i am sure the download counts is only a effect, that gcc2 hybride the only
vesion is, he is preffer, so users download this version.
When Haiku preffer the x86_64 and a new compiler gcc or llvm/clang then
will more people this download.
I think its a good to have one supportet version as rolling release, for me
i am happy when we have x86_64 and llvm/clang and some bsd userland tools
(POSIX singeluser).
I have no problem when we brake the beos r5 stuff.
I have no problem, when we release and frezz the API for Applikation
developers
2016-11-30 23:04 GMT+01:00 kallisti5 <kallisti5@xxxxxxxxxxx>:
On 2016-11-30 15:02, kallisti5 wrote:
I'm reviewing the download site and cleaning up some language around
changing to the anyboot iso.
Our current download pages are a complex maze of "use the stable, use
the nightly, use the stable, use the nightly!"
Are we at the point that we want to no longer recommend R1A4 to users?
It's *so* old now (4 years!).
We need a release... or to recommend nightlies.
Now on to the bigger elephant...
I personally still *really* want to see R1 to be x86 gcc4h (cough gcc5h).
I think we missed the gcc2h boat years ago.
gcc2h in 2006 made sense.
gcc2h in 2016 makes 0 sense.
The list of software that even compiles on gcc2 is shrinking
rapidly... why are we doing this to ourselves?
gcc4h (gcc5h now) means you can run BeOS apps, and have a native
modern compiler. The downside is users
won't be able to run BeOS replicants, drivers, or screensavers on R1.
At this point in time, I can't think of a single "must have 1999 BeOS
replicant" (that isn't open sourced).
There may be folks out there however i'm not thinking of.
So once again, why are we doing this to ourselves?
As a visual aide, here are our rough download counts from the nightly
web server for all of 2016 through today:
135052 x86_gcc2_hybrid
38904 x86_64
18909 x86
17389 x86_hybrid
15026 x86_gcc2
The higher x86_gcc2_hybrid is no surprise since our pages highly recommend
it.
x86_64 being higher is interesting, most likely savvy tech users laughing
at
our gcc2 recommendation :-)
-- Alex
