Merge octave 4.0.2-1 (universe) from Debian testing (main)

Bug #1563100 reported by Amr Ibrahim
18
This bug affects 3 people
Affects Status Importance Assigned to Milestone
octave (Ubuntu)
Fix Released
Undecided
Unassigned

Bug Description

Please merge bugfix release octave 4.0.2-1 (universe) from Debian testing (main)
----------------------------------------------------------------------

* The package comes from Debian testing, which means that:
 - It compiles and up to date on all architectures it has previously been compiled for in unstable;
 - It does not have release-critical bugs;
 - All of its dependencies are satisfiable by packages already in testing, which is what the LTS is derived from;
 - Installing the package does not break any packages currently in testing
-------------------------------------------------------------------------

Bugs Fixed in GNU Octave 4.0.1: http://www.gnu.org/software/octave/bug-fixes-4-0-1.html

    Allow strsplit to use '+' as a delimiter
    svds.m: Special handling for case of sigma=0
    Fix splines and remove unnecessary calculations
    Fix printf regression
    Avoid overflow while reshaping large sparse matrices
    Fix memory corruption allocating to sparse arrays
    Disable char/int8_t function overloads where not permitted
    Fix cut/paste error in rande documentation
    Fix regressions caused by ismatrix definition change
    doc: Update copyright date in octave.texi
    Fix regressions caused by ismatrix definition change
    Fix regression for coordinate transforms on 3-D arrays
    Fix regressions caused by ismatrix definition change
    doc: Correct example errobar code and plot
    __getlegenddata__.m: Fix handling of multiple child elements
    fzero.m: Correctly choose tolerance (eps) based on class of fun and X0
    Fix behavior of warning ("error") call
    Fixed isequal for comparison of objects with overridden size
    Fix compilation of classdef with the clang compiler
    doc: Fix incorrect description of Hessenberg decomposition
    set all lexical_feedback variables in init function
    Stop segfault when find function called with 6 outputs
    textread.m, textscan.m: Properly process single-quoted endofline parameter
    Fix rotated graphics when printing through ghostscript
    drawnow: Add a delay to let the GUI thread lock the mutex when printing
    Fix segfault with 'dbclear all'
    Force left-to-right alignment for the whole GUI
    validateattributes.m: Fix typo on documentation
    doc: Fix typo in short-circuit operators documentation
    datevec.m: Properly handle FFF millisecond code
    Fix issue with gnuplot, subplot, and 2 y-axes
    plot.m: Use actual property names in the description FMT argument
    normrnd.m: Return correct result for vector case when std = 0
    prevent doc browser from searching for an empty string
    improve focus detection of gui
    Fix segfault when complex double matrix calls ZGETRF
    Add uiXXXX documentation
    quantile.m: Fix operation along a singleton dimension
    __gnuplot_drawnow__.m: Use "screenpixelsperinch" instead of constant value
    methods.m: Rry Java class names if getMethods fails on Java objects
    textread.m, textscan.m: Always remove headerlines args before invoking strread
    Include stdint.h in mex.h file
    Handle hggroup objects "buttondownfcn" when children are clicked
    Fix selection of one pixel high(wide) images
    annotation.m: Document the "string" property of textbox and textarrow annotations
    mkoctfile: Apply default C++ compiler flags when linking oct-file
    Use backslash as windows file separator for canonicalize_file_name
    Generate correct ezplot for 2-input functions
    doc: Make excplicit that asctime and ctime values end in a newline
    Initialize interpreter correctly with script and --traditional
    do not leave debug modus when entering a command at debug prompt
    __finish__.m: Change from function to script to stay in base workspace
    Don't put default xtick vector size in documentation
    Fix 'descending' sort of Range objects
    strread.m: Clarify delimiter and whitespace usage
    strread.m: Properly preprocess all forms of string format specifier
    strfunc: Document function does not accept anonymous functions
    Fix hang when using errorbar with empty dataset
    doc: Clarify behavior of length function
    Allow sorting columns in workspace view
    Pass all fields in face/vertex structure through to patch()
    polar.m: Display plot titles on polar plots
    ellipke.m: Use correct definition of elliptic integral in documentation
    unpack.m: Stop hang when unpacking on to existing filename
    Allow NEWS to be displayed properly in QTextBrowser
    Fix crash when selecting an image
    qt toolkit: Set keyboard focus in canvas
    Change mxCreateNumericArray to be Matlab compatible for ndims < 2
    Fix conversion from string cell array into java's String
    Fix load/save of integers with -hdf5
    ellipke.m: Modify to accept row vectors
    Return correct hex value for printf when used with string inputs
    run.m: Return to original directory on Windows
    Convert double Octave vectors to double Java vectors
    Also load settings and translations in --no-gui mode
    Avoid crash when trying to annotate an empty figure
    Allow copying a figure to clipboard even in --no-gui mode
    Allow saving/copying other figures than the currentfigure
    linsolve.m: Fix regression when calling linsolve with 2 arguments
    fftshift.m, ifftshift.m: Restore support for N-dimensional arrays
------------------------------------------------------------------------

Sorry I can't work on this merge because I don't have any packaging skills.
--------------------------------------------------------------------------

Ubuntu delta:
  * No-change rebuild for glpk soname change.
  * No-change rebuild for qhull transition.
  * No-change rebuild against suitesparse 1:4.4.5-2
  * No-change rebuild against new graphicsmagick
  * No-change rebuild against libgraphicsmagick-q16-3
  * No-change rebuild for hdf5 soname change.
  * No-change rebuild against libgraphicsmagick++11
  * No-change rebuild against libgraphicsmagick++11
  * Adjust hard-coded Breaks on octave modules to match Ubuntu, rather than
    Debian, rebuild versions.
  * Cherry-pick patch from upstream to disable the GUI on GLES platforms, where
    it fails to build.
---------------------------------------------------------------------------

Changelog entries since current yakkety version 4.0.0-3ubuntu10:

octave (4.0.1-1) unstable; urgency=medium

  * Imported Upstream version 4.0.1
  * d/p/drop_versioned_binaries.diff: add back hunk for src/main.in.cc.

 -- Sébastien Villemot <email address hidden> Tue, 22 Mar 2016 22:10:04 +0100

octave (4.0.1~rc4-1) experimental; urgency=low

  * Imported Upstream version 4.0.1~rc4
    + This version fixes the interaction with gnuplot v5 (Closes: #805295).
    + Also, it also fixes a HDF5 regression (Closes: #815597).
  * d/p/drop_versioned_binaries.diff: Refresh for new upstream version
  * d/control:
    + Use secure URIs in the Vcs-* fields
    + Bump Standards-Version to 3.9.7
  * Install the PDF files of the octave-doc package in /usr/share/doc/octave.
    This is now recommended by the Debian Policy, version 3.9.7 (section
    12.3).
  * d/rules: Avoid triggering the build of oct-gperf.h
  * d/octave.lintian-overrides: Add override for shlib-calls-exit
  * d/p/octave-pkgbuilddir.patch: New patch.
    This patch replaces the patch enable_dirs_in_pkg. It introduces a
    better change in pkg/private/install.m that allows packages to be
    built in place and avoid unreproducible builds due to temporary build
    directories appearing in the .oct files.
  * d/octave-doc.install: Drop useless call to dh-exec

 -- Rafael Laboissiere <email address hidden> Mon, 14 Mar 2016 16:01:37 +0100

octave (4.0.1~rc2-1) experimental; urgency=low

  [ Rafael Laboissiere ]
  * Imported Upstream version 4.0.1~rc2
  * d/octave.menu: Drop menu file per the tech-ctte decision on Bug#741573
  * d/control: Drop useless build-dependency on gperf
  * d/s/lintian-overrides:
    + Add override for source-contains-autogenerated-gperf-data
    + Add override for unknown variable in d/octave-common.install
    + Add override for source-contains-prebuilt-java-object
    + Add override for dh-exec-useless-usage in d/octave.dirs
  * Add Lintian overrides for empty directories
  * d/liboctave3v5.lintian-overrides: Remove obsolete Lintian override.
    Lintian identifies now the "v5" suffix related to the GCC-5 /
    libstdc++6 transition.
  * d/octave-info.lintian-overrides: Remove obsolete Lintian override.
    The info files are shipped again in compressed form since version
    4.0.0-rc3-1 of the package.
  * d/liboctave3v5.lintian-overrides: Add file
  * d/rules: Avoid useless call of ldconfig in the triggers script
  * d/octave-htmldoc.menu: Remove obsolete file

  [ Sébastien Villemot ]
  * Drop octave-dbg package, now rely on automatic debug packages.

 -- Sébastien Villemot <email address hidden> Tue, 19 Jan 2016 17:27:49 +0100

octave (4.0.1~rc1-1) experimental; urgency=low

  [ Rafael Laboissiere ]
  * Imported Upstream version 4.0.1~rc1
  * d/p/xbackslash-texinfo.patch: Drop patch (applied upstream)
  * d/p/dont-use-included-texinfo-texmfcnf.patch: Drop patch (applied upstream)

 -- Sébastien Villemot <email address hidden> Wed, 23 Dec 2015 10:43:56 +0100

octave (4.0.0-5) unstable; urgency=medium

  * d/control: replace libhdf5-mpich2-dev by libhdf5-mpich-dev.

 -- Sébastien Villemot <email address hidden> Thu, 01 Oct 2015 10:05:15 +0200

octave (4.0.0-4) unstable; urgency=medium

  [ Sébastien Villemot ]
  * Rename liboctave3 to liboctave3v5 for g++-5 transition. (Closes: #797992)
  * d/control: promote octave-info to Recommends of octave.
    It is needed by the GUI help browser. (Closes: #793524)

  [ Mike Miller ]
  * Downgrade default-jre-headless from Depends to Recommends.
    Thanks to Alois Schloegl for the suggestion (Closes: #742503)
  * Disable java on hppa, m68k, and sparc64 ports. Fixes FTBFS on those archs.
    (Closes: #792869)

 -- Sébastien Villemot <email address hidden> Sat, 05 Sep 2015 00:03:52 +0200

Revision history for this message
Launchpad Janitor (janitor) wrote :

Status changed to 'Confirmed' because the bug affects multiple users.

Changed in octave (Ubuntu):
status: New → Confirmed
description: updated
summary: - Please merge bugfix release octave 4.0.1-1 (universe) from Debian
+ [FFe] Please merge bugfix release octave 4.0.1-1 (universe) from Debian
testing (main)
summary: - [FFe] Please merge bugfix release octave 4.0.1-1 (universe) from Debian
+ [FFE] Please merge bugfix release octave 4.0.1-1 (universe) from Debian
testing (main)
Changed in octave (Ubuntu):
status: Confirmed → New
summary: - [FFE] Please merge bugfix release octave 4.0.1-1 (universe) from Debian
+ [FFe] Please merge bugfix release octave 4.0.1-1 (universe) from Debian
testing (main)
summary: - [FFe] Please merge bugfix release octave 4.0.1-1 (universe) from Debian
+ Please merge bugfix release octave 4.0.1-1 (universe) from Debian
testing (main)
Revision history for this message
Daniel Holbach (dholbach) wrote :

I will unsubscribe the sponsors team, as there is no patch to be reviewed.

Revision history for this message
Amr Ibrahim (amribrahim1987) wrote :

Could someone work on this merge? I don't have any packaging skills to do it myself.

Revision history for this message
Steve Langasek (vorlon) wrote :

Subscribing the entire development team to the bug is not an appropriate way to get this looked at, sorry. Unsubscribing ubuntu-release and ubuntu-devel.

Revision history for this message
Amr Ibrahim (amribrahim1987) wrote :

Ping!

Changed in octave (Ubuntu):
status: New → Confirmed
Revision history for this message
Amr Ibrahim (amribrahim1987) wrote :

Could someone work on this? It also fixes LP: #1568647.

summary: - Please merge bugfix release octave 4.0.1-1 (universe) from Debian
+ Please merge bugfix release octave 4.0.2-1 (universe) from Debian
testing (main)
description: updated
summary: - Please merge bugfix release octave 4.0.2-1 (universe) from Debian
- testing (main)
+ Merge bugfix release octave 4.0.2-1 (universe) from Debian testing
+ (main)
Mike Miller (mtmiller)
description: updated
summary: - Merge bugfix release octave 4.0.2-1 (universe) from Debian testing
- (main)
+ Merge octave 4.0.2-1 (universe) from Debian testing (main)
Revision history for this message
Gianfranco Costamagna (costamagnagianfranco) wrote :

I merged and uploaded it in yakkety!

Changed in octave (Ubuntu):
status: Confirmed → Fix Released
Revision history for this message
Amr Ibrahim (amribrahim1987) wrote :

Thanks a lot, LocutusOfBorg.

@Mike Miller, could you update the Octave PPA with 4.0.2 for Trusty and Xenial?

Revision history for this message
Mike Miller (mtmiller) wrote :

Yes, and Precise also. After I figure out why I can't `bzr branch ubuntu:octave`.

Gianfranco, I'm interested in your comment on bug #1583128 that 4.0.2 introduces a library change so it's not eligible for an SRU. As an upstream Octave developer, I can say that our intentions are for patch-level releases to be compatible for this very reason, but something may have slipped through. Do you have a library symbol diff that shows why this is the case? You can mail me privately.

Revision history for this message
Amr Ibrahim (amribrahim1987) wrote :

I would love to see 4.0.2 SRUed to Xenial if possible.

Revision history for this message
Amr Ibrahim (amribrahim1987) wrote :

As I filed this bug originally for Xenial.

Revision history for this message
Gianfranco Costamagna (costamagnagianfranco) wrote :

@Amr, this needs a separate tracking on #1583128,
this bug was about merging, and the merge has been done (it is ongoing) for yakkety.
feel free to convert it in an SRU bug if you want to try to get the new release into xenial or wherever.
BTW "diff from 4.0.0-3ubuntu10 to 4.0.2-1ubuntu1 (3.5 MiB)", seems a lot, but I didn't check the diff, just the size

@Mike Miller, unfortunately the fault isn't on your side, the package is exposing a library, that uses/used c++ string variables (exposing them to the packages depending on the library).
That c++ breakage came with libstdc++5 transition, and a lot of packages were renamed in libfooSONAMEv5, breaking libfooSONAME and conflicting with it (because the library file is the same, so they can't coexist.

if you want to see the rationale, feel free to reply/look at the Debian bug report
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=797992
(message #22 shows a clear issue on the package).

So, at the end, if somebody reverts the gcc rename for xenial, and tries to see if the reverse dependencies breaks or not... an SRU might be possible (specially because for xenial the package didn't get the transition done, and this is probably a bug in the current status quo)

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.