The overly greedy blue-comment star matcher would make all C and C++
code in doxygen look good, but it messes with non-doxygen block comments.
By changing to this `containedin` blocking, we pickup the trailing
`)` and `}` characters that seem to be skipped if just doing the
ticked-region contained requirement.
Now the `'*'` characters in fixed-width blocks show up correctly,
with respect to how I write code. (I think doxygen also swallows
them anyhow, so the highlighting is more consistent with doxygen's
parser.)
Also on loading a file, everything looks better now.
I made `cpp.cplusplus` a hybrid filetype which tricks the vim
highlighting into believing that we're really in raw C++ for the
purposes of the core vim highlighting rules. My `cplusplus` type
adds more bits to the core stuff, but mostly exists to let
doxygen slip in the way that I want it to.
This "cplusplus" filetype permits the sourcing of `cpp.vim`
and `doxygen.vim` as lower-level files from itself. This should
permit `doxygen.vim` to source from `cpp.vim` and avoid endless
recursion, when handling code-in-comments formatting.
This now permits simpler installation and setup of my environment,
when the core configuration is stored at `~/cshenv`. If stored
elsewhere, the file would have to be edited. I'll think up an automatic
solution to this later.
This is kept, for now, in `~/.gitconfig` thus:
```
[commit]
gpgSign = true
```
This might be migrated into the universal gitconfig, but
that would require key setup for ephemeral cases.
For ages I've had problems with 256-color xterm and vim. I never
used a proper colorscheme, but just modified the default I was
"magically" given. This changes all of that to make us now have
a proper color scheme -- the "adam" scheme.
Colors now work independent of console color depth, I think.
There are remaining quirks, but I'll have to add the requirements
as proper color mappings to my color scheme file.
In generating CSHENV prompts, I used stty(1) to find the width
of the current terminal, at sourcing time. SYSV and BSD use
different output formats. Presently I have tcshrc.prompt.defs
default to 80 columns when I don't know how to parse stty(1)'s
output. AIX is known to be a SYSV variant, and I've added a
case to handle this.
