[vtk-developers] Request for a gentler tone in commit message verbiage...

David Lonie david.lonie at kitware.com
Wed Apr 29 14:12:07 EDT 2015


On Wed, Apr 29, 2015 at 1:50 PM, David Cole <DLRdave at aol.com> wrote:

> The commit message here:
>
>
> https://gitlab.kitware.com/vtk/vtk/commit/7247da0ecf23c2c52e90acc86dda2e2bbf33461a
>
> claims there was some sort of problem with make_pair and double. What
> exactly was the problem? With what exact version of MSVC? (I don't buy
> that there's a general problem with make_pair and double...)
>
> And then the follow on commit message:
>
>
> https://gitlab.kitware.com/vtk/vtk/commit/e3d0653de443435533ebfb2db3fa6513e3b5c86b
>
> seems a bit accusatory towards poor old closed-source MSVC.


> This code compiles just fine for me with Visual Studio 2013:
>
>     bool TestMakePair()
>     {
>         const double value = 0.0;
>         vtkTextProperty *prop = NULL;
>
>         std::pair<double, vtkTextProperty*> pairOne =
>             std::make_pair(value, prop);
>
>         return true;
>     }
>

Now try std::make_pair<double, vtkTextProperty*>(value, NULL).


> Of course, make_pair deduces its types from the supplied arguments, so
> passing "NULL" as the second arg is not exactly well defined across
> all OSes and compilers. Not sure why you would expect that.
>

Telling it the first argument is explicitly a double and the second is a
pointer, and then having it complain about the double with one syntax and
the pointer in the other (both while being explicitly declared in the
template arguements) is a tad ridiculous, IMO.


> Anyhow.... This email is simply a request to remove the disparaging
> remarks about "Windows" and MSVC from the comments, and to be kinder
> and gentler in future commit messages. Writing thoughtful explanatory
> commit messages is nearly as important as writing the code for a
> widely used public open source project.
>

Not sure they can be removed. I don't think this warrants rewriting the
commit history.


> Nobody likes dealing with "the other platforms" that they don't work
> on daily, but we all love how VTK works awesomely on so many
> platforms. It's worth the extra effort to get code just right so it
> compiles cleanly everywhere, and


I don't mind, actually, But if something ridiculous happens, I'm going to
find it ridiculous. Besides, "putting in the extra effort to get code
right" is exactly what those commits are doing...?


> commentary like this just turns me
> off.
>

Sorry to turn you off, Dave. I meant no offense, but apologize for any that
was taken. Writing code that works on every platform but one, and then
fails in two distinct ways on different version of the primary compiler for
that platform is surely frustrating, no? If GCC was behaving similarly, I'd
have similar frustrations.

I'm not trying to play favorites and disparage anyone's favorite platform.
There are annoyances with every compiler (admittedly, some more than
others!), and I view my commit messages as harmless venting. I still don't
see them as expressing anything offensive, just frustration at a couple of
idiosyncrasies. Less than absolutely professional? Sure. Restrained
expression of legitimate frustration? I'd say "sure" to that as well.
Offensive? Well, that's a bit of a stretch IMO. Different strokes.

Regardless, I'll make an effort to adopt a more neutral tone in my commit
messages.

Have a great day,
Dave
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20150429/4e242bc7/attachment.html>


More information about the vtk-developers mailing list