalid case: > > ``` > build --copt=-v \ > # comment \ > --copt=-invalid > ```...

> In the single token case, we could use an empty string. Then the server would error if the single token is not a valid command name. That would push...