On 05/05/15 15:19, Martin Lucina wrote:
On Tuesday, 05.05.2015 at 11:19, Antti Kantee wrote:
Do we want to keep the name "app-tools"? AFAIR, it was coined (by
Ian?) before we'd really settled in on other terminology. Is simply
rumprun-tools better?
I have no opinion. It's more of an "internal" name and seems as good as
any.
If nobody has an opinion, I'm pulling the "consistency matters to me"
card (and to probably everyone new to the project who's trying to do
some internal hacking and is wondering what's an "app" and why we need
"app-tools").
I'm a bit unhappy with "baremetal". It's long and not very
descriptive. At least we'll get rid of the extra "bmk" term, which
is good, but a better term for "baremetal" would be nice to invent.
How about just "metal" ?
It's shorter, but not really better. I was tossing around "hw" in my
head earlier, which is even shorter, but it's still a bad name since the
platform is not only for hardware.
I'd like to get the renaming done ASAP, so if anyone has objections to
the proposed scheme please speak up now! Final call :-)
How many real programs have you tested the scheme with?
I've tested incomplete, hand-made wrappers using this scheme with our
"configure" test and AFAIR also PHP but that was some time ago. I will do a
more thorough test and try and find some software that runs native
binaries during the build before committing anything.
Ok. A dozen or so varied programs should be sufficient. All of
php/nginx/mysql are already delightfully different.
Does anyone have suggestions on what should be tested?
I will also add in functionality to generate "Toolchain" files for CMake
and test with MySQL. These should make software using CMake also just work
(assuming cross-compilation support in the build system).
Sounds cool, though I have no idea what that means ;)
