I looked at the coding guidelines for C++ on the wiki but I didn’t see this topic addressed.
I know there is a preference not to use C-style explicit casting such as (double) int_variable. Is it okay to use C-style functional casting such as double (int_variable)?
Or, is the preference for the very long form C++ style static_cast<double> (int_variable)?
The current decision seems to prefer static_cast<double> (), as -Wold-style-cast is active by default, see https://hg.savannah.gnu.org/hgweb/octave/file/54509b7fddd3/configure.ac#l2684.
Personally, I like the longer notation too, as it documents what the programmer expects to happen (static_cast, const_cast, reinterpret_cast).
IIUC, the C-style casting and functional casting are basically equivalent (with the same difficult to deduce rules what actually happens with the cast argument).
I’d personally prefer the long C++ style cast. Even if that is quite verbose. But I believe that verbosity is necessary to make the code clearer and possibly avoid coding errors.
Using the longer C++ style cast syntax also makes it easier to find (and maybe eliminate) them. If possible, I would like to avoid casting unless it is really necessary so being able to locate them is helpful to me.
Do we want to maintain the coding conventions on the wiki? The motivation for this thread was that this design choice was undocumented, but it seems like there is a preference for the long form.
Documenting coding conventions would be good. A wiki page also seems like a good place to start. Do we have one already?
Yes, the guidelines for contributing patches, along with style guides for C/C++, m-files, and Fortran were removed from the Octave manual and now live on the Octave Wiki only. C++ style guide is at https://wiki.octave.org/C%2B%2B_style_guide. I made the change.