Richard Russon

@ragu-manjegowda Hmm... I'm surprised that ever worked :-) There's a `10` (one zero) in the middle of your `set index_format...` that shouldn't be there. If I remove that, it works...

@Tonus1 Yes, that's the same problem. The parser is working on a conditional `%` that looks like its end marker.

Thanks for your patience. I _was_ working on a fix, but then disappeared on a 3 month hiking expedition :-) I'll make sure it's fixed for the next release.

If roccoblues is willing, I suggest: 1. Test the retval of every `scanf()` This means making a decision, for each case, on how to handle an error. `ASSERT()` is ok...

> in the current changes those [atoi functions] couldn't be used right? We _can_ use them, I was suggesting that we sort out the error paths first. s/we/you/ :-)

A quick skim of the changes and they look rational :-) How did you make the changes? Manually, [Coccinelle](https://github.com/neomutt/coccinelle), or some other tool? Edit: I'd have used vim-fu :-)

OK, I've read through **all** the changes and they look good... There's just one problem, the `typeof` you've added to `mutt/array.h`. Until C23, `typeof`s a GCC extension (supported by Clang)....

Thanks! I shall have a read and a play as soon as I can.

I had a go at implementing a _flag_ solution. Unfortunately, `struct ConfigDef` (where I'd be storing the flag) is passed `const` to the variable-changing functions.

I'd _prefer_ **not** to kill lua. Scripting has potential -- think of the difference between a text editor and Vim. Keeping nntp is harder for me to justify. It _does_...