[Bug 1964741] [NEW] date command produces "invalid date" on day of DST change when calculating a time between 02:00 and 02:59
Public bug reported: When calculating a timestamp for a time between 02:00 and 02:59 in some locales, date will report "invalid date". This only occurs if the system time is currently during the day that standard time transitions to daylight time. Example: OS Version information: Distributor ID: Ubuntu Description:Ubuntu 20.04.4 LTS Release:20.04 Codename: focal System Locale is: America/Chicago System time the command was issued: Sun 13 Mar 2022 10:15:18 PM CDT Command issued: date +%s -d 'tomorrow 02:50' System Response: date: invalid date 'tomorrow 02:50' If the command issued is changed such that the calculated time is anything other than a time between 02:00 and 02:59 the output is normal. ** Affects: coreutils (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1964741 Title: date command produces "invalid date" on day of DST change when calculating a time between 02:00 and 02:59 To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/coreutils/+bug/1964741/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 310778] [NEW] Can't connect to wireless network with intrepid
Public bug reported: Ubuntu intrepid is not connecting to my wireless network. gutsy connected fine. $ lsb_release -rd Description:Ubuntu 8.10 Release:8.10 Updates current as of 2008/12/22. My wireless card is a Lucent Technologies Orinoco Gold. The following shows up in lsmod: orinoco_cs 21516 1 orinoco48020 1 orinoco_cs hermes 14976 2 orinoco_cs,orinoco pcmcia 43052 1 orinoco_cs I'm expecting to be able to connect to my wireless network using the following gui procedure. However, my network is not being connected to: 1. Left-click the little icon in the upper right that looks like two monitors with a red thing in front of them; it puts up a box that says "No network connection" when you hover over it. 2. Select "Connect to Hidden Wireless Network..." 3. Network name: use the network's SSID. Wireless Security: WEP 40/128-bit Key Key: enter the 40-bit (10 hex digit) key Check "show key" to check for typos. WEP Index: 1 (Default) Authentication: Open System 4. Click "Connect". 5. The little icon turns into an animation with a bluish clockwise rotating blur. 6. After a minute or so the animation stops and a notification pops up by the icon: Disconnected: the network connection has been disconnected. I'm sorry that's all the information I have to put in this report. ** Affects: ubuntu Importance: Undecided Status: New -- Can't connect to wireless network with intrepid https://bugs.launchpad.net/bugs/310778 You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 183671] Re: Race with drawing vs. vt switching
And if you use a lock, SIGUSR2 will need to be blocked before locking so the handler won't deadlock if the signal is delivered to a thread while it holds the lock. Or it may be possible to just have the drawing code block SIGUSR2 to prevent a vt switch (if that works), and have the signal handler set or clear a volatile indicating whether drawing is allowed or not. I'm not sure how the vt switching and signals work together, but I'm sure someone knows. -- Race with drawing vs. vt switching https://bugs.launchpad.net/bugs/183671 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 183671] Race with drawing vs. vt switching
Public bug reported:
Binary package hint: usplash
The bogl functions use global variable bogl_drawing to indicate they may
be writing into the framebuffer, e.g.:
extern int bogl_drawing;/* Currently drawing? */
/* Set pixel (X,Y) to color C. */
void
bogl_tcfb_pixel (int x, int y, int c)
{
bogl_drawing = 1;
assert (x >= 0 && x < bogl_xres);
assert (y >= 0 && y < bogl_yres);
put_var (bogl_frame + y * bogl_line_len, x, cmap_lookup(c), bpp);
bogl_drawing = 0;
}
This variable is used by vt_switch, which is a signal handler, to try to avoid
switching the vt when code may be drawing into the framebuffer:
/* Signal handler called whenever the kernel wants to switch to or
from our tty. */
static void
vt_switch (int sig unused)
{
signal (SIGUSR2, vt_switch);
/* If a BOGL drawing function is in progress then we cannot mode
switch right now because the drawing function would continue to
scribble on the screen after the switch. So disable further
drawing and schedule an alarm to try again in .1 second. */
if (bogl_drawing)
{
draw_disable ();
signal (SIGALRM, vt_switch);
{
struct itimerval duration;
duration.it_interval.tv_sec = 0;
duration.it_interval.tv_usec = 0;
duration.it_value.tv_sec = 0;
duration.it_value.tv_usec = 10;
if (-1 == setitimer (ITIMER_REAL, &duration, NULL))
bogl_fail ("can't set timer: %s", strerror (errno));
}
return;
}
There is a race here. A drawing function may be called, and about to set
bogl_drawing to one. At this point vt_switch may be called. It will see
bogl_drawing is zero, the vt will be switched, and when control returns to the
drawing function it will "scribble on the screen after the switch".
Also, bogl_drawing is not declared volatile, and neither are the
variables written to by draw_disable, and even if they were, C doesn't
have a memory model that specifies the interactions between threads and
variables so nothing is certain.
Making this work properly would require a lock protecting one or more
variables that indicate whether it's ok to draw (i.e., we're on the
right vt), and whether it's ok to switch the vt (i.e., we're not
drawing). It also requires figuring out what to do if a drawing
operation is attempted when drawing is not ok, e.g., it may not be ok
just to block until the vt is switched back because in the meantime the
usplash fifo could fill up and block writers, and it may not be ok just
to blow the operation off because then the framebuffer won't have the
proper information when the vt is switched back.
** Affects: usplash (Ubuntu)
Importance: Undecided
Status: New
--
Race with drawing vs. vt switching
https://bugs.launchpad.net/bugs/183671
You received this bug notification because you are a member of Ubuntu
Bugs, which is the bug contact for Ubuntu.
--
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 113755] Re: pccard network interface is lost on resume (+ fix)
Thanks for the comment. I wonder how many people have this problem?
Somehow everything manages to work fine when I run debian unstable on
the same system, and its 55-down-interfaces.sh sets INTERFACES after
pccardctl eject:
#!/bin/sh
pccardctl eject
# Get rid of any currently running dhclients
killall dhclient dhclient3 2>/dev/null
# Find the currently running network interfaces...
INTERFACES=`/sbin/ifconfig | awk '/^[^ ]+/ {print $1}'`
# And shut them down
for x in $INTERFACES; do
ifdown $x;
ifconfig $x down;
done
--
pccard network interface is lost on resume (+ fix)
https://bugs.launchpad.net/bugs/113755
You received this bug notification because you are a member of Ubuntu
Bugs, which is the bug contact for Ubuntu.
--
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 113755] Re: pccard network interface is lost on resume (+ fix)
The ifdown/ifup doesn't always fix the problem either. Sadly, I'll be setting my kids' computer up with Debian instead of Ubuntu. -- pccard network interface is lost on resume (+ fix) https://bugs.launchpad.net/bugs/113755 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 72458] Re: Firefox icon is too big
My five year old dragged the icon to the desktop and immediately said "this icon is too big". If it's too big for a preschooler, it's too big. He's running 7.04. He's not sure whether that's Feisty or what, and doesn't know how to figure it out. The firefox icon can be shrunk to about 44x44 with right-click stretch icon, but that's still big compared to other icons that can be shrunk to 32x32 or smaller. I had to edit firefox.desktop and set Icon=/usr/share/firefox/icons/mozicon50.xpm for him. It would be great if Nautilus allowed icons to be shrunk further, and if it pre-shrunk icons to a reasonable size, and if it allowed the icon file to be easily changed, but this probably isn't the right place to suggest that. -- Firefox icon is too big https://bugs.launchpad.net/bugs/72458 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 113755] Re: pccard network interface is lost on resume (+ fix)
The fix I gave may not be correct, and it may not work all the time. I checked my Debian system which does not have this problem, and it has an "allow-hotplug eth1" line in /etc/network/interfaces. Added this line to the Ubuntu system does not fix the problem. In any case, having resume bring up the same set of interfaces that were active on suspend seems most correct, so an allow-hotplug "fix" doesn't seem correct for any system. My current fix is to add a resume.d script to do "ifdown eth1; ifup eth1". -- pccard network interface is lost on resume (+ fix) https://bugs.launchpad.net/bugs/113755 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 113755] pccard network interface is lost on resume (+ fix)
Public bug reported:
Binary package hint: acpi-support
When I suspend then resume on an A31 Thinkpad with a Lucent Wavelan
card, the Wavelan card's network interface is no longer configured.
The problem is in /etc/acpi/suspend.d/55-down-interfaces.sh. "pccardctl
eject" results in the "eth1" interface going away, so when ifconfig is
run to set INTERFACES, the resulting list does not include eth1, so the
corresponding script /etc/acpi/resume.d/62-ifup.sh doesn't bring eth1
back up on resume. This can be fixed by running ifconfig to set
INTERFACES before running pccardctl. Here's what my stock 55-down-
interfaces.sh looks like:
#!/bin/sh
pccardctl eject
# Get rid of any currently running dhclients
killall dhclient dhclient3 2>/dev/null
# Find the currently running network interfaces...
INTERFACES=`/sbin/ifconfig | awk '/^[^ ]+/ {print $1}'`
# And shut them down
for x in $INTERFACES; do
ifdown $x;
ifconfig $x down;
done
This problem has unfortunately resulted in my kids thinking Linux is as
junky as Windows.
** Affects: acpi-support (Ubuntu)
Importance: Undecided
Status: Unconfirmed
--
pccard network interface is lost on resume (+ fix)
https://bugs.launchpad.net/bugs/113755
You received this bug notification because you are a member of Ubuntu
Bugs, which is the bug contact for Ubuntu.
--
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
