Bug#905826: ILP64 interface of openblas?

2018-10-13 Thread Milan Bouchet-Valat
ckaging script. > On Sat, Oct 13, 2018 at 11:38:42AM +0200, Milan Bouchet-Valat wrote: > OK, thanks, I guess it's fine as long as the default is OpenBLAS. > BTW, it would be very useful to build with USE_BLAS64=1 > OPENBLAS_SYMBOLSUFFIX=64_ to use an ILP64 BLAS. Without this, >

Bug#905826: libjulia depends on libblas.so.3 instead of openblas?

2018-10-13 Thread Milan Bouchet-Valat
OK, thanks, I guess it's fine as long as the default is OpenBLAS. BTW, it would be very useful to build with USE_BLAS64=1 OPENBLAS_SYMBOLSUFFIX=64_ to use an ILP64 BLAS. Without this, packages using BinaryBuilder/BinaryProvider which call BLAS won't install. This affects notably Arpack.jl, which i

Bug#905826: libjulia depends on libblas.so.3 instead of openblas?

2018-10-11 Thread Milan Bouchet-Valat
I second this (and I'm sure this reflects the opinion of upstream in general). It's really important that after installing the julia package one gets the same performance as with official upstream binaries. It's a trap to use the Netlib reference BLAS/LAPACK by default, which is known to be terribl

Bug#910784: Add optimized JULIA_CPU_TARGETs

2018-10-11 Thread Milan Bouchet-Valat
Package: julia Version: 0.7.0-2 Severity: normal Hi, I've noticed that the package currently builds a single sysimage for the baseline CPU target (i.e. pentium4 and x86-64). Recent Julia versions allow specifying multiple targets via JULIA_CPU_TARGET. This allows building one image for the baseli

Bug#832798: [Pkg-julia-devel] Bug#839668: julia: New upstream release

2018-04-20 Thread Milan Bouchet-Valat
Any news? Unless you willing to backport libgit2 patches to use MbedTLS (which appear to be ready upstream) soon, it would probably be better to remove the julia package, which is now totally outdated. Julia 0.4.7 can only be confusing for users as the language has changed a lot since then.

Bug#545714: deluser does not remove empty main group when the suer is member of it

2011-11-23 Thread Milan Bouchet-Valat
Le mercredi 23 novembre 2011 à 09:52 +0100, Marc Haber a écrit : > retitle #545714 deluser does not remove empty main group when the user is > member of it > thanks > > Hi Milan, > > On Wed, May 26, 2010 at 11:03:19AM +0200, C. Gatzemeier wrote: > > Does that answer your question so you can > >

Bug#604242: [Adduser-devel] Bug#604242: [patch] Allow dots in username by default

2011-11-23 Thread Milan Bouchet-Valat
users-admin has been replaced with a gnome-control-center applet in GNOME 3, and it uses useradd, so I don't really care now. Feel free to close the bug. (Note though that useradd seems to allow dots in usernames on most platforms, which probably suffer from the same potential problems as Debian.)

Bug#604242: [Adduser-devel] Bug#604242: [patch] Allow dots in username by default

2010-11-21 Thread Milan Bouchet-Valat
Le dimanche 21 novembre 2010 à 17:42 +, Stephen Gran a écrit : > - $configref->{"name_regex"} = "^[a-z][-a-z0-9_]*\$"; > - $configref->{"name_regex_system"} = "^[A-Za-z][-A-Za-z0-9_]*\$"; > + $configref->{"name_regex"} = "^[a-z][-a-z0-9_.]*\$"; > + $configref->{"name_regex_system"} = "^[A-Z

Bug#604242: [patch] Allow dots in username by default

2010-11-21 Thread Milan Bouchet-Valat
Package: adduser Version: 3.112 I don't think there's a reason not to allow usernames to contain dots (.), as they are also allowed to contain underscores (_) and dashes (-). useradd accepts this. Attached is a patch that makes dots be allowed just as other special chars. One could also add the p

Bug#594775: ALSA modem detection in init script broken in non-English locales

2010-08-29 Thread Milan Bouchet-Valat
Le dimanche 29 août 2010 à 14:53 +0200, أحمد المحمودي a écrit : > Hello, > > Thanks for the patch, I have applied it now. Nice, thanks for your responsiveness! > On Sun, Aug 29, 2010 at 01:05:36PM +0200, Milan Bouchet-Valat wrote: > > For reference, people with amd64 versions

Bug#594775: ALSA modem detection in init script broken in non-English locales

2010-08-29 Thread Milan Bouchet-Valat
Package: sl-modem Version: 2.9.11~20100718-1 Reported against Ubuntu, with a patch available at http://bazaar.launchpad.net/% 7Enalimilan/ubuntu/lucid/sl-modem/sl-modem-regex/diff/21 Hope this helps! The /etc/init.d/sl-modem-daemon script checks for the presence of modems by parsing the output o

Bug#545714: deluser does not remove empty main group when the suer is member of it

2009-09-08 Thread Milan Bouchet-Valat
Package: adduser Version: 3.110 Hi! I've hit a problem when trying to fix a bug in the system-tools-backends/gnome-system-tools, and I discovered a strange behavior from deluser. If you add an user to its main group, deluser won't automatically remove that group when deleting the user, even if he

Bug#525731: Don't start s-t-b at boot time

2009-04-26 Thread Milan Bouchet-Valat
Package: system-tools-backends Version: 2.6.0-2 Severity: minor I'm wondering why the system-tools-backends are loaded at boot time. They're only used through D-Bus, so activation is here to start them when needed, isn't it? I've tried stopping them and running the gnome-system-tools, and that

Bug#519810: Please build gnunet-server using --with-microhttpd=yes

2009-03-15 Thread Milan Bouchet-Valat
Package: gnunet-server Version: 0.8.0b-5 Well, I guess you know everything! ;-) Building gnunet with the option --with-microhttpd=yes at configure time will enable the HTTP transport, which is really useful. That will add a dependency on libmicrohttpd5. You should then change the line TRANSPORTS

Bug#499759: Known limitation in GTK+

2009-01-05 Thread Milan Bouchet-Valat
This is a known limitation in GTK+, not a bug in Evince. See http://bugzilla.gnome.org/show_bug.cgi?id=384940 -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#499496: closed by Daniel Baumann <[EMAIL PROTECTED]> (Bug#499496: fixed in gnunet-gtk 0.8.0b-3)

2008-09-19 Thread Milan Bouchet-Valat
Le vendredi 19 septembre 2008 à 11:54 +0200, Daniel Baumann a écrit : > Milan Bouchet-Valat wrote: > > Though, I fear I still have to complain : why did you only choose to > > suggest gnunet-server? > > recommends is too strong in this case, gnunet-server should really n

Bug#499496: closed by Daniel Baumann <[EMAIL PROTECTED]> (Bug#499496: fixed in gnunet-gtk 0.8.0b-3)

2008-09-19 Thread Milan Bouchet-Valat
I always admire you responsiveness ! ;-) Though, I fear I still have to complain : why did you only choose to suggest gnunet-server? The interest of *recommending* it was that on standard desktops, gnunet-server will be automatically installed when the user doesn't know what to do. On the cont

Bug#499496: gnunet-gtk should recommend gnunet-server

2008-09-19 Thread Milan Bouchet-Valat
Package: gnunet-gtk Version: 0.8.0b-2 I think gnunet-gtk should really recommend gnunet-server. When base users want to use GNUnet, they get the GUI, and then they experience problems trying to use it. They're not aware of the fact that gnunetd is not installed. We can't make gnunet-server a depe