Re: Is there going to be a post-5.9 ncurses release in the near future?

[Fotis Georgatos]

Hi Thomas,

I guess you'd be interested then to eventually use the first column facilities of the EasyBuild tool:

http://easybuild.readthedocs.org/en/latest/eb_list_toolchains.html?highlight=list%20toolchains

very convenient and clean way to test against multiple versions of gcc, icc, clang and what have you.

Basically, you can largely automate the production of software builds

and arbitrarily change later things underneath, such as build modes, dependencies etc.

I hope it helps you in some way - in return, I hope for baptising (i.e. tagging) the recent release 5.0xA or such ;-)

cheers, F.

On 1 May 2015 at 11:24, Thomas Dickey <address@hidden> wrote:

On Wed, Apr 29, 2015 at 02:29:19PM +0100, Fotis Georgatos wrote:
> Hi Thomas,
>
> yes, thanks for the update!
>
> Am I wrong to think that distros have not really caught up widely on
> following up the patch-dates?
> IMHO, patch-dates appear as more of a developer thingy, rather than release
> thingy.

partly - some packagers ignore whatever versions I mark, and add development
changes to an existing package indefinitely.  That's why I added version
options to the utilities as well as the curses_version() function to the
library...

> Anyway, I wonder how you test ncurses across different distros/unices etc;
> what are your preferred test modes and expected coverage?
> I'm asking this, to see to  see if we may copy good practices elsewhere...

probably not much to learn: I have a lot of virtual machines which I turn
on as needed to check portability of changes.  On a given machine, I run
scripts to compile a half-dozen or so configurations, and keep the logs
in local rcs so that I can spot obvious regressions.  The machines I select
and the tests I run are ad hoc, to focus on what I'm working on.  (I have
about 90 machines in my local DNS, but only visit a dozen or so machines in
a week).

On a different dimension, I test-compile with different compilers (clang,
icc, as well as occasional visits to a guest account on some Unix boxes).

> cheers,
> Fotis
>
>
> On 28 April 2015 at 02:03, Thomas Dickey <address@hidden> wrote:
>
> > On Mon, Apr 27, 2015 at 12:07:49PM +0100, Fotis Georgatos wrote:
> > > Hi,
> > >
> > > In relation to C++ and ncurses, ref. (I'm not expert in this):
> > >
> > https://github.com/hpcugent/easybuild-easyconfigs/pull/225#issuecomment-16603167
> > >
> > > is there going to be any release of the ncurses/5.9 or ncurses/5.A type?!
> >
> > I am working toward that goal, yes.  I have at least one more iteration
> > on the configure/build-scripts after fixing some regressions from the
> > current phase (which is to get some better testing of packaged versions
> > of ncurses).
> >
> > > I'm impartial to ncurses practices, however I am with the impression that
> > > many/most linux distros have been largely stuck to 5.9:
> > >
> > http://www.rpmfind.net/linux/rpm2html/search.php?query=ncurses&submit=Search+
> >
> > Actually, "stuck" is what I'm working to address with the symbol versioning
> > and so forth.  But if you notice, those "5.9" packages have a variety of
> > patch-dates...
> >
> > --
> > Thomas E. Dickey <address@hidden>
> > http://invisible-island.net
> > ftp://invisible-island.net
> >
> > -----BEGIN PGP SIGNATURE-----
> > Version: GnuPG v1.4.10 (GNU/Linux)
> >
> > iEYEARECAAYFAlU+3EQACgkQcCNT4Pfkjtv05QCfdKxzW44wIW/6psgo/oKJdfEN
> > iRgAoIPsxGvahz8JN0v8DA2U6FTUIxSN
> > =QbGG
> > -----END PGP SIGNATURE-----
> >
> >
>
>
> --
>
>
> echo "sysadmin know better bash than english"|sed s/min/mins/ \
>   | sed 's/better bash/bash better/' # signal detected in a CERN forum

--
Thomas E. Dickey <address@hidden>
http://invisible-island.net
ftp://invisible-island.net

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)

iEYEARECAAYFAlVDVEoACgkQcCNT4PfkjttpkQCfS1eJCLFTJR+UCpUpcCHG/ELa
KWUAn3jZus9KkfPp8gJV4sUMfmp7kEjO
=fZPA
-----END PGP SIGNATURE-----

--

 

echo "sysadmin know better bash than english"|sed s/min/mins/ \
  | sed 's/better bash/bash better/' # signal detected in a CERN forum