Is there any timeline regarding the implementation of Fortran 202x?
As another poster sometimes says, that depends on how many people volunteer or how much money is raised for paid work on gfortran. Koenig commented in comp.lang.fortran on what he may work on.
Is there a way to donate/raise money? The recent thread on Catalyzing gfortran development didnāt gain much traction.
I think if everyone of us can contribute just 5$ a month it would be great donation. Imagine 200 of us contributing 5$ a month for an year itās 12000$. Letās conduct a poll on how many are willing to donate 5$ a month to Fortran development.
C++ dudes are also complaining.
Install GCC 12 on MacOS here: Releases Ā· fxcoudert/gfortran-for-macOS Ā· GitHub
If youāre using Spack package manager you can install it using the following command:
spack install gcc@12.1.0
Edit: fixed version in command (was 12.0.1)
Homebrew (both MacOS and Linux) stays at 11.3.
I have detected a new warning with GFortran 12: when a variable may be used uninitialized, it tells you on which line it was declared:
/home/osboxes/gtk-fortran/src/gtk-hl-tree.f90:1871:34:
1871 | call g_value_unset(pstring)
| ^
Warning: āpstringā may be used uninitialized [-Wmaybe-uninitialized]
/home/osboxes/gtk-fortran/src/gtk-hl-tree.f90:1837:33:
1837 | type(c_ptr) :: store, pstring, col, rlist, renderer, pmodel, model
| ^
note: āpstringā was declared here
GFortran 12.1.1 is now available in Fedora Rawhide.
I believe it already did, but maybe itās on by default now instead of requiring an option?
Yes, youāre right as I have found the same warning in that March 2021 post :
https://fortran-lang.discourse.group/t/interesting-output-from-gfortran-when-using-parameterized-derived-types/919
The GCC and GFortran documentations are now rendered with Sphinx & Furo:
Updated 2022-11-18: the doc has been reverted to its old rendering. We can just hope this is temporary.
This change is merged for GCC 13 with -O2. The GCC 13.1 stable release should be out in its usual March~April timeframe.
At the moment, in the changelog page the word āFortranā appears only one time, related to OpenMP: āFortran now supports non-rectangular loop nests, which were added for C/C++ in GCC 11.ā
Iām curious, are those the only changes in GFortran? I vaguely recall some Fortran patches submitted by @kargl for import host association, did they not make it in, or is the Fortran Changelog section out of date?
I guess youāre right, this document does not yet include changes in the non-C/C++ compilers of the collection, as the stable release date is still far away. If I make a print preview of the GCC 12 release page, I obtain around 10 pages, with Ada, D, Fortran, etc. For GCC 13, only 3.5 pages by nowā¦
Concerning GFortran, I donāt know if it is the right place to look at, but I have found that changelog:
https://gcc.gnu.org/git/?p=gcc.git;a=blob_plain;f=libgfortran/ChangeLog;hb=refs/heads/master
We can also search āFortranā related commits here:
https://gcc.gnu.org/git/?p=gcc.git&a=search&h=HEAD&st=commit&s=fortran
Another solution to look at changes:
Assuming you have Git installed, you can check out the GCC sources using the following command:
git clone git://gcc.gnu.org/git/gcc.git SomeLocalDir
And since GCC 12 the -O2
optimization option has flag -ftree-vectorize
switched on by defalt.
Is GitHub - gcc-mirror/gcc the best entry point for submitting pull requests?
Thatās a shame, If you have the will and time to point me/others to those patches we could open merge requests and get them into GFortran. It is a real shame for peopleās completed work to not be recognised and used because of a tool change.
Thanks, I did not see the fortran
directory in the overcrowded gcc
dirā¦
https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/fortran/ChangeLog;h=08ed6caefa238d6c766455b8ae3dd695c6f40717;hb=HEAD