On Fri, Jun 15, 2007 at 05:26:41PM -0700, Daniel Burrows [EMAIL PROTECTED]
was heard to say:
So, I need to find a way to do wget_wch() in my main thread, while
still unblocking when a background thread posts a message to my global
queue. I guess this means I have to make some sort of dummy
I've finally gotten a chance to look into this again, and I may have
a meaningful clue as to what's happening.
I ran an strace, and although I can't reproduce the race on this
single-CPU laptop, I *did* catch a background thread writing to stdout!
It's the input thread, which does nothing but
On 2007-03-27 08:11:51 -0700, Daniel Burrows wrote:
Thomas,
I've gone over an strace provided by Vincent, and it looks pretty
clear that all the write() system calls are from a single thread -- and
the writes that overlap each other on stdout are actually just two
writes, being called
Thomas,
I've gone over an strace provided by Vincent, and it looks pretty
clear that all the write() system calls are from a single thread -- and
the writes that overlap each other on stdout are actually just two
writes, being called in sequence. The fact that they end up interleaved
seems
On 2007-03-15 20:23:14 -0700, Daniel Burrows wrote:
Well, I can't reproduce this, and I don't see many obvious places
in the code that would be causing it.
With aptitude 0.4.4-4, this is more difficult to reproduce. For
instance, I've tried to reproduce the problem, and I had to launch
On Fri, Mar 16, 2007 at 09:13:48AM +0100, Vincent Lefevre [EMAIL PROTECTED]
was heard to say:
On 2007-03-15 20:23:14 -0700, Daniel Burrows wrote:
I tried adding some serialization to make the drawing code
explicitly safe (previously it relied on being called only from the
main thread; all
Well, that's just fascinating. What I see in there immediately is the
menu bar and quickhelp overlaying themselves:
Actions Undo Package RAectsioolnsve r U n doS ear cPhac k agOep t ioRness o
lvVeir e wSs eHaelrpcESCh[ K Options Views HelpESC[KESC[2;1HC-T:C -MeT:n uM
e n?u : ?:H
Hi,
Well, I can't reproduce this, and I don't see many obvious places
in the code that would be causing it. I tried adding some serialization
to make the drawing code explicitly safe (previously it relied on being
called only from the main thread; all the call sites I could find *should*
be
On Wed, Mar 14, 2007 at 01:20:26AM +0100, Vincent Lefevre [EMAIL PROTECTED]
was heard to say:
At startup, aptitude sends random characters to the terminal,
including escape sequences. Some of them have the effect to send
terminal data to the printer (this is not always reproducible),
which
On 2007-03-14 06:56:13 -0700, Daniel Burrows wrote:
Could you try running aptitude | tee /tmp/aptitude.out and sending
the resulting file to this bug? I'm not sure I can decode terminal
escape sequences, but that will at least rule out obvious things like
human-readable messages being
On 2007-03-14 01:20:26 +0100, Vincent Lefevre wrote:
Package: aptitude
Version: 0.4.4-3
FYI, I have the same problem with 0.4.4-4.
--
Vincent Lefèvre [EMAIL PROTECTED] - Web: http://www.vinc17.org/
100% accessible validated (X)HTML - Blog: http://www.vinc17.org/blog/
Work: CR INRIA - computer
Package: aptitude
Version: 0.4.4-3
Severity: grave
Tags: security
Justification: user security hole
At startup, aptitude sends random characters to the terminal,
including escape sequences. Some of them have the effect to send
terminal data to the printer (this is not always reproducible),
which
12 matches
Mail list logo