Re: [Monotone-devel] Typesafe VA_ARGS replacement for database::execute/fetch

[Vinzenz 'evilissimo' Feenstra]

Christof Petig schrieb:

> Vinzenz 'evilissimo' Feenstra wrote:
>   
> > I've tested the modifications now and corrected them. The source
> > compiles fine now.
> > So this should be the final diff for it :)
> >     
>
> What a mess ...
>
> If I compare the patch to the subject of this thread I feel a bit sick.
>
> While Nathaniel stated that std::string, CoW and % are ok with him, you
> rewrote your patch to use non-typesafe VA_ARGS using execute/fetch
> variants. I hope that you did not invest too much time into this
> modifications.
>
> I really liked to be able to remove .c_str() from the query parameter on
> several calls because I dumped VA_ARGS.
>
> And this misunderstanding/debate actually prevents sqlite3.binary from
> being included IIRC.
>
> Sigh
>    Christof
>
> I would vote for your original % (and std::string ?) based approach.
> Should I take the time and (ab)use your patches to my liking? Perhaps we
> should sort this out first.
Hi Chris,

It was meant as final diff for the VA_ARGS stuff, not for the whole 
issue. Although, it is buggy how I've figured out today.

Nathaniel now is thinking that the % operator way is better since the 
bugs that are in there came from missed text() suroundations. This won't 
happen if we would choose the operator% way since it is more typesafe 
and the compiler would complain about it.

So this couldn't happen again. This evening I'll going on to implement 
the op% solution again :)

BR
Vinzenz