On Sat, 2 Jan 2021 at 23:26, Felipe Contreras
<felipe.contreras@xxxxxxxxx> wrote:
>
> Martin Ågren wrote:
> > On Sat, 2 Jan 2021 at 18:43, Thomas Ackermann via GitGitGadget
> > <gitgitgadget@xxxxxxxxx> wrote:
> >
> > [Snip several typo fixes in Documentation/, all of which I agree with.]
> >
> > > diff --git a/contrib/buildsystems/CMakeLists.txt b/contrib/buildsystems/CMakeLists.txt
> > > index c151dd7257f..8f77baa678f 100644
> > > --- a/contrib/buildsystems/CMakeLists.txt
> > > +++ b/contrib/buildsystems/CMakeLists.txt
> >
> > Hmm... This does not match the "doc:" prefix of the patch.
> >
> > > @@ -442,7 +442,7 @@ endif()
> > > check_c_source_compiles("
> > > #include <regex.h>
> > > #ifndef REG_STARTEND
> > > -#error oops we dont have it
> > > +#error oops we don't have it
> > > #endif
> >
> > I don't think this is correct. This omission of the single quote somehow
> > looks like it's done on purpose. I don't build using this system, but I
> > tried making some silly code like this in another file, which I actually
> > do use:
> >
> > #if 0
> > #error might or mightn't work
> > #endif
>
> But this works:
>
> #error "might or mightn't work"
Yeah, I pondered writing something like that. But then I thought I would
also need to write something about how, at the least, it should still be
a separate patch, but also how even if it were a separate patch, I would
probably sort it under code churn (IMVHO).
The error is not really intended for anyone's eyes, at least assuming my
reading is correct. So I refrained from suggesting to Thomas that here's
how you could go about making such a change.
And now I ended up laying out my thoughts anyway. :-) Of course, others
may well disagree with me about this being code churn.
Martin
