Linux Pan crashes on exit on x64 system.

classic Classic list List threaded Threaded
15 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Linux Pan crashes on exit on x64 system.

Joe Zeff-2
I've just had to reinstall my system and took advantage of the fact to
upgrade from X86-PAE to X64.  Now, it crashes on exit every time I run
it and the posts in the last group I accessed aren't marked as read the
next time I run it.  (I've corrected that by editing my .newsrc-1 file.)
  There is a record of it here:
https://retrace.fedoraproject.org/faf/reports/522761/  Please note that
this has been occurring since 12/30/2014, but it hasn't even been
assigned yet.  If there's an appropriate place to report it, please let
me know, and what needs to be included.

_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Joe Zeff-2
On 03/24/2017 01:38 PM, Joe Zeff wrote:
> There is a record of it here:
> https://retrace.fedoraproject.org/faf/reports/522761/  Please note that
> this has been occurring since 12/30/2014, but it hasn't even been
> assigned yet.

Correction: There's a bugzilla I reported two days ago at
https://bugzilla.redhat.com/show_bug.cgi?id=1435042

_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Duncan-42
In reply to this post by Joe Zeff-2
Joe Zeff posted on Fri, 24 Mar 2017 13:38:17 -0700 as excerpted:

> I've just had to reinstall my system and took advantage of the fact to
> upgrade from X86-PAE to X64.  Now, it crashes on exit every time I run
> it and the posts in the last group I accessed aren't marked as read the
> next time I run it.  (I've corrected that by editing my .newsrc-1 file.)
>   There is a record of it here:
> https://retrace.fedoraproject.org/faf/reports/522761/  Please note that
> this has been occurring since 12/30/2014, but it hasn't even been
> assigned yet.  If there's an appropriate place to report it, please let
> me know, and what needs to be included.

Your followup RH bugzilla report from two days ago:

https://bugzilla.redhat.com/show_bug.cgi?id=1435042


I see pkovar has already CCed the bug, so should get whatever fix and
include it in pan upstream. =:^)


x64... is unfortunately somewhat ambiguous.

Is this x64 the newish unholy and problematic 64-bit/32-bit amalgam
that's 64-bit kernel and IIRC some kernel calls but otherwise 32-bit
userspace?  If so, that may be the problem as that's fairly new and
uncommon and won't have been debugged very well (similar to amd64 aka
x86_64 in ~ 2005).

IMO, FWIW, this is a too-late "solution" to a problem that time has
already pretty much worked out on its own.  If it would have been
introduced back about the time normal amd64 aka x86_64 was introduced, it
might have gone somewhere and may well have inherited the x86 successor
crown much as amd64 did in its absence, but by now the problems of amd64
have pretty much been worked out and it's more common than the now legacy
32-bit x86, and I simply don't see it being worth the trouble or going
much of anywhere.

Or is it the more conventional 64-bit amd64/x86_64, that can also run
legacy 32-bit x86, and normally does as shipped by most distros in what
is known as a multilib installation?  I believe this is called x64 in at
least some MS contexts, but it's normally known as amd64 or x86_64 in
Linux contexts.

Of course x64 was also sometimes used to refer to Intel's now rather
legacy itanium (aka itanic, due to the cost, inefficiency and
unworkability that pretty much doomed it from the start), but the chances
of you just getting that sort of hardware now (as opposed to trying to
continue to run it after a purchase in the 200x's... if you had the money
then... or perhaps purchasing it second-hand for pennies on the dollar
more recently simply as a curiosity) are pretty small.


Meanwhile...  Not being a dev I don't get much from cores/backtraces and
can't really help you with the crash itself.  But something I *can* help
with...

Note that pan saves group read-messages state, writing it to the newsrc
files, when you switch groups.

While pan on amd64 has been stable for me for quite some time now, I've
been running it since pan 0.11-something (then on 32-bit x86) on gnome-1
back in 2002, and neither it, nor X, nor the hardware I've been running
it on, has always been stable during that now decade and a half.

So long ago I developed the habit of keeping some idea in my head of how
many messages I had read that hadn't yet been saved, and once it gets to
a level that marking all those messages read again on restart is going to
be a hassle (generally every 5-500 messages or every 2-50 threads
depending on the group and how fast I'm going), I'll quickly switch out
to some other group and back again, triggering pan's exit-group save-read-
messages logic.

That way, if pan crashes, regardless of whether it's due to pan itself,
or X, or the kernel, or hardware, I'll only have a few messages/threads
to mark-read again when I restart it.

Of course if you have the option set to mark all messages in a group read
when you switch groups, you may want to rethink your choice on that,
first. =:^)

--
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman


_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Joe Zeff-2
On 03/25/2017 03:52 PM, Duncan wrote:
> Is this x64 the newish unholy and problematic 64-bit/32-bit amalgam
> that's 64-bit kernel and IIRC some kernel calls but otherwise 32-bit
> userspace?

No.  I think that's called PAE, at least for Fedora.  It's a proper
X86-64 bit system.

_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Zan Lynx
In reply to this post by Duncan-42


On 3/25/2017 4:52 PM, Duncan wrote:
> x64... is unfortunately somewhat ambiguous.
>
> Is this x64 the newish unholy and problematic 64-bit/32-bit amalgam
> that's 64-bit kernel and IIRC some kernel calls but otherwise 32-bit
> userspace?  If so, that may be the problem as that's fairly new and
> uncommon and won't have been debugged very well (similar to amd64 aka
> x86_64 in ~ 2005).

See
http://stackoverflow.com/questions/7635013/difference-between-x86-x32-and-x64-architectures

x86 is the traditional Intel 32 bit architecture. x64 is used for the
AMD64, which is what all common AMD and Intel 64 bit processors are
today. x32 is AMD64 but only using 32 bit pointers. IA64 is Itanium.

--
                 Knowledge is Power -- Power Corrupts
                         Study Hard -- Be Evil


_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Duncan-42
Zan Lynx posted on Sat, 25 Mar 2017 17:06:04 -0600 as excerpted:

> On 3/25/2017 4:52 PM, Duncan wrote:
>> x64... is unfortunately somewhat ambiguous.
>>
>> Is this x64 the newish unholy and problematic 64-bit/32-bit amalgam
>> that's 64-bit kernel and IIRC some kernel calls but otherwise 32-bit
>> userspace?  If so, that may be the problem as that's fairly new and
>> uncommon and won't have been debugged very well (similar to amd64 aka
>> x86_64 in ~ 2005).
>
> See
> http://stackoverflow.com/questions/7635013/difference-between-x86-x32-
and-x64-architectures
>
> x86 is the traditional Intel 32 bit architecture. x64 is used for the
> AMD64, which is what all common AMD and Intel 64 bit processors are
> today. x32 is AMD64 but only using 32 bit pointers. IA64 is Itanium.

But the x64 name, due to its ambiguous use for other archs, never took
off at least in the Linux context, where what you call x64 is generally
referred to as either amd64 or x86_64.

Meanwhile, x32 is what I referred to as that unholy come-lately amalgam
(but couldn't remember the name of), and of course ia64 is itanic/itanium.

So using x64 in the Linux context marks you as possibly a Microsofty
(where I believe it /is/ commonly used for x86_64), and in any case, as
someone who doesn't know enough about what they're talking about in the
Linux context to use the more common x86_64 or amd64 terms, if indeed
that's what they're talking about at all, not ia64 or x32.

--
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman


_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Duncan-42
In reply to this post by Joe Zeff-2
Joe Zeff posted on Sat, 25 Mar 2017 15:59:12 -0700 as excerpted:

> On 03/25/2017 03:52 PM, Duncan wrote:
>> Is this x64 the newish unholy and problematic 64-bit/32-bit amalgam
>> that's 64-bit kernel and IIRC some kernel calls but otherwise 32-bit
>> userspace?
>
> No.  I think that's called PAE, at least for Fedora.  It's a proper
> X86-64 bit system.

As Zan pointed out, it's x32 I was referring to with the unholy amalgam
reference.

PAE is different.  It's still a normal 32-bit x86 system in general, but
using the PAE feature of more modern x86 hardware to extend the total
kernel-addressable address space to 64 GiB, up from the normal 4 GiB.

But per-process space remains stuck at 4 GiB total, with a usual 2 GiB
userspace and 2 GiB kernelspace (2:2) split.  Using other kernel config
options this split can be made 1:3 or 3:1, or with a bit more cost to
kernel/userspace switching times, userspace and kernelspace can each be
given their separate 4 GiB address spaces, but even with both PAE and
separate user/kernelspace, a 32-bit process remains limited to 4 GiB
userspace.

Full 64-bit removes that userspace limit, letting both userspace and
kernelspace address huge amounts of memory well beyond anything even big
corps tend to be able to afford at this point (tho AFAIK some are getting
close to hitting the 40-bit hardware limit the original x86_64 hardware
shipped with, using the other bits for other things, and I believe the
present hardware limit has now been upped to 48 bits).

Running 32-bit processes on a 64-bit kernel ups the limit for the kernel
(no more 64-GiB system limit as imposed by 32-bit but with PAE), so you
can run more processes, but not for the still 32-bit processes.  Running
full 64-bit processes as well ups the limit for them as well.

What my "unholy amalgam" x32 does is somewhat better than normal 32-bit
x86 on an x86_64 kernel, as it allows the still 32-bit apps to use other
hardware features (more hardware registers, etc) otherwise reserved for
64-bit only due to having to maintain backward compatibility in the 32-
bit space, and they use some newer 64-bit kernel calls, but they're still
limited to 32-bit address pointers and thus 4 GiB per app, not so much of
a problem for most apps especially back when amd64 aka x86_64 was first
introduced, but slowly becoming more of a problem as individual apps, now
broken free of the previous 2-4 GiB address space limit of 32-bit, grow
to sometimes take advantage of that available extra memory.

Which means it would have been great back in the day when most apps were
still assuming 32-bit, and were yet to be ported to amd64/x86_64 anyway,
but isn't so great now, since it means more porting costs now, for only a
relatively small gain in lower per-app memory costs, and arguably a loss
in terms of limits due to the reimposition of the 32-bit 4 GiB memory cap.

(Tho I'm not sure, arguably with the 64-bit x32 kernel not using PAE, if
setup correctly, x32 apps might be able to use it, thus allowing per-app
access, if at a slight overhead cost, to 64 GiB.  But I still think the
cost and hassle aren't worth it since full x86_64 took over.)

And of course there's now arm64 coming on quite strong as well, with
arguably more interest in porting to and supporting it than yet another
x86 variant.

--
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman


_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Petr Kovar-2
In reply to this post by Duncan-42
On Sat, 25 Mar 2017 22:52:52 +0000 (UTC)
Duncan <[hidden email]> wrote:

> Joe Zeff posted on Fri, 24 Mar 2017 13:38:17 -0700 as excerpted:
>
> > I've just had to reinstall my system and took advantage of the fact to
> > upgrade from X86-PAE to X64.  Now, it crashes on exit every time I run
> > it and the posts in the last group I accessed aren't marked as read the
> > next time I run it.  (I've corrected that by editing my .newsrc-1 file.)
> >   There is a record of it here:
> > https://retrace.fedoraproject.org/faf/reports/522761/  Please note that
> > this has been occurring since 12/30/2014, but it hasn't even been
> > assigned yet.  If there's an appropriate place to report it, please let
> > me know, and what needs to be included.
>
> Your followup RH bugzilla report from two days ago:
>
> https://bugzilla.redhat.com/show_bug.cgi?id=1435042
>
>
> I see pkovar has already CCed the bug, so should get whatever fix and
> include it in pan upstream. =:^)

Heh, I just happen to be CC'ed on all Pan bugs at
https://bugzilla.redhat.com/ because it's usually me updating the pan
package for Fedora. :)

Anyway, the crash on exit problem seems to be related to the gnome-keyring
code and shouldn't happen if you compile pan with --enable-gkr=no.

Joe, I've slightly updated the instructions on
http://pan.rebelbase.com/bugs/#backtrace, would you be able to follow them?
If you use the official Fedora pan package, you also need to install
debuginfo if it's not already installed:

dnf debuginfo-install pan

I think your crash could be caused by the same problem as the one shown
in this backtrace I generated the other day:

https://pastebin.com/C5AbH53v

The problem seems to be with the way gnome-keyring is freeing memory on
exit:

https://github.com/GNOME/pan2/blob/master/pan/gui/pan.cc#L1145

Anybody with C++ skills wants to take this one? :)

Cheers,
pk

_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Joe Zeff-2
On 03/27/2017 09:36 AM, Petr Kovar wrote:
> Joe, I've slightly updated the instructions on
> http://pan.rebelbase.com/bugs/#backtrace, would you be able to follow them?
> If you use the official Fedora pan package, you also need to install
> debuginfo if it's not already installed:
>
> dnf debuginfo-install pan

I don't compile if I can install.  Compiling is work, and I'm retired.
And, I don't think dnf includes a command "debuginfo-install."  I'll see
about getting what you need installed, assuming Fedora provides it, and
try again.

_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Joe Zeff-2
In reply to this post by Petr Kovar-2
On 03/27/2017 09:36 AM, Petr Kovar wrote:
> dnf debuginfo-install pan

dnf list \*pan\* | grep debug

returns nothing.

_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Duncan-42
In reply to this post by Petr Kovar-2
Petr Kovar posted on Mon, 27 Mar 2017 18:36:42 +0200 as excerpted:

> Anyway, the crash on exit problem seems to be related to the
> gnome-keyring code and shouldn't happen if you compile pan with
> --enable-gkr=no.

That would be why I've not seen it here. =:^)

I run (a lite) kde-plasma here, not gnome, and don't pull in anything
gnome-ish I can get away without.

(Probably no one but gentooers care about the following, but FWIW...)  
Being gentoo, that means I can get away with installing a lot less than
normal (tho I do need build-time deps too), because most of the build-
time options that pull in more deps are exposed via USE flags, and my
policy is to leave USE flags off unless I have some reason (like a
depending package requiring it) I need them.  I even set -* at the
beginning of my make.conf/use file, in ordered to avoid all the flags my
profile would normally turn on, and have literally justified to myself
every single flag I have on, both system-wide and for individual packages
where they differ from system defaults.

--
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman


_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Joe Zeff-2
On 03/27/2017 07:10 PM, Duncan wrote:
> I run (a lite) kde-plasma here, not gnome, and don't pull in anything
> gnome-ish I can get away without.

And I run Xfce, but can't seem to get completely away from a few bits of
gnome such as the key-ring.

_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Duncan-42
Joe Zeff posted on Mon, 27 Mar 2017 19:18:48 -0700 as excerpted:

> On 03/27/2017 07:10 PM, Duncan wrote:
>> I run (a lite) kde-plasma here, not gnome, and don't pull in anything
>> gnome-ish I can get away without.
>
> And I run Xfce, but can't seem to get completely away from a few bits of
> gnome such as the key-ring.

Well, xfce is at least gtk-based, as is gnome.  It may even be gtk3-based
(the alternative being gtk2), as is gnome, I don't know.  But
particularly if it's gtk3 based, there's a lot of common deps between the
two and installing the gnome keyring would be a natural as at least some
xfce and general gtk apps are likely to use it.

By contrast, kde is qt based, with kde-frameworks5 on which plasma5 is
based being qt5 based (was qt3 for kde3, qt4 for kde4...).  So it's a
different toolkit and installing anything gnome-based here will probably
be a lot more expensive in terms of additional dependencies than it will
be for you, as your desktop is already gtk-based, while mine is qt-based.

In fact, at one point back in the late kde3 era, I was very close to
being able to rid my system of gtk entirely, with pan being the only
major gtk-based app I didn't have a good qt-based substitute for.

Unfortunately kde4 then happened and the kde4 devs refused to continue
supporting the still working kde3 even while kde4 was still very **VERY**
broken, and even after promising there'd be support as long as there were
users.

I did manage to get (and keep) the kde4 desktop, called plasma in kde4
and after, working, but I was left with little choice but to find
actually working replacements for most of the rest of kde.  I hacked
together a multi-key-hotkey/menu system of sorts for myself, primarily
using bash, and found independent or qt-based solutions for several other
things, but irony of irony, it'd now be much easier to dump kde and qt
entirely than gtk, as for kde/plasma the only real big kde/qt thing I'd
have to find a properly working alternative for would be the plasma
desktop itself, while on the gtk side, in addition to still running pan
for news, I now run claws-mail for both mail and feeds, and firefox for
the web, all of which I use enough to be "big things" in terms of
switching, and all of which are gtk-based.

Tho most of them (including pan) remain gtk2-based.  I guess firefox
itself is both gtk2 and gtk3 based now, but the others are gtk2-based,
perhaps with a gtk3 option (as pan has), but no sign of dropping gtk2 as
an option as well, out a couple years, anyway.

Tho of course chromium is qt-based and now its proprietary counterpart
chrome is the most popular browser on the web, so I could switch to it,
if necessary.

Beyond a couple years, wayland will probably be gaining dominance over X,
and unless someone adopts gtk2 and ports/updates it for wayland support,
eventually gtk2 and all the apps that haven't switched to either gtk3 or
qt5 will be on their slow deathward spiral.  But firefox is mostly gtk3
ported and of course their's chromium as an alternative, pan has a gtk3
option altho gtk2 continues to be the recommended build, AFAIK, claws is
working on a gtk3 port, and qt5 has wayland support which plasma is well
on its way to supporting as well.  So wayland will bring changes, but it
looks like I should be in fair shape even if many of my big apps are
still gtk2-based ATM.

--
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman


_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Petr Kovar-2
In reply to this post by Joe Zeff-2
On Mon, 27 Mar 2017 11:38:54 -0700
Joe Zeff <[hidden email]> wrote:

> On 03/27/2017 09:36 AM, Petr Kovar wrote:
> > Joe, I've slightly updated the instructions on
> > http://pan.rebelbase.com/bugs/#backtrace, would you be able to follow them?
> > If you use the official Fedora pan package, you also need to install
> > debuginfo if it's not already installed:
> >
> > dnf debuginfo-install pan
>
> I don't compile if I can install.  Compiling is work, and I'm retired.
> And, I don't think dnf includes a command "debuginfo-install."

First hit on Google:

http://dnf-plugins-core.readthedocs.io/en/latest/debuginfo-install.html

Cheers,
pk

_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Linux Pan crashes on exit on x64 system.

Joe Zeff-2
On 03/28/2017 07:27 AM, Petr Kovar wrote:
> First hit on Google:
>
> http://dnf-plugins-core.readthedocs.io/en/latest/debuginfo-install.html

Installing the pan debuginfo now.

_______________________________________________
Pan-users mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/pan-users
Loading...