Hi,
I've updated the caffe package in the git repo, one of the major changes is that
`python-caffe-cpu` was changed to `python3-caffe-cpu`.
On 2 June 2016 at 07:00, Ghislain Vaillant wrote:
>
>
> The build time testsuite, autopkgtest and piuparts serve different
> purposes.
Package:wnpp
Severity: wishlist
Owner: lumin
* Package name: lua-torch-cwrap
Version : 0~20160222-gdbd0a62
Upstream Author : Torch Developers
* URL : https://github.com/torch/cwrap
* License : BSD-3-Clause
Package:wnpp
Severity: wishlist
Owner: lumin
* Package name: lua-torch-paths
Version :
0~20160203-g68d579a
Upstream Author : Torch Developers
* URL : https://github.com/torch/
paths
* License : BSD-3-Clause
Package:wnpp
Severity: wishlist
Owner: lumin
* Package name: lua-torch-torch7
Version : 0~20160604-g69d7a01
Upstream Author : Torch Developers
* URL : https://github.com/torch/
torch7
* License : BSD-3-Clause
Package: wnpp
Severity: normal
I request assistance with maintaining the julia package.
Specifically I need a ppc64el porter (or anyone who has root access to
a ppc64el box) to help me:
1. Apply patch[1] to Debian's llvm-toolchain-6.0 (= 1:6.0.1-4) and build it.
2. Install the resulting
Package: lintian
Version: 2.5.98
Severity: normal
Justification: lintian requests a bug report
I got the following lintian E while building the source package
for tensorflow. The false-positive is requested by the lintian message.
This lintian report can be reproduced from this git repo:
Hi Tom,
Have you found any sponsor for the fzf package?
And are you still interested in maintaining it?
> https://github.com/tomfitzhenry/pkg-fzf
Best,
Package: wnpp
Severity: wishlist
Owner: Mo Zhou
* Package name: vim-julia
Version : 0.0~git20180821.120a0b6
Upstream Author : Carlo Baldassi and other contributors
* URL : https://github.com/JuliaEditorSupport/julia-vim
* License : Expat
Programming Lang
control: retitle -1 RFP: pscircle -- visualizing Linux processes in a form of
radial tree
control: owner -1 wnpp.debian.org
This software looks not quite mature at the time of ITP submission.
> Currently I can build it manually on my daily Debian experimental
> system (amd64) and another unclean chroot (amd64). However I'm
> still not sure whether the other can build it successfully like I do.
Preliminary lintian-clean binary packages are available on debomatic-amd64:
Hi Dmitry,
Thanks for the update!
On Fri, Aug 31, 2018 at 03:41:01PM +0300, Dmitry Eremin-Solenikov wrote:
> Hello,
>
> I've uploaded new 1.19.0.2-1 version to mentors.d.o.
> I've added manpages, fixed copyright info, fixed alternatives
> and enabled auto-tests. Could you please review it?
Hi mentors,
Does anyone have idea about the following very last blocker for libtensorflow?
We are really ready to upload TF once the C API unit tests passed without fatal
error.
procedure to reproduce::
1. download libtensorflow-cc1.10 and libtensorflow-dev from
debomatic-amd64 and
On Sun, Sep 09, 2018 at 12:25:54PM +0200, Giulio Paci wrote:
> Hi Mo,
> thank you for your offer and review.
> Unfortunately I do not know when I will be able to handle these issues as I
> have very limited time at the moment.
>
> And the latest standards version is 4.2.1 .
>
> As you can
control: owner -1 !
control: tags -1 +moreinfo
Hi Giulio,
I can sponsor this package. However this package looks old and needs
updating before the upload.
1. please have a look and fix the following lintian warnings:
-
│ W: mitlm
control: retitle -1 RFS: tensorflow/1.10.1+dfsg-A1 [ITP]
control: tag -1 -moreinfo
Hello science team and mentors,
I did a right choice to write the python+ninja build system
from scratch (I call this build system TF-Shogun in the source code).
Now I started to sort out any possible FTBFS with
Package: dpkg
Version: 1.19.0.5+b1
Severity: normal
Justification: triggers confusing FTBFS under a certain condition
Procedure to reproduce:
1. clone https://salsa.debian.org/science-team/tensorflow
2. checkout lumin/A1u30 or equivalently
a2d2212c63bc23ada20bef0eeb6284e6f3a022ec
3. build
control: close -1
I'll sponsor myself shortly as long as nothing goes wrong in the last build.
This will be an ~300MB initial upload.
On Fri, Sep 07, 2018 at 01:23:03PM +, Mo Zhou wrote:
> Hi mentors,
>
> Does anyone have idea about the following very last blocker for libtensorf
control: owner -1 !
Hi Ghislain,
I can sponsor this now. Should wait for you to update the package
to the latest upstream verison, or check and upload it from git
repo as is?
I think the packaging repo is this one:
https://salsa.debian.org/python-team/modules/python-jsonrpc
BTW, why don't you
control: close -1
Justification: not a bug.
I'm aware of the breakage before uploading lua-moses to unstable
since the affected packages are maintained by me. This is not
a regression in lua-moses but simply an API bump.
I'll deal with lua-torch-nn soon, by updating it's function calls
related
Control: tag -1 +pending -wontfix
Hi Milan,
Things have changed a bit since I submitted this bug.
The 1.0.1-1 package waiting in NEW queue is linked against libblas.so.3
(which is a symlink pointing to openblas). The dependency of the
resulting package can be satisfied by installing any package
Package: julia
Version: 1.0.1-2
Hi Sebastien, do you have any plan to add ILP64 interface to OpenBLAS?
I acknowledge that bumping BLAS interface from LP64 to ILP64 is really a
hardwork under Debian's context, and will take a long time to transit.
If there is not much necessity to provide ILP64
t; https://github.com/adsr/mle/tree/debian
Oops. Thanks for the hint. Then you should put it to the
Vcs-Git and Vcs-Browser fields in debian/control.
> Thanks for the tip about the "Recommends" field. I will add optional
> runtime deps there.
>
> Adam
>
> On
Control: tag -1 +moreinfo
Hi Yaroslav,
I think this bug has been fixed already. At least I cannot reproduce
this issue with the latest 1.9.2-1 version (uploaded several hours ago)
May I close this bug?
and ILP64 interface
I will send a RFC proposal about the NEW ILP64 interface on science list.
[1] you can find it on salsa under science-team's namespace
On Fri, Oct 19, 2018 at 10:18:53AM +0200, Sébastien Villemot wrote:
> Hi,
>
> Le samedi 13 octobre 2018 à 13:22 +, Mo Zho
Package: wnpp
Severity: wishlist
Owner: Mo Zhou
* Package name: termbox
Version : 1.1.2+dfsg
Upstream Author : nsf
* URL : https://github.com/nsf/termbox
* License : Expat
Programming Lang: C, Python
Description : Library for writing text-based user
Package: libgmp-dev
Version: 2:6.1.2+dfsg-3
Severity: important
Tag: patch
Justification: affects Julia's functionality on non-x86 arches
This patch is cherry-picked from upstream repo:
https://gmplib.org/repo/gmp/rev/c59c3879f982
Related discussions:
Hi Adam,
Thanks for this package, the copyright file looks good to me.
However there are still a couple of problems:
1. Can we avoid shipping with embedded code copy[1] of lua, termbox and
uthash? . lua 5.3 and uthash are already in the archive. termbox needs
to be packaged.
2. I'd
Package: libopenblas-dev
Version: 0.3.3+ds-1
Severity: wishlist
Hi Sebastien,
The range between priority numbers of OpenBLAS and Atlas is a bit narrow.
I think it isn't enough for furture usage when BLIS landed onto the
archive, especially when we wanted to provide libraries in different
On Sat, Oct 13, 2018 at 06:03:48PM -0400, a...@php.net wrote:
> Yes, I can make Lua and uthash package deps instead of embedding them.
> Is it ok to embed termbox for now, or should we consider that a
> blocker?
Try to build against this termbox package, it's almost finished:
Package: libtbb-dev
Version: 2017~U7-8
Severity: important
X-Debbugs-CC: steven.cap...@gmail.com,
debian-science-maintain...@lists.alioth.debian.org,
debian-scie...@lists.debian.org
tbb packaging repo is hosted under debian-science team's namespace on
Salsa. However the maintainer is not set
On Tue, Oct 23, 2018 at 08:42:31PM +0100, Steven Capper wrote:
> Hi Mo,
> I maintain the package and we had it hosted on Debian Science. I would
> like to keep maintaining the package, but I am happy to co-maintain
> the package with others.
Thank you for the quick reply. I'm happy to co-maintain
Package: libdsfmt-dev
Version: 2.2.3+dfsg-3
Severity: important
https://salsa.debian.org/julia-team/julia/blob/master/deps/patches/dSFMT.c.patch
https://salsa.debian.org/julia-team/julia/blob/master/deps/patches/dSFMT.h.patch
This change is similar to changes about BLAS's 64-bit-index version.
Package: wnpp
Severity: wishlist
* Package name: rover
Version : 0.1
Upstream Author : myself
* URL : https://salsa.debian.org/debian/rover
* License : GPL-3+
Programming Lang: Python
Description : text-based light-weight frontend for
Package: openrc
Version: 0.38-2
Severity: serious
On my amd64 virtual machine rc-status failes every time I run it.
Some times it just reports
double free or corruption (!prev)
And sometimes it reports
malloc(): smallbin double linked list corrupted
I think this is 100% reproducible by
Hi Sylvestre,
F.Y.I.
> * llvm 6.0 upstream won't have a new upstream release * I have been
> focusing my Debian effort on 7 for a while, so, packaging is a much
> better shape
Julia upstream backported many patches from llvm-7 to their vendored
llvm-6 source. However it's llvm-7 support is
control: tags -1 -patch +moreinfo
Hi Antonio,
I cannot find the patch anymore. Do we still need to fix the ZoL
upstream deb -> Debian deb transition problem?
control: close -1
This is a fake bug.
Patches are already applied by Peter 3 years ago.
control: retitle -1 RFP: nvidia-cudnn -- NVIDIA CUDA Deep Neural Network library
control: owner -1 w...@debian.org
control: close 887728
I'm no longer interested in reading Nvidia's license.
Packaging scripts are avaialble on Salsa.
Package: python3-sklearn
Version: 0.20.0+dfsg-1
Severity: serious
https://buildd.debian.org/status/package.php?p=scikit-learn
Package: wnpp
Severity: wishlist
Owner: lu...@debian.org
* Package name: zsh-autosuggestions
Version : 0.4.3
Upstream Author : Thiago de Arruda, Eric Freese
* URL : https://github.com/zsh-users/zsh-autosuggestions
* License : Expat
Programming Lang: zsh
control: tags -1 +moreinfo
Please hold on the CUDA version for a while. Some important applications
are still not fully ready for CUDA 10. Since recently Nvidia started to
produce buggy software and buggy hardware, I think we should slow down
our pace and only update CUDA when necessary, or when
Package: wnpp
Severity: normal
I plan to drop this package after releasing
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and I'm moving to Pytorch.
I cannot maintain more than 3 deep learning frameworks at the same time,
so I wish someone to take over the maintenance.
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and I'm moving to Pytorch.
I cannot maintain more than 3 deep learning frameworks at the same time,
so I wish someone to take over the maintenance.
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and I'm moving to Pytorch. I
cannot maintain more than 3 deep learning frameworks at the same time,
so I wish someone to take over the maintenance.
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and I'm moving to Pytorch. I
cannot maintain more than 3 deep learning frameworks at the same time,
so I wish someone to take over the maintenance.
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and I'm moving to Pytorch.
I cannot maintain more than 3 deep learning frameworks at the same time,
so I wish someone to take over the maintenance.
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and
Package: wnpp
Severity: normal
LuaTorch has been deprecated by upstream and
Hi anbe and nvidia team,
During debconf Ximin and I have figured out a workaround to make NVCC
work with GCC-8. Using this workaround I have successfully compiled
src:caffe-contrib.
CXXFLAGS += -std=c++03 -D__GNUC__=8
We need to make GCC-8 disguise to be GCC-6, and downgrade CXX standard
to
requirement to (>= 3.5)
+ * Downgrade debhelper compat level to 10.
+ * Enable control.bpo.py script in rules.
+ * Backport to stretch.
+
+ -- Mo Zhou Tue, 25 Sep 2018 08:19:29 +
+
intel-mkl (2019.0.117-2) unstable; urgency=medium
* Export HOME=/tmp/ to fix FTBFS due to the failure that rpm
Package: wnpp
Severity: wishlist
X-Debbugs-CC: debian-scie...@lists.debian.org
Owner: Mo Zhou
* Package name: blis
Version : 0.4.1
Upstream Author : The University of Texas at Austin, HP Enterprise, AMD Inc.
* URL : https://github.com/flame/blis
* License
Package: gcc-8
Version: 8.2.0-7
Severity: important
This happend on our buildd and Ubuntu buildfarm, and I reproduced it in my
qemu-armhf chroot:
qemu: Unsupported syscall: 385
Invalid Thumb instruction at 0xff5ef780: 0xee4c, 0x000e
signal (4): Illegal instruction
in expression
Update on the way to reproduce:
Repo: https://salsa.debian.org/julia-team/julia
tag: debian/1.0.0-3
Way to reproduce:
1. same as the last mail, revert
https://salsa.debian.org/julia-team/julia/commit/b89c05c87f27d23513cb3f90cd534f775daea289
2. remove -marm from rules.
On Sun, Sep 23, 2018
Package: libnvblas9.2
Version: 9.2.148-1
Severity: wishlist
X-Debbugs-CC: debian-scie...@lists.debian.org
Nvidia says nvBLAS is a "drop-in" BLAS library however it merely
provides several level-3 BLAS symbols. That being said, nvBLAS
is indeed a valid candidate for libblas.so.3 if explicitly
Package: nvidia-cuda-toolkit
Version: 9.2.148-1
Severity: wishlist
Justification: new upstream release
https://docs.nvidia.com/cuda/cuda-toolkit-release-notes/index.html
Some notable points from the release notes:
1. CUDA 10.0.130, requires nvidia >= 410.48
2. CUDA 10.0 adds support for the
Package: intel-mkl
Version: 2018.3.222-3
Severity: wishlist
https://software.intel.com/en-us/articles/intel-math-kernel-library-release-notes-and-new-features
I should also write a watch file monitoring this page:
https://pypi.org/project/mkl/
simple benchmark result here:
https://github.com/flame/blis/issues/255
On Mon, Sep 24, 2018 at 07:06:44AM +0000, Mo Zhou wrote:
> Package: wnpp
> Severity: wishlist
> X-Debbugs-CC: debian-scie...@lists.debian.org
> Owner: Mo Zhou
>
> * Package name: blis
> Version
Package: intel-mkl
Version: 2019.0.117-1
Severity: wishlist
2019.0.117-1~bpo9+1
On Tue, Oct 23, 2018 at 08:42:31PM +0100, Steven Capper wrote:
> I maintain the package and we had it hosted on Debian Science. I would
> like to keep maintaining the package, but I am happy to co-maintain
> the package with others.
I added myself as a co-maint.
> Out of interest, what would
- Forwarded message -
Hi Rich,
I investigated into this issue a bit and it looks like a result of
messy system where systemd-sysv and insserv are co-installed.
In insserv/sid, the postinst process will nolonger fail even if the
same error occurs. The error will disappear if you remove
Source: libunwind
Version: 1.2.1-8
Severity: normal
Hi Adrian,
Julia upstream applied several patches[1] to their vendored libunwind,
but it seems that none of these patches was applied to Debian's
libunwind package.
This is a list of patches:
(1) libunwind-arm-dyn.patch
(2)
control: severity -1 wishlist
control: tag -1 wontfix
Moving headers out of mkl/ subsirectory would mess up user's /usr/include
Hi,
In addition, upstream is likely to release libunwind 1.3 after some
holidays in january[1]. The timing sounds bad for the Buster freeze,
but I think a freeze exception is still feasible.
[1] https://github.com/libunwind/libunwind/issues/97
Package: git-buildpackage
Version: 0.9.13
Severity: important
X-Debbugs-CC: Mattia Rizzolo
Dear maintainer,
I noticed that the pristine-tar tarball id for the source is incorrect
if the set of orig-tarballs are imported with
gbp-import-orig XXX.orig.tar.xz --components YYY
For example,
Package: libopenmpi3
Version: 3.1.3-7
Severity: serious
Dear maintainer,
You missed a space in the script, which resulted in
The following packages will be upgraded:
libopenmpi3 openmpi-common
2 upgraded, 0 newly installed, 0 to remove and 4 not upgraded.
210 not fully installed or removed.
I guess sshfs 3.X is hopeless for Buster.
Because there is no any related transition on the tracker.
And fuse3 breaks gnome-core.
Anyone who cannot wait for the official package could try mine:
https://salsa.debian.org/lumin/sshfs-fuse (3.5.1)
On Sun, Dec 30, 2018 at 12:53:18PM +0100, Guido Günther wrote:
> > I noticed that the pristine-tar tarball id for the source is incorrect
> > if the set of orig-tarballs are imported with
> >
> > gbp-import-orig XXX.orig.tar.xz --components YYY
> >
> > For example, after importing the opencv
Hi anbe,
Do you still have any idea about the way to reproduce this failure?
I've added an autopkgtest script to zfs to test the dkms build
against linux-headers-$(dpkg --print-architecture), and it seems
that I cannot reproduce this issue in a clean and minimal chroot.
Package: fish
Version: 2.7.1-3+b1
X-Debbugs-CC: Tristan Seligmann
Hi Tristan,
fish is my default login shell and I wish to use fish 3.0 and push
fish 3.0.X into Buster. So following the ITS process I'm filing this
bug.
not buildd.
Builds for 3.4.5+dfsg-1~exp1 were successful so I don't expect any
problem for 3.4.5+dfsg-1~exp1+b1 .
On Mon, Dec 31, 2018 at 10:14:20AM +, Mo Zhou wrote:
> I've uploaded opencv 3.4.5 to experimental which fixed some issues found
> in the previous version. This time
On Sun, Jan 06, 2019 at 07:56:11PM +0100, Fabio Pedretti wrote:
> I don't think a transition is needed, now that fuse and fuse3 are
> co-installable. The packages that can migrate to fuse3 could do, the others
Source of "co-installable"???
I just installed fuse3, then apt removed gnome-core and
On Sun, Jan 13, 2019 at 08:06:57PM +0100, Emilio Pozuelo Monfort wrote:
>
> What is the status with the rdeps? I looked at two bugs and they worry me:
I haven't had enough time to test rdeps for another round. But I guess
the situation would be similar to the first round.
> #915544 suggests the
lua-torch-cutorch is deprecated by upstream and orphaned by me:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=916281
caffe-contrib FTBFS because libopencv-dev cannot be installed,
libopencv-dev cannot be installed because there is a temporary FTBFS
against hdf5 where libvtk6-dev is
Package: ftp.debian.org
Severity: normal
rm src:lua-torch-cutorch
It's deprecated by upstream, and has already been replaced by newer
alternatives. Since I'm not going to maintain it anymore and I don't
expect anyone to take it over, I'm requesting a removal.
Hi Norbert,
> It seems that the current Julia does not like the mkl blas libraries.
Our prebuild Julia package cannot depend on MKL. It assumes that the
user use openblas by default.
For MKL users, there is a CUSTOM_MKL flag provided by julia's
debian/rules file, with which one can easily
See comments I left in github issues.
On Fri, Dec 21, 2018 at 03:34:40PM -0500, a...@php.net wrote:
> On Wed, 17 Oct 2018 13:28:40 +0000 Mo Zhou wrote:
> > ...
> > Try to build against this termbox package, it's almost finished:
> > https://salsa.debian.org/debian/ter
On Wed, Dec 19, 2018 at 11:32:36AM +0100, Emilio Pozuelo Monfort wrote:
> I was going to ack this, but I noticed that opencv failed to build on some
> architectures:
>
> https://buildd.debian.org/status/package.php?p=opencv=experimental
>
> Please look at that before we start this.
opencv was
OpenCV 3.4.4 is green on all official architectures after uploading
manually built mips{,64}el packages by using Yunqiang's machine.
Shall we move on?
On Mon, Dec 24, 2018 at 01:49:45AM +, Mo Zhou wrote:
> On Wed, Dec 19, 2018 at 11:32:36AM +0100, Emilio Pozuelo Monfort wrote:
&g
Package: wnpp
Severity: wishlist
Owner: Mo Zhou
* Package name: gym
Version : 0.10.9
Upstream Author : OpenAI
* URL : https://gym.openai.com/
* License : MIT/Expat
Programming Lang: Python
Description : A toolkit for developing and comparing
Package: ros-opencv-apps
Version: 1.12.0-1
Severity: important
Tags: patch
https://patch-diff.githubusercontent.com/raw/ros-perception/opencv_apps/pull/71.diff
Package: wnpp
Severity: normal
X-Debbugs-CC: debian-scie...@lists.debian.org
As time goes by, the DL framework is more and more likely an educational
code base. It's architecture and design is old, compared to PyTorch or
TensorFlow 2.0.
The last time I use caffe for neural network training is
control: tags -1 +fixed-in-experimental
3.4.4+dfsg-1~exp1
Hi Anbe,
Do you have any plan about Buster's CUDA version?
My Recommendation is 9.2 since 9.1 is buggy.
My most familiar use case of CUDA is PyTorch, which
omitted CUDA 9.1: https://pytorch.org/
I've fixed the opencv build failure in experimental due to the path
change of ant.
Fixed in version: 3.4.4+dfsg-1~exp1
control: tags -1 +fixed-in-experimental
3.4.4+dfsg-1~exp1
Source: gst-plugins-bad1.0
Version: 1.14.4-1
Severity: important
builddlog attached.
gst-plugins-bad1.0_1.14.4-1_amd64-2018-12-06T04:04:31Z.build.xz
Description: application/xz
Source: libkf5kface
Version: 17.08.1
Severity: important
builddlog attached.
libkf5kface_17.08.1-1_amd64-2018-12-06T04:21:30Z.build.xz
Description: application/xz
Source: mldemos
Version: 0.5.1+git.1.ee5d11f-4
Severity: important
builddlog attached.
mldemos_0.5.1+git.1.ee5d11f-4_amd64-2018-12-06T04:29:45Z.build.xz
Description: application/xz
Source: php-facedetect
Version: 1.1.0+git20170801-2
Severity: important
buildlog attached.
php-facedetect_1.1.0+git20170801-2_amd64-2018-12-06T06:23:12Z.build.xz
Description: application/xz
Package: rover
Version: 0.3.1
Rover will crash if I search with *
It's clear that we need to deal with malformed regex if there is any.
Traceback (most recent call last):
File "/usr/bin/rover", line 389, in
rv.refresh_selections()
File "/usr/bin/rover", line 204, in refresh_selections
1 - 100 of 434 matches
Mail list logo