Bug#1064982: gnuplot-qt: gnuplot displays a window with nothing in it

2024-02-29 Thread Vincent Lefevre 
Another thing: This problem actually appears even without FVWM's
ManualPlacement, i.e. if the window is displayed immediately. So
I suppose that it is likely to occur with other window managers.

A simple command to reproduce the issue:

  echo 'set terminal wxt; plot x' | gnuplot -persist

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1064982: gnuplot-qt: gnuplot displays a window with nothing in it

2024-02-29 Thread Vincent Lefevre 
On 2024-02-29 12:34:17 +0100, Vincent Lefevre wrote:
> Indeed, this was caused by the upgrade of the pango library. I've
> just posted on the upstream bug:
> 
> Actually, the problem also occurs with gnuplot 5. It has actually
> appeared with the 1.52 pango library: downgrading the pango packages
> to 1.51 makes the problem disappear, both with gnuplot 5 and 6. Now,
> I don't know whether this is a bug in pango or a bug in gnuplot that
> is triggered only with the new version of pango (for instance, it
> might be possible that pango 1.52 is faster than 1.51, making the
> problem appear, since it seems to be a race condition).

I've identified the "problematic" commit in Pango, and I would tend
to say that this is the reason:

https://gitlab.gnome.org/GNOME/pango/-/commit/89442dae443eba2aa0f0a526b4d6d39c0c9b13c6

According to the commit message, a new thread was created "for every
single fontconfig call". Now "a single fontconfig thread per fontmap"
is used. So I suppose that this makes Pango faster, triggering the
race condition in gnuplot.

I've updated the upstream gnuplot bug and opened a bug in Pango,
hoping confirmation:

https://gitlab.gnome.org/GNOME/pango/-/issues/784

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1064982: gnuplot-qt: gnuplot displays a window with nothing in it

2024-02-29 Thread Vincent Lefevre 
Control: found -1 5.4.4+dfsg1-2

On 2024-02-29 03:43:33 +0100, Vincent Lefevre wrote:
> Now, I can reproduce the bug on my other machine. I upgraded some
> packages, and I don't know whether this is related (only pango1.0
> packages seem related).

Indeed, this was caused by the upgrade of the pango library. I've
just posted on the upstream bug:

Actually, the problem also occurs with gnuplot 5. It has actually
appeared with the 1.52 pango library: downgrading the pango packages
to 1.51 makes the problem disappear, both with gnuplot 5 and 6. Now,
I don't know whether this is a bug in pango or a bug in gnuplot that
is triggered only with the new version of pango (for instance, it
might be possible that pango 1.52 is faster than 1.51, making the
problem appear, since it seems to be a race condition).

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1064982: gnuplot-qt: gnuplot displays a window with nothing in it

2024-02-28 Thread Vincent Lefevre 
Control: forwarded -1 https://sourceforge.net/p/gnuplot/bugs/2693/

I've reported the bug upstream.

On 2024-02-29 04:07:17 +0100, Vincent Lefevre wrote:
> For instance, if I run
> 
>   { echo 'plot x' ; sleep 3 ; } | gnuplot -persist
> 
> there are no issues if I place the window before the 3 seconds,
> but the problem appears if I wait for more than 3 seconds.

But note that ssh and strace make the problem disappear
(even if I wait for more than 3 seconds).

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1064982: gnuplot-qt: gnuplot displays a window with nothing in it

2024-02-28 Thread Vincent Lefevre 
Control: retitle -1 with GNUTERM=wxt, "gnuplot -persist" hangs if gnuplot 
cannot draw before stdin is closed, e.g. with FVWM's ManualPlacement

On 2024-02-29 03:47:41 +0100, Vincent Lefevre wrote:
> On 2024-02-29 03:43:33 +0100, Vincent Lefevre wrote:
> > This can be reproduced with the following script:
> > 
> > #!/bin/sh
> > 
> > gnuplot -persist < > plot '-' using 1:2 t '' with line
> > 0 0
> > 10 10
> > e
> > EOF
> > 
> > No issues with GNUTERM=qt, but with GNUTERM=wxt, it doesn't work.
> 
> Or just:
> 
> $ echo 'plot x' | gnuplot -persist

This may be related to my window manager FVWM (and its configuration),
where I choose the window position before it appears:

  ManualPlacement (aka active placement).  The user is required to
  place every new window manually.  The window  only  shows  as  a
  rubber  band  until a place is selected manually.  The window is
  placed when a mouse button or any key except Escape is  pressed.
  Escape  aborts  manual  placement which places the window in the
  top left corner of the screen.  If mouse  button  2  is  pressed
  during the initial placement of a window (respectively Shift and
  mouse  button  1 in case Mwm emulation has been enabled with the
  Emulate command), the user is asked to resize the window too.

For instance, if I run

  { echo 'plot x' ; sleep 3 ; } | gnuplot -persist

there are no issues if I place the window before the 3 seconds,
but the problem appears if I wait for more than 3 seconds.

The problem seems to be due to the fact that the standard input gets
closed before gnuplot can actually draw. Indeed,

  { echo 'plot x' ; exec >&- ; sleep 3 ; } | gnuplot -persist

triggers the problem even if I place the window before the 3 seconds.

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1064982: gnuplot-qt: gnuplot displays a window with nothing in it

2024-02-28 Thread Vincent Lefevre 
On 2024-02-29 03:43:33 +0100, Vincent Lefevre wrote:
> This can be reproduced with the following script:
> 
> #!/bin/sh
> 
> gnuplot -persist < plot '-' using 1:2 t '' with line
> 0 0
> 10 10
> e
> EOF
> 
> No issues with GNUTERM=qt, but with GNUTERM=wxt, it doesn't work.

Or just:

$ echo 'plot x' | gnuplot -persist

> I suspect a race condition, as using strace makes the problem
> disappear. Using ssh (even "ssh localhost", then the test script)
> also makes the problem disappear.

Ditto with the above command.

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1064982: gnuplot-qt: gnuplot displays a window with nothing in it

2024-02-28 Thread Vincent Lefevre 
On 2024-02-28 15:01:17 -0800, Dima Kogan wrote:
> Can you see if other wxt applications work on a system that's
> exhibiting this problem?

The issue seems specific to gnuplot under particular conditions.
See below.

Now, I can reproduce the bug on my other machine. I upgraded some
packages, and I don't know whether this is related (only pango1.0
packages seem related).

This can be reproduced with the following script:

#!/bin/sh

gnuplot -persist < - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1064982: gnuplot-qt: gnuplot displays a window with nothing in it

2024-02-28 Thread Vincent Lefevre 
On 2024-02-28 10:39:49 -0800, Dima Kogan wrote:
> Hi. I'd like to get more clarity.
> 
> - You see the issue when you try to plot anything at all?
> 
> - You say "plot x" and you get a plot window, but it's all white, or
>   something?
> 
> - Only with the "qt" terminal?

It was the wxt terminal (which I normally use all the time).

I'll do more tests tomorrow since I can't reproduce the problem
remotely (or locally on a different machine).

I could also look at the strace and xtrace output.

When I got the problem, using FVWM's Delete command to remove
the window did not have any effect (so this is not just a
display issue, the process was frozen in some state); FVWM's
Destroy command worked. These commands are described as follows:

Delete
Sends a message to a window asking that it remove itself,
frequently causing the application to exit.

Destroy
Destroys an application window, which usually causes the
application to crash and burn.

The problem occurred first at 2024-02-28T19:12:24 CET.
At that time, journalctl shows:

Feb 28 19:12:24 cventin dbus-daemon[1489]: [session uid=1000 pid=1489] 
Activating via systemd: service name='org.freedesktop.portal.Desktop' 
unit='xdg-desktop-portal.service' requested by ':1.27' (uid=1000 pid=7164 
comm="/usr/bin/gnuplot -persist")
Feb 28 19:12:25 cventin systemd[1462]: Starting xdg-desktop-portal.service - 
Portal service...
Feb 28 19:12:25 cventin dbus-daemon[1489]: [session uid=1000 pid=1489] 
Activating via systemd: service name='org.freedesktop.portal.Documents' 
unit='xdg-document-portal.service' requested by ':1.28' (uid=1000 pid=7168 
comm="/usr/libexec/xdg-desktop-portal")
Feb 28 19:12:25 cventin systemd[1462]: Starting xdg-document-portal.service - 
flatpak document portal service...
Feb 28 19:12:25 cventin dbus-daemon[1489]: [session uid=1000 pid=1489] 
Activating via systemd: service 
name='org.freedesktop.impl.portal.PermissionStore' 
unit='xdg-permission-store.service' requested by ':1.29' (uid=1000 pid=7174 
comm="/usr/libexec/xdg-document-portal")
Feb 28 19:12:25 cventin systemd[1462]: Starting xdg-permission-store.service - 
sandboxed app permission store...
Feb 28 19:12:25 cventin dbus-daemon[1489]: [session uid=1000 pid=1489] 
Successfully activated service 'org.freedesktop.impl.portal.PermissionStore'
Feb 28 19:12:25 cventin systemd[1462]: Started xdg-permission-store.service - 
sandboxed app permission store.
Feb 28 19:12:25 cventin dbus-daemon[1489]: [session uid=1000 pid=1489] 
Successfully activated service 'org.freedesktop.portal.Documents'
Feb 28 19:12:25 cventin systemd[1462]: Started xdg-document-portal.service - 
flatpak document portal service.
Feb 28 19:12:25 cventin xdg-document-portal[7174]: Ignoring invalid max threads 
value 4294967295 > max (10).
Feb 28 19:12:25 cventin xdg-desktop-por[7168]: Choosing gtk.portal for 
org.freedesktop.impl.portal.Lockdown as a last-resort fallback
Feb 28 19:12:25 cventin xdg-desktop-por[7168]: The preferred method to match 
portal implementations to desktop environments is to use the portals.conf(5) 
configuration file
Feb 28 19:12:25 cventin dbus-daemon[1489]: [session uid=1000 pid=1489] 
Activating via systemd: service name='org.freedesktop.impl.portal.desktop.gtk' 
unit='xdg-desktop-portal-gtk.service' requested by ':1.28' (uid=1000 pid=7168 
comm="/usr/libexec/xdg-desktop-portal")
Feb 28 19:12:25 cventin systemd[1462]: Starting xdg-desktop-portal-gtk.service 
- Portal service (GTK/GNOME implementation)...
Feb 28 19:12:25 cventin dbus-daemon[1489]: [session uid=1000 pid=1489] 
Activating via systemd: service name='org.a11y.Bus' 
unit='at-spi-dbus-bus.service' requested by ':1.31' (uid=1000 pid=7189 
comm="/usr/libexec/xdg-desktop-portal-gtk")
Feb 28 19:12:25 cventin systemd[1462]: Starting at-spi-dbus-bus.service - 
Accessibility services bus...
Feb 28 19:12:25 cventin dbus-daemon[1489]: [session uid=1000 pid=1489] 
Successfully activated service 'org.a11y.Bus'
Feb 28 19:12:25 cventin systemd[1462]: Started at-spi-dbus-bus.service - 
Accessibility services bus.
Feb 28 19:12:25 cventin at-spi-bus-launcher[7197]: dbus-daemon[7197]: 
Activating service name='org.a11y.atspi.Registry' requested by ':1.0' (uid=1000 
pid=7189 comm="/usr/libexec/xdg-desktop-portal-gtk")
Feb 28 19:12:25 cventin at-spi-bus-launcher[7197]: dbus-daemon[7197]: 
Successfully activated service 'org.a11y.atspi.Registry'
Feb 28 19:12:25 cventin at-spi-bus-launcher[7203]: SpiRegistry daemon is 
running with well-known name - org.a11y.atspi.Registry
Feb 28 19:12:25 cventin dbus-daemon[1489]: [session uid=1000 pid=1489] 
Successfully activated service 'org.freedesktop.impl.portal.desktop.gtk'
Feb 28 19:12:25 cventin systemd[1462]: Started xdg-desktop-portal-gtk.service - 
Portal service (GTK/GNOME implementation).
Feb 28 19:12:25 cventin rtkit-daemon[811]: Supervising 3 threads of 1 processes 
of 1 users.
Feb 28 19:12:25 cventin rtkit-daemon[811]: Supervising 3 threads of 1 processes 
of 1

Bug#1064982: gnuplot-qt: gnuplot displays a window with nothing in it

2024-02-28 Thread Vincent Lefevre 
On 2024-02-28 19:20:58 +0100, Vincent Lefevre wrote:
> Before the upgrade to gnuplot 6, everything was fine. But after
> the upgrade, I get a window where nothing is drawn, i.e. I just
> get what's *behind* the window. That's always reproducible on
> this machine.
> 
> I've tried GNUTERM=wxt only. I'll do more tests tomorrow.

Interesting. This is not reproducible when the display is on
another machine (i.e. I connect via ssh). Perhaps an issue
related to the local X server?

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1064982: gnuplot-qt: gnuplot displays a window with nothing in it

2024-02-28 Thread Vincent Lefevre 
Package: gnuplot-qt
Version: 6.0.0+dfsg1-1
Severity: grave
Justification: renders package unusable

Before the upgrade to gnuplot 6, everything was fine. But after
the upgrade, I get a window where nothing is drawn, i.e. I just
get what's *behind* the window. That's always reproducible on
this machine.

I've tried GNUTERM=wxt only. I'll do more tests tomorrow.

-- System Information:
Debian Release: trixie/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'stable-updates'), (500, 
'stable-security'), (500, 'stable-debug'), (500, 'proposed-updates-debug'), 
(500, 'unstable'), (500, 'testing'), (500, 'stable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 6.6.15-amd64 (SMP w/12 CPU threads; PREEMPT)
Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE, 
TAINT_UNSIGNED_MODULE
Locale: LANG=C.UTF-8, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

Versions of packages gnuplot-qt depends on:
ii  gnuplot-data 6.0.0+dfsg1-1
ii  libc62.37-15
ii  libcairo21.18.0-1+local1
ii  libedit2 3.1-20230828-1
ii  libgcc-s114-20240221-2.1
ii  libgd3   2.3.3-9+b1
ii  libglib2.0-0 2.78.4-1
ii  liblua5.4-0  5.4.6-3
ii  libpango-1.0-0   1.52.0+ds-1
ii  libpangocairo-1.0-0  1.52.0+ds-1
ii  libqt5core5a 5.15.10+dfsg-7
ii  libqt5gui5   5.15.10+dfsg-7
ii  libqt5network5   5.15.10+dfsg-7
ii  libqt5printsupport5  5.15.10+dfsg-7
ii  libqt5svg5   5.15.10-2+b1
ii  libqt5widgets5   5.15.10+dfsg-7
ii  libstdc++6   14-20240221-2.1
ii  libwebp7 1.3.2-0.4
ii  libwebpmux3  1.3.2-0.4
ii  libwxbase3.2-1   3.2.4+dfsg-3
ii  libwxgtk3.2-13.2.4+dfsg-3
ii  libx11-6 2:1.8.7-1

gnuplot-qt recommends no packages.

Versions of packages gnuplot-qt suggests:
ii  gnuplot-doc  6.0.0+dfsg1-1

-- no debconf information

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1057344: libgmp10: major formatted output function bug with %c and the value 0

2023-12-03 Thread Vincent Lefevre 
Package: libgmp10
Version: 2:6.2.1+dfsg1-1.1
Severity: grave
Tags: security upstream
Justification: user security hole
Forwarded: https://gmplib.org/list-archives/gmp-bugs/2023-December/005420.html
X-Debbugs-Cc: Debian Security Team 

I've reported the following bug upstream. Debian/stable is affected
(at least on the testcase below, but the various issues are probably
related).

With GMP 6.3.0, the formatted output functions do not handle %c
with the value 0 correctly. For gmp_sprintf, the return value is
incorrect. For gmp_asprintf and gmp_vasprintf, this is either a
buffer overflow (according to the GMP manual: "The block will be
the size of the string and null-terminator.") or, in case this
is an error in the GMP manual, possible memory corruption when
freeing the allocated memory, if the custom memory allocation
function cares about the size parameter.

Testcase for gmp_sprintf:


#include 
#include 

static void test (int flag)
{
  char s[3] = { 1, 1, 1 };
  int r;

  r = (flag ? sprintf : gmp_sprintf) (s, "%c", 0);
  printf ("%4s: r = %d, s = { %d %d %d }\n",
  flag ? "libc" : "gmp", r, s[0], s[1], s[2]);
}

int main (void)
{
  test (0);
  test (1);
  return 0;
}


which currently gives:

 gmp: r = 0, s = { 0 0 1 }
libc: r = 1, s = { 0 0 1 }

MPFR has various issues concerning %c with the value 0, but an
attempt to fix them fails due to

  length = gmp_vasprintf (...);
[...]
  mpfr_free_str (s);

which is similar to GMP's tests/misc/t-printf.c file, which contains

  got_len = gmp_vasprintf (, fmt, ap);
[...]
  (*__gmp_free_func) (got, strlen(got)+1);

But replacing

  mpfr_free_str (s);

by

  mpfr_free_func (s, length + 1);

i.e. using the return value length instead of strlen(s), also fails.
I suppose that this is related to the incorrect return value.

-- System Information:
Debian Release: 12.2
  APT prefers stable-updates
  APT policy: (500, 'stable-updates'), (500, 'stable-security'), (500, 
'stable-debug'), (500, 'stable')
Architecture: amd64 (x86_64)

Kernel: Linux 6.1.0-13-amd64 (SMP w/1 CPU thread; PREEMPT)
Kernel taint flags: TAINT_WARN
Locale: LANG=POSIX, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

Versions of packages libgmp10 depends on:
ii  libc6  2.36-9+deb12u3

libgmp10 recommends no packages.

libgmp10 suggests no packages.

-- no debconf information

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#884538: gnuplot-qt: corrupted double-linked list

2023-04-25 Thread Vincent Lefevre 
On 2017-12-16 14:45:11 +0100, Vincent Lefevre wrote:
> Package: gnuplot-qt
> Version: 5.2.2+dfsg1-2
> Severity: important
> 
> "/usr/bin/gnuplot -persist" crashed after the following warning:
> 
> (gnuplot:30420): GLib-WARNING **: corrupted double-linked list detected
> 
> I can reproduce neither the crash nor the warning. But since this
> seems to be due to memory corruption, this is not very surprising.
[...]

Perhaps the same issue as bug 841137 (one year before, in 2016).
See also bug 909223 (also a corrupted double-linked list, which
made gnuplot-qt crash) in 2018 and bug 794799 in 2015.

But AFAIK, my latest gnuplot crash was bug 909223 in 2018. All these
crashes occurred on my HP ZBook 15 G2 laptop, which I'm still using,
and I haven't changed my config. So perhaps the issue has been fixed
(in gnuplot or some library).

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1027303: gnuplot-doc: The Info manual contains a spurious Texinfo comment "@c <3 [...]"

2022-12-29 Thread Vincent Lefevre 
Package: gnuplot-doc
Version: 5.4.4+dfsg1-2
Severity: minor
Tags: upstream patch
Forwarded: https://sourceforge.net/p/gnuplot/bugs/2576/

The generated Info manual (gnuplot.info) contains a spurious Texinfo
comment "@c <3 - all terminal stuff is pulled from the .trm files" in
Section 4.1 (complete list of terminals).

What happens is that the docs/doc2texi.el converter transforms the
line starting with "<3" into a Texinfo comment by inserting "@c "
at the beginning of the line (function d2t-get-terminals) with the
intent that this line will not appear in the Info manual (as being
a Texinfo comment). However, the input has not been converted to
the Texinfo format yet, so that function d2t-braces-atsigns
(consisting in the first step of the conversion to Texinfo), which
is called later, assumes that "@c" is real text and adds a "@" in
front of it (to escape the "@" since it is regarded as real text).
As a consequence, one gets a line

  @@c <3 -- all terminal stuff is pulled from the .trm files

in the generated Texinfo file gnuplot.texi, and this line will be
rendered as

  @c <3 -- all terminal stuff is pulled from the .trm files

in the Info manual.

I'm attaching a patch that makes doc2texi.el remove this <3 line
instead of attempting to comment it out.

-- System Information:
Debian Release: bookworm/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'stable-updates'), (500, 
'stable-security'), (500, 'unstable'), (500, 'testing'), (500, 'stable'), (1, 
'experimental')
merged-usr: no
Architecture: amd64 (x86_64)

Kernel: Linux 6.0.0-6-amd64 (SMP w/8 CPU threads; PREEMPT)
Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE, 
TAINT_UNSIGNED_MODULE
Locale: LANG=POSIX, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

-- no debconf information

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)
Description: Remove spurious Texinfo comment from gnuplot.info.
  The docs/doc2texi.el converter transforms the line starting
  with "<3" into a Texinfo comment by inserting "@c " at the
  beginning of the line (function d2t-get-terminals). However,
  the input has not been converted to the Texinfo format yet,
  so that function d2t-braces-atsigns, which is called later,
  assumes that "@c" is real text and adds a @ in front of it.
  As a consequence, one gets a line
@@c <3 -- all terminal stuff is pulled from the .trm files
  in the generated Texinfo file gnuplot.texi, and this line
  will be rendered as
@c <3 -- all terminal stuff is pulled from the .trm files
  in the Info manual.
  This patch makes doc2texi.el remove the "<3" line instead of
  attempting to comment it out.
Author: Vincent Lefevre 
Bug: https://sourceforge.net/p/gnuplot/bugs/2576/
Last-Update: 2022-12-30

Index: gnuplot-5.4.4+dfsg1/docs/doc2texi.el
===
--- gnuplot-5.4.4+dfsg1.orig/docs/doc2texi.el
+++ gnuplot-5.4.4+dfsg1/docs/doc2texi.el
@@ -482,8 +482,8 @@ the end of `d2t-get-terminals'.")
 (save-excursion
   (when (re-search-forward "^<3" (point-max) t)
 	(beginning-of-line)
-	(insert "@c ")
-	(forward-line 1)
+	(let ((eol (save-excursion (end-of-line) (point-marker
+	  (delete-region (point-marker) eol))
 	(dolist (elem list)
   (and d2t-verbose (message "%s ..." elem))
   (setq file (concat d2t-terminal-directory elem ".trm"))
-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#989029: info: scroll-backward is buggy

2022-12-28 Thread Vincent Lefevre 
Control: forwarded -1 https://sourceforge.net/p/gnuplot/bugs/2575/

On 2022-12-29 02:45:35 +0100, Vincent Lefevre wrote:
> The forward was for the missing terminal types. I have not reported
> this second bug upstream yet.

I've now reported the bug upstream.

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#989029: info: scroll-backward is buggy

2022-12-28 Thread Vincent Lefevre 
Control: retitle -1 gnuplot-doc: spurious menu at the end of Section 4 of 
gnuplot.info breaks navigation in info
Control: notforwarded -1

The forward was for the missing terminal types. I have not reported
this second bug upstream yet.

On 2022-12-29 00:51:19 +0100, Vincent Lefevre wrote:
[...]
> This solves the inclusion of the terminal types, but unfortunately,
> the buggy menu is a separate problem. In the gnuplot.info.gz file
> with the patch:
> 
> [...]
> 4.1.54 xlib
> ---
> The 'xlib' terminal driver supports the X11 Windows System.  It
> generates gnuplot_x11 commands, but sends them to the output file
> specified by 'set output '''.  'set term x11' is equivalent to
> 'set output "|gnuplot_x11 -noevents"; set term xlib'.  'xlib' takes the
> same set of options as 'x11'."
> * Menu:
> * Bugs::
> File: gnuplot.info,  Node: Bugs,  Next: Concept_Index,  Prev: Terminal_types, 
>  Up: Top
> 5 Bugs
> **
> [...]
> 
> i.e. the menu is still there, so that going backward from Section 5
> goes to Section 5.2 instead of 4.1.54.

I've attached a patch for this issue: this incorrect menu is added
by 3 lines in doc2texi.el, with the comment

  ; `texinfo-all-menus-update' seems
  ; to miss this one.  how odd.

Perhaps this was added as a workaround to an old bug. With my
patch that removes these 3 lines, the navigation is now correct
(going backward from Section 5 "Bugs" sends to the end of
Section 4 as expected), and I cannot see any new issue.

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)
Description: Do not generate a spurious menu in gnuplot.info.
  A spurious menu with a "Bugs" menu item is generated at the end
  of Section 4 (just before Section 5 "Bugs"). This is incorrect
  and breaks navigation in "info": going backward from Section 5
  sends to Section 5.2 instead of 4.1.54.
Bug-Debian: https://bugs.debian.org/989029
Author: Vincent Lefevre 
Last-Update: 2022-12-29

Index: gnuplot-5.4.4+dfsg1/docs/doc2texi.el
===
--- gnuplot-5.4.4+dfsg1.orig/docs/doc2texi.el
+++ gnuplot-5.4.4+dfsg1/docs/doc2texi.el
@@ -415,9 +415,6 @@ particular conversion chore."
 (search-forward "@node")
 ;; (beginning-of-line)
 ;; (insert "\n\n" d2t-main-menu "\n\n")
-(search-forward "@node Bugs")	; `texinfo-all-menus-update' seems
-(beginning-of-line)			; to miss this one.  how odd.
-(insert "@menu\n* Bugs::\t\t\t\n@end menu\n\n")
 (goto-char (point-max))
 (insert d2t-texi-footer))
   (load-library "texinfo") ;; now do the hard stuff with texinfo-mode
-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#989028: gnuplot.info: missing terminal types

2022-12-28 Thread Vincent Lefevre 
On 2021-05-24 01:21:56 +0200, Vincent Lefevre wrote:
> After "info gnuplot", go to "Terminal_types::". One gets:
> 
> 
> 4 Terminal types
> 
> 
> * Menu:
> 
> * complete_list_of_terminals::
> 
> 
> The next page is just:
> 
> 
> 4.1 complete list of terminals
> ==
> 
> Gnuplot supports a large number of output formats.  These are selected
> by choosing an appropriate terminal type, possibly with additional
> modifying options.  See *note terminal::.
> 
>This document may describe terminal types that are not available to
> you because they were not configured or installed on your system.  To
> see a list of terminals available on a particular gnuplot installation,
> type 'set terminal' with no modifiers.
> 
>Terminals marked 'legacy' are not built by default in recent gnuplot
> versions and may not actually work.  @c <3 - all terminal stuff is
> pulled from the .trm files
> 
> * Menu:
> 
> * Bugs::
> 
> 
> without the complete list of terminals, contrary to the manual
> in HTML.

I'm attaching the patch I had posted to bug 989029. I thought that
these two bugs (missing terminal types and broken navigation) were
due to the same cause, and I merged them, but these are really two
independent bugs, though they occur at the same place in the manual.
So I unmerged them a few hours ago.

This patch fixes the missing terminal types by providing a correct
relative pathname for the .trm files. Note that it is specific to
Debian, since the relative pathname depends on the location of the
build directory (a general fix would be more complex than this
one-line change).

Note: the menu with the "Bugs" item seen above is bug 989029 (this
menu is the cause of the broken navigation in "info", and I will
post a patch there).

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)
Description: Fix missing terminal types in gnuplot.info.
  This patch is specific to Debian, since the d2t-terminal-directory
  relative pathname depends on the location of the build directory
  (a general fix would be more complex).
Bug-Debian: https://bugs.debian.org/989028
Author: Vincent Lefevre 
Last-Update: 2022-11-22

Index: gnuplot-5.4.4+dfsg1/docs/doc2texi.el
===
--- gnuplot-5.4.4+dfsg1.orig/docs/doc2texi.el
+++ gnuplot-5.4.4+dfsg1/docs/doc2texi.el
@@ -116,7 +116,7 @@
 ;;; You may need to customize these variables:
 (defvar d2t-doc-file-name "gnuplot.doc"
   "Name of the gnuplot.doc file.")
-(defvar d2t-terminal-directory (expand-file-name "../term/")
+(defvar d2t-terminal-directory (expand-file-name "../../../term/")
   "Location of .trm files in gnuplot source tree.")
 
 
-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#989029: info: scroll-backward is buggy

2022-12-28 Thread Vincent Lefevre 
Control: unmerge -1
Control: retitle -1 gnuplot-doc: spurious menu at the end of Section 4 breaks 
navigation

On 2022-11-22 15:08:10 +0100, Vincent Lefevre wrote:
> doc2texi.el has
> 
> (defvar d2t-terminal-directory (expand-file-name "../term/")
>   "Location of .trm files in gnuplot source tree.")
> 
> If I understand correctly, this is relative to the build directory
> (since the current working directory is there), while this should have
> been relative to the source directory. A general fix is not obvious:
> one would have to pass $(srcdir) to the file, unless there is some
> Emacs variable that already contains the information. However, for
> Debian, the build directory is at some fixed place in the source tree,
> so that this can be fixed by just adapting the relative path.
> 
> Patch attached.

This solves the inclusion of the terminal types, but unfortunately,
the buggy menu is a separate problem. In the gnuplot.info.gz file
with the patch:

[...]
4.1.54 xlib
---
The 'xlib' terminal driver supports the X11 Windows System.  It
generates gnuplot_x11 commands, but sends them to the output file
specified by 'set output '''.  'set term x11' is equivalent to
'set output "|gnuplot_x11 -noevents"; set term xlib'.  'xlib' takes the
same set of options as 'x11'."
* Menu:
* Bugs::
File: gnuplot.info,  Node: Bugs,  Next: Concept_Index,  Prev: Terminal_types,  
Up: Top
5 Bugs
**
[...]

i.e. the menu is still there, so that going backward from Section 5
goes to Section 5.2 instead of 4.1.54.

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#989029: info: scroll-backward is buggy

2022-11-22 Thread Vincent Lefevre 
Control: forwarded -1 https://sourceforge.net/p/gnuplot/bugs/2568/

I've just reported a bug upstream ("with a separate build directory,
the info manual is not built correctly").

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#989029: info: scroll-backward is buggy

2022-11-22 Thread Vincent Lefevre 
Control: tags -1 upstream patch

On 2022-11-22 14:46:28 +0100, Vincent Lefevre wrote:
> Removed the upstream tag as I can't reproduce the issue by compiling
> upstream's gnuplot 5.4.5, and Makefile.am and doc2texi.el in docs have
> not changed (and the 5.4.5 release notes don't mention any fix about
> the info manual).

Actually, it *is* an upstream bug. But I can reproduce it only with
a separate build directory (this is what Debian does). I suppose
that the path to the .trm files is incorrect in such a case.

doc2texi.el has

(defvar d2t-terminal-directory (expand-file-name "../term/")
  "Location of .trm files in gnuplot source tree.")

If I understand correctly, this is relative to the build directory
(since the current working directory is there), while this should have
been relative to the source directory. A general fix is not obvious:
one would have to pass $(srcdir) to the file, unless there is some
Emacs variable that already contains the information. However, for
Debian, the build directory is at some fixed place in the source tree,
so that this can be fixed by just adapting the relative path.

Patch attached.

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)
Index: gnuplot-5.4.4+dfsg1/docs/doc2texi.el
===
--- gnuplot-5.4.4+dfsg1.orig/docs/doc2texi.el
+++ gnuplot-5.4.4+dfsg1/docs/doc2texi.el
@@ -116,7 +116,7 @@
 ;;; You may need to customize these variables:
 (defvar d2t-doc-file-name "gnuplot.doc"
   "Name of the gnuplot.doc file.")
-(defvar d2t-terminal-directory (expand-file-name "../term/")
+(defvar d2t-terminal-directory (expand-file-name "../../../term/")
   "Location of .trm files in gnuplot source tree.")
 
 
-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#989029: info: scroll-backward is buggy

2022-11-22 Thread Vincent Lefevre 
Control: tags -1 - upstream

Removed the upstream tag as I can't reproduce the issue by compiling
upstream's gnuplot 5.4.5, and Makefile.am and doc2texi.el in docs have
not changed (and the 5.4.5 release notes don't mention any fix about
the info manual).

In the gnuplot.info generated using upstream's tarball:

[...]
   Terminals marked 'legacy' are not built by default in recent gnuplot
versions and may not actually work.  @c <3 - all terminal stuff is
pulled from the .trm files

* Menu:

* aifm::
* aqua::
* be::
* epscairo::
[...]

In the gnuplot.info generated using the Debian source package:

[...]
   Terminals marked 'legacy' are not built by default in recent gnuplot
versions and may not actually work.  @c <3 - all terminal stuff is
pulled from the .trm files

* Menu:

* Bugs::

[...]

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1024630: gnuplot should not build-depend on emacs-nox only

2022-11-22 Thread Vincent Lefevre 
Source: gnuplot
Version: 5.4.4+dfsg1-2
Severity: normal

The debian/control file has

Build-Depends-Indep: latex2html,
[...]
 emacs-nox

However, the emacs variants conflicts with each other, so that
if one uses a different emacs variant, one cannot build gnuplot.

Since any emacs variant can be used, it would be better to just
build-depend on "emacs-nox | emacs".

-- System Information:
Debian Release: bookworm/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'stable-updates'), (500, 
'stable-security'), (500, 'unstable'), (500, 'testing'), (500, 'stable'), (1, 
'experimental')
merged-usr: no
Architecture: amd64 (x86_64)

Kernel: Linux 6.0.0-4-amd64 (SMP w/8 CPU threads; PREEMPT)
Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE, 
TAINT_UNSIGNED_MODULE
Locale: LANG=POSIX, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#989029: info: scroll-backward is buggy

2022-11-22 Thread Vincent Lefevre 
Control: found -1 5.4.1+dfsg1-1
Control: found -1 5.4.4+dfsg1-2
Control: forcemerge -1 989028

On 2022-11-22 13:12:00 +0100, Vincent Lefevre wrote:
> The main issue is actually that the list of terminals is absent
> from the info manual (the HTML and PDF manuals seem OK), which
> makes the info manual *very* incomplete (this part is about 20%
> of the PDF manual) as there are many options specific to
> terminals, which are thus not documented; and in
[...]

and this was bug 989028, which I had reported just before this one.

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#989029: info: scroll-backward is buggy

2022-11-22 Thread Vincent Lefevre 
Control: retitle -1 gnuplot-doc: the list of terminals is absent from the info 
manual, also breaking navigation
Control: severity -1 important
Control: tags -1 - wontfix

The wontfix tag had to be removed...

The main issue is actually that the list of terminals is absent
from the info manual (the HTML and PDF manuals seem OK), which
makes the info manual *very* incomplete (this part is about 20%
of the PDF manual) as there are many options specific to
terminals, which are thus not documented; and in

  4.1 complete list of terminals

this list has been replaced by "Bugs::" in the menu, which also
breaks navigation, as explained below.

Moreover, due to the absent list of terminals, the terminal index
is obviously empty.

I suspect a bug in doc2texi.el, which generates the gnuplot.texi
file, from which the info manual is built by texinfo. More
precisely, the inclusion of the .trm files is buggy.

On 2022-11-22 10:34:26 +0100, Hilmar Preuße wrote:
> Control: reassign -1 gnuplot-doc
> 
> Am 24.05.2021 um 01:36 teilte Vincent Lefevre mit:
> 
> Hello,
> 
> as discussed in that bug the issue seems only to occur w/ gnuplot-doc so
> the issue is specific to the gnuplot documentation. Please be so kind to
> have a look at the issue.
> 
> Hilmar
> 
> > The info manual says:
> > 
> >  ('scroll-backward')
> > 
> >   Shift the text in this window down.  The inverse of
> >   'scroll-forward'.  If you are at the start of a node,  takes
> >   you to the "previous" node, so that you can read an entire manual
> >   from finish to start by repeating .  The default scroll size
> >   can be changed by invoking the
> >   ('scroll-backward-page-only-set-window') command with a numeric
> >   argument.
> > [...]
> > 
> > But this doesn't behave like that in the gnuplot manual from the
> > gnuplot-doc 5.4.1+dfsg1-1 Debian package: Go to "Bugs". This gives
> > 
> > 
> > 5 Bugs
> > **
> > 
> > Please e-mail bug reports to the gnuplot-bugs mailing list or upload the
> > report to the gnuplot web site on SourceForge.  Please give complete
> > information on the version of gnuplot you are using and, if possible, a
> > test script that demonstrates the bug.  See 'seeking-assistance'.
> > 
> > * Menu:
> > 
> > * known_limitations::
> > * External_libraries::
> > 
> > 
> > Then do a scroll-backward (with the Backspace or PageUp key).
> > This gives:
> > 
> > 
> > 5.2 External libraries
> > ==
> > [...]
> > 
> > 
> > which is not the "previous" node, but some later node.

The issue comes from the fact that

  4.1 complete list of terminals

has "Bugs::" in the menu, while the "Bugs" node is not a child of
this node ("Bugs" is Section 5, not Section 4.1.x). This is not
supported by "info".

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1019061: gnuplot-data not available

2022-09-03 Thread Vincent Lefevre 
Control: retitle -1 gnuplot and gnuplot-data not available
Control: reassign -1 src:gnuplot 5.4.4+dfsg1-1

This is actually a problem for the binary packages gnuplot-qt,
gnuplot-x11 and gnuplot-nox. The gnuplot metapackage is not
available either, while it is still mentioned at

  https://packages.debian.org/source/unstable/gnuplot

(but this one would not alone prevent an upgrade).

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#1019061: gnuplot-data not available

2022-09-03 Thread Vincent Lefevre 
Control: severity -1 serious

Hi,

On 2022-09-03 15:51:49 +0200, alberto wrote:
> Dear Maintainer,
> 
> it looks like gnuplot-data is not available anymore (at least from version
> 5.4.4-gfsg1-1). This makes gnuplot (any flavour) uninstallable.
> 
> I'm surprised I could not find any report on this on the debian bugtracking
> system. Is there anything I'm missing about this package?

I've noticed the same issue for several days. I thought that this
was just temporary, but apparently, after 14 days, it isn't.

> In any case, if I'm not wrong, this "bug" (maybe just a delay in packaging
> the new version?) is far more than "important" because it makes the package
> unusable.

I agree. It is not installable. So this should be a RC bug.
I'm increasing the severity.

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#994405: libgmp10:i386: buffer overflow due to integer overflow in mpz/inp_raw.c on 32-bit machines

2021-11-14 Thread Vincent Lefevre 
On 2021-11-14 14:15:25 +0100, Anton Gladky wrote:
> well, I was thinking that upstream should request a CVE. Neverheless
> I could not reproduce the issue with the modern GCC-versions.
> Even on 32bit-systems.

I can still reproduce the segmentation fault under Debian/unstable.
Simplified testcase:

#include 
#include 

int main (void)
{
  mpz_t s;
  mpz_init (s);
  mpz_inp_raw (s, stdin);
  return 0;
}

Compile with gcc -m32 and execute:

  printf 12345 | ./testcase

Note that even if you don't get a segmentation fault, there may be
other erratic behaviors, such as silent memory corruption (which may
be even worse).

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#994405: libgmp10:i386: buffer overflow due to integer overflow in mpz/inp_raw.c on 32-bit machines

2021-09-16 Thread Vincent Lefevre 
On 2021-09-16 21:23:34 +0200, Anton Gladky wrote:
> Thanks for the bug report. We will fix it when CVE (if any) will be
> assigned and upstream patch will be available.

FYI, an upstream patch is now available here:

  https://gmplib.org/list-archives/gmp-bugs/2021-September/005087.html

> Though, the integer overflows are not making the package unusable in
> most cases.

Yes, but they may introduce security issues, in particular here
because the behavior depends on data from a file, which may be
untrusted. That said, here it is probably wise to check that the
size is not too large in order to prevent the address space from
being exhausted.

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#994405: libgmp10:i386: buffer overflow due to integer overflow in mpz/inp_raw.c on 32-bit machines

2021-09-15 Thread Vincent Lefevre 
Package: libgmp10
Version: 2:6.2.1+dfsg-2
Severity: grave
Tags: security upstream
Justification: user security hole
X-Debbugs-Cc: Debian Security Team 

mpz_inp_raw segfaults (SEGV_MAPERR) on large sizes. I suspect that
this is due to an integer overflow in mpz/inp_raw.c:

  abs_xsize = BITS_TO_LIMBS (abs_csize*8);

See discussion
  https://gmplib.org/list-archives/gmp-bugs/2021-September/005077.html

and my comment
  https://gmplib.org/list-archives/gmp-bugs/2021-September/005086.html

I have not checked, but abs_xsize would be smaller than expected,
thus

  xp = MPZ_NEWALLOC (x, abs_xsize);

would allocate less than expected, thus I suppose that

  cp = (char *) (xp + abs_xsize) - abs_csize;

points to a location that is *before* the buffer.

-- System Information:
Debian Release: bookworm/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'stable-security'), (500, 
'unstable'), (500, 'testing'), (500, 'stable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 5.10.0-8-amd64 (SMP w/12 CPU threads)
Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE, 
TAINT_UNSIGNED_MODULE
Locale: LANG=POSIX, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

Versions of packages libgmp10:i386 depends on:
ii  libc6  2.32-2

libgmp10:i386 recommends no packages.

libgmp10:i386 suggests no packages.

-- no debconf information

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#989028: gnuplot.info: missing terminal types

2021-05-23 Thread Vincent Lefevre 
Package: gnuplot-doc
Version: 5.4.1+dfsg1-1
Severity: normal

After "info gnuplot", go to "Terminal_types::". One gets:


4 Terminal types


* Menu:

* complete_list_of_terminals::


The next page is just:


4.1 complete list of terminals
==

Gnuplot supports a large number of output formats.  These are selected
by choosing an appropriate terminal type, possibly with additional
modifying options.  See *note terminal::.

   This document may describe terminal types that are not available to
you because they were not configured or installed on your system.  To
see a list of terminals available on a particular gnuplot installation,
type 'set terminal' with no modifiers.

   Terminals marked 'legacy' are not built by default in recent gnuplot
versions and may not actually work.  @c <3 - all terminal stuff is
pulled from the .trm files

* Menu:

* Bugs::


without the complete list of terminals, contrary to the manual
in HTML.

-- System Information:
Debian Release: 11.0
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'testing-security'), (500, 
'stable-updates'), (500, 'unstable'), (500, 'testing'), (500, 'stable'), (1, 
'experimental')
Architecture: amd64 (x86_64)

Kernel: Linux 5.10.0-7-amd64 (SMP w/8 CPU threads)
Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE, 
TAINT_UNSIGNED_MODULE
Locale: LANG=POSIX, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

-- no debconf information

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#964062: libfplll-dev:amd64: missing static library libfplll.a

2020-06-30 Thread Vincent Lefevre 
Package: libfplll-dev
Version: 5.3.2-1
Severity: important

The static library libfplll.a is missing. This makes projects that
require static linking fail to build.

-- System Information:
Debian Release: bullseye/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'stable-updates'), (500, 
'unstable'), (500, 'testing'), (500, 'stable'), (1, 'experimental')
Architecture: amd64 (x86_64)

Kernel: Linux 5.7.0-1-amd64 (SMP w/8 CPU cores)
Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE, 
TAINT_UNSIGNED_MODULE
Locale: LANG=POSIX, LC_CTYPE=C.UTF-8 (charmap=UTF-8), LANGUAGE=POSIX 
(charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

Versions of packages libfplll-dev:amd64 depends on:
ii  libfplll65.3.2-1
ii  libgmp-dev   2:6.2.0+dfsg-6
ii  libmpfr-dev  4.0.2-1

libfplll-dev:amd64 recommends no packages.

libfplll-dev:amd64 suggests no packages.

-- no debconf information

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#863217: libgmp10:amd64: gmp_snprintf tries to allocate 18 EB on long strings

2020-01-21 Thread Vincent Lefevre 
Control: tags -1 fixed-upstream

On 2017-05-23 21:43:06 +0200, Vincent Lefevre wrote:
> Consider the following program:
> 
> #include 
> #include 
> 
> int main(void)
> {
>   int r;
>   long n = -1;
> 
>   r = gmp_snprintf (NULL, 0, "%2147483600s%100s%ln", "", "", );
>   printf ("%d %ld\n", r, n);
>   return 0;
> }
> 
> On my Debian/unstable x86_64 machine (GMP 6.1.2 provided by the Debian
> package), I get after 273 seconds:
> 
> GNU MP: Cannot allocate memory (size=18446744071562067968)

With upstream's GMP 6.2.0, I get:

-1 -1

So the bug is fixed upstream.

-- 
Vincent Lefèvre  - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#925927: libgmp10:amd64: failure with Clang's memory sanitizer: use-of-uninitialized-value

2019-03-28 Thread Vincent Lefevre 
Control: tags -1 upstream
Control: forwarded -1 
https://gmplib.org/list-archives/gmp-bugs/2019-March/004525.html

I also reported the bug upstream.

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#925927: libgmp10:amd64: failure with Clang's memory sanitizer: use-of-uninitialized-value

2019-03-28 Thread Vincent Lefevre 
Package: libgmp10
Version: 2:6.1.2+dfsg-4
Severity: important

The following program

#include 

int main (void)
{
  mpz_t z;

  mpz_init (z);
  mpz_set_ui (z, 1);
  mpz_dump (z);
  mpz_clear (z);

  return 0;
}

compiled with: clang-7 -fsanitize=memory -o tst tst.c -lgmp

gives the following error:

cventin:~> ./tst
Uninitialized bytes in __interceptor_puts at offset 1 inside [0x70100010, 2)
==17647==WARNING: MemorySanitizer: use-of-uninitialized-value
#0 0x7fbf03e6fe6a in __gmpz_dump 
(/usr/lib/x86_64-linux-gnu/libgmp.so.10+0x18e6a)
#1 0x49518a in main (/home/vlefevre/tst+0x49518a)
#2 0x7fbf03aeb09a in __libc_start_main 
(/lib/x86_64-linux-gnu/libc.so.6+0x2409a)
#3 0x41e2b9 in _start (/home/vlefevre/tst+0x41e2b9)

SUMMARY: MemorySanitizer: use-of-uninitialized-value 
(/usr/lib/x86_64-linux-gnu/libgmp.so.10+0x18e6a) in __gmpz_dump
Exiting
zsh: exit 77./tst

In particular, this failure prevents one from using -fsanitize=memory
with software that uses GMP.

-- System Information:
Debian Release: buster/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'stable-updates'), (500, 
'unstable'), (500, 'testing'), (500, 'stable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.19.0-4-amd64 (SMP w/12 CPU cores)
Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE, 
TAINT_UNSIGNED_MODULE
Locale: LANG=POSIX, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), LANGUAGE=POSIX 
(charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

Versions of packages libgmp10:amd64 depends on:
ii  libc6  2.28-8

libgmp10:amd64 recommends no packages.

libgmp10:amd64 suggests no packages.

-- no debconf information

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#844484: Warning: slow font initialization + graph not displayed in the window

2019-01-07 Thread Vincent Lefevre 
Control: found -1 5.2.5+dfsg1-1

Another issue is that the warning message doesn't end with a newline
character.

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#917256: gnuplot-qt: should provide gnuplot-x11 in order to avoid conflicting dependencies

2018-12-24 Thread Vincent Lefevre 
Package: gnuplot-qt
Version: 5.2.5+dfsg1-1
Severity: important

If I understand correctly, gnuplot-qt is like gnuplot-x11, but also
has Qt support. Thus it should provide gnuplot-x11 in order to satisfy
some dependencies.

As an example, if I want to install maxima, apt wants to remove
gnuplot-qt and replace it by gnuplot-x11, thus losing Qt support:

# apt install maxima
[...]
The following additional packages will be installed:
  gnuplot-x11 maxima-share
Suggested packages:
  texmacs maxima-doc xmaxima maxima-emacs
The following packages will be REMOVED:
  gnuplot-qt gnuplot-qt-dbgsym
The following NEW packages will be installed:
  gnuplot-x11 maxima maxima-share

The reason is that maxima recommends gnuplot-x11.

-- System Information:
Debian Release: buster/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'stable-updates'), (500, 
'unstable'), (500, 'testing'), (500, 'stable'), (1, 'experimental')
Architecture: amd64 (x86_64)

Kernel: Linux 4.19.0-1-amd64 (SMP w/8 CPU cores)
Locale: LANG=POSIX, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), LANGUAGE=POSIX 
(charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

Versions of packages gnuplot-qt depends on:
ii  gnuplot-data 5.2.5+dfsg1-1
ii  libc62.28-3
ii  libcairo21.14.10-1
ii  libedit2 3.1-20181209-1
ii  libgcc1  1:8.2.0-13
ii  libgd3   2.2.5-5
ii  libglib2.0-0 2.58.1-2
ii  liblua5.3-0  5.3.3-1
ii  libpango-1.0-0   1.42.1-1
ii  libpangocairo-1.0-0  1.42.1-1
ii  libqt5core5a 5.11.1+dfsg-6
ii  libqt5gui5   5.11.1+dfsg-6
ii  libqt5network5   5.11.1+dfsg-6
ii  libqt5printsupport5  5.11.1+dfsg-6
ii  libqt5svg5   5.11.1-2
ii  libqt5widgets5   5.11.1+dfsg-6
ii  libstdc++6   8.2.0-13
ii  libwxbase3.0-0v5 3.0.4+dfsg-7
ii  libwxgtk3.0-0v5  3.0.4+dfsg-7
ii  libx11-6 2:1.6.7-1

gnuplot-qt recommends no packages.

Versions of packages gnuplot-qt suggests:
ii  gnuplot-doc  5.2.5+dfsg1-1

-- no debconf information

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#905102: gnuplot-doc: gnuplot.pdf does not have images

2018-07-31 Thread Vincent Lefevre 
Package: gnuplot-doc
Version: 5.2.2+dfsg1-2
Severity: minor

The Gnuplot PDF manual gnuplot.pdf does not have images. The reason is
that upstream chose to use picins, which is not provided in Debian.
There are several solutions:

1. Ask to provide picins if the license issue can be solved:

http://www.tug.org/pipermail/tex-live/2007-December/015186.html

2. Use a patch to use wrapfig instead of picins. There were bugs in 2008:

https://sourceforge.net/p/gnuplot/bugs/655/

   but perhaps they have been solved, and perhaps this could still be
   useful even though the layout is not as good as with picins.

3. Use gnuplot.pdf from upstream, as suggested here:

https://sourceforge.net/p/gnuplot/feature-requests/423/

   if there are no documentation patches (alternatively, provide both
   the upstream gnuplot.pdf and the rebuilt one).

-- System Information:
Debian Release: buster/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'stable-updates'), (500, 
'unstable'), (500, 'testing'), (500, 'stable'), (1, 'experimental')
Architecture: amd64 (x86_64)

Kernel: Linux 4.17.0-1-amd64 (SMP w/8 CPU cores)
Locale: LANG=POSIX, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), LANGUAGE=POSIX 
(charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

-- no debconf information

-- 
debian-science-maintainers mailing list
debian-science-maintainers@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers