jma's questions

When I apt-get dist-upgrade, several 4.x kernels are reliably generated.

jeff@starshine:boot $ ls -lt
total 135676
-rw-r--r-- 1 root root 68578468 Dec 23 09:27 initrd.img-5.0.0-37-generic
-rw-r--r-- 1 root root 18585279 Dec 23 09:26 initrd.img-4.4.0-59-generic
-rw-r--r-- 1 root root 18585112 Dec 23 09:26 initrd.img-4.10.0-42-generic
-rw-r--r-- 1 root root 18585676 Dec 23 09:26 initrd.img-4.13.0-46-generic
drwxr-xr-x 5 root root     1024 Dec 12 12:56 grub
-rw------- 1 root root  8806136 Nov 13 22:36 vmlinuz-5.0.0-37-generic
-rw-r--r-- 1 root root   224480 Nov 13 20:35 config-5.0.0-37-generic
-rw------- 1 root root  4440495 Nov 13 20:35 System.map-5.0.0-37-generic
-rw-r--r-- 1 root root   182704 Jan 28  2016 memtest86+.bin
-rw-r--r-- 1 root root   184380 Jan 28  2016 memtest86+.elf
-rw-r--r-- 1 root root   184840 Jan 28  2016 memtest86+_multiboot.bin
drwx------ 2 root root    12288 Dec 17  2015 lost+found
drwx------ 3 root root     4096 Jan  1  1970 efi
jeff@starshine:boot $ 

Sometimes this causes the upgrade to run out of space. The 4.x kernels appear no longer to exist on my machine:

jeff@starshine:boot $ dpkg --list | grep linux- | grep -Ev '^rc'
ii  binutils-x86-64-linux-gnu                 2.32-7ubuntu4                                       amd64        GNU binary utilities, for x86-64-linux-gnu target
ii  linux-base                                4.5ubuntu1                                          all          Linux image base package
ii  linux-firmware                            1.178.7                                             all          Firmware for Linux kernel drivers
ii  linux-generic                             5.0.0.37.39                                         amd64        Complete Generic Linux kernel and headers
ii  linux-headers-5.0.0-37                    5.0.0-37.40                                         all          Header files related to Linux kernel version 5.0.0
ii  linux-headers-5.0.0-37-generic            5.0.0-37.40                                         amd64        Linux kernel headers for version 5.0.0 on 64 bit x86 SMP
ii  linux-headers-generic                     5.0.0.37.39                                         amd64        Generic Linux kernel headers
ii  linux-image-5.0.0-37-generic              5.0.0-37.40                                         amd64        Signed kernel image generic
ii  linux-image-generic                       5.0.0.37.39                                         amd64        Generic Linux kernel image
ii  linux-libc-dev:amd64                      5.0.0-37.40                                         amd64        Linux Kernel Headers for development
ii  linux-modules-5.0.0-37-generic            5.0.0-37.40                                         amd64        Linux kernel extra modules for version 5.0.0 on 64 bit x86 SMP
ii  linux-modules-extra-5.0.0-37-generic      5.0.0-37.40                                         amd64        Linux kernel extra modules for version 5.0.0 on 64 bit x86 SMP
ii  linux-signed-generic                      5.0.0.37.39                                         amd64        Complete Signed Generic Linux kernel and headers (dummy transitional package)
ii  linux-sound-base                          1.0.25+dfsg-0ubuntu5                                all          base package for ALSA and OSS sound systems
ii  syslinux-common                           3:6.04~git20171011.af7e95c3+dfsg1-6                 all          collection of bootloaders (common)
ii  syslinux-legacy                           2:3.63+dfsg-2ubuntu9                                amd64        Bootloader for Linux/i386 using MS-DOS floppies
jeff@starshine:boot $ 

Poking through grub configs, I've not found what is triggering these extra images to be created. Can anyone offer any pointers?

Since upgrading to Ubuntu 19.04 (wm=i3, not gnome desktop), when my machine is idle for long enough the screens blank and neither the keyboard nor the mouse will awaken them. Up until that time, the host works normally.

I can still ssh into the machine, so it is not suspended. Indeed, via ssh I can see that what I do on the keyboard is still being accepted and acted upon by the host.

As an experiment, I put hooks in the pm-* utilities so that they would log their invocations and refuse to run (i.e., do nothing successfully). I also tried disabling xautolock to see if this is somehow related to i3-lock, but I still see the same behaviour.

If I quit (or kill) the i3 session, lightdm somehow wakes the screens.

I suspect this is not BIOS related, since pre-upgrade the host was behaving correctly.

So, in brief, this appears to be simply the screens being put into power-saving mode and not being wake-able.

FWIW, my graphics is this:

  *-display                 
       description: VGA compatible controller
       product: Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller
       vendor: Intel Corporation
       physical id: 2
       bus info: pci@0000:00:02.0
       version: 06
       width: 64 bits
       clock: 33MHz
       capabilities: msi pm vga_controller bus_master cap_list rom
       configuration: driver=i915 latency=0
       resources: irq:34 memory:f7800000-f7bfffff memory:e0000000-efffffff ioport:f000(size=64) memory:c0000-dffff

Any suggestions for what to look at or how to debug this?

My laptop sometimes doesn't suspend/sleep when I close the lid. (I can tell because the fan still runs a bit, for example.) Sometimes it does but won't wake. Sometimes it sort of does, sort of doesn't. Sort of doesn't in the sense that the battery is just way weaker than it should be when I wake it. The kern.log shows things like this:

Jan 17 22:37:17 starshine kernel: [32612.982283] pci_bus 0000:04: Allocating resources
Jan 17 22:37:17 starshine kernel: [32612.982312] pci_bus 0000:06: Allocating resources
Jan 17 22:37:17 starshine kernel: [32612.982376] pci_bus 0000:07: Allocating resources
Jan 17 22:37:17 starshine kernel: [32612.982416] pci_bus 0000:08: Allocating resources
Jan 18 20:00:23 starshine kernel: [32623.080391] PM: Syncing filesystems ... done.
Jan 18 20:00:23 starshine kernel: [32623.344971] PM: Preparing system for sleep (mem)
Jan 18 20:00:23 starshine kernel: [32623.345230] Freezing user space processes ... (elapsed 0.002 seconds) done.
Jan 18 20:00:23 starshine kernel: [32623.347439] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
Jan 18 20:00:23 starshine kernel: [32623.348916] PM: Suspending system (mem)
Jan 18 20:00:23 starshine kernel: [32623.348957] Suspending console(s) (use no_console_suspend to debug)

Any suggestions what to look at to make this a bit more sane?

(I know I may have just presented three different problems, or not, but I suspect the strategy for looking for solutions will be similar.)

Thanks.

When I connect to a wifi network automatically or by choosing it with nm-applet, and assuming I've successfully connected with WPA2 previously, the connection is fine.

If I connect by typing nmcli device wifi connect network-name, then I'm asked for the password (via nm's dialog popup box), at which point the connection succeeds.

Is there a way I can tell nmcli about the secrets that nm-applet clearly knows?

[This concerns ubuntu 17.04 but appears to be the case since ubuntu 16.04 at least.]

I have two storage devices in my machine. When I installed ubuntu, it offered to format the SSD and install to it, which was what I wanted. I didn't offer to do anything with the second. Fair enough. I can see that it's there with lshw.

So I can use LUKS to format the second drive and mount it at /home2. On my desktop, this arrangement isn't too annoying, because the machine is up most of the time. But having to do an explicit mount of /home2 isn't great on my laptop, there are just too many boots in its life.

I don't think I have a preference on LUKS vs ecryptfs.

What I'd prefer is to see the second storage device mounted when I login with my password decrypting it, as it does my home directory. I don't see how to do this, though. Any pointers?

Thanks.

In python3 / matplotlib, I'd like a plot with a label:

plt.plot(np.sin(np.linspace(1,10)), label='fête')
plt.legend(loc='upper left')
plt.show()

Works great. Now let's try with xkcd mode:

plt.xkcd()
plt.plot(np.sin(np.linspace(1,10)), label='fête')
plt.legend(loc='upper left')
plt.show()

No error, but the label is 'f?te'. It seems likely that this is a font problem rather than a matplotlib problem, based on reaction to this question on stackoverflow.

That said, I tried writing the word 'fête' in libreoffice with font humor sans, and no UTF-8'ish problems apparent (i.e., accents display correctly).

ubuntu-bug really, really seems to want me to ask this as a question on askbuntu.com rather than reporting it as a bug. So I am.

My machine (for which ubuntu-bug has not attached any details for you, but two salient details are that it's an nv graphics card and 2 x 1920 x 1200 heads) takes roughly a half second to change window focus when I move the mouse to a new window. This is very long, enough that I have to think to wait. A reasonable goal in human interface design would be under a tenth of a second.

It's worth noting that I am using sloppy focus mode (focus follows mouse-enter events).

I'm not quite sure how to phrase this as a question, unless the question is "what's the right way to report this as a bug?" or maybe "any suggestions how to tune compiz, and how can we make it so ubuntu users don't have to tune compiz?".

Thanks.

Suppose I have a spreadsheet in which I track my fruit purchases. It might look like this:

Date        Item    Amount
1-Apr-2012  Apple   .35
2-Apr-2012  Pear    .40
3-Apr-2012  Apple   .36

I want a summary range that looks like this:

Apple     .71
Pear      .40

In other words, I've picked unique "Item" values and summed over them. This seems like something a spreadsheet ought to be able to do, but I've not been able to find it in the gnumeric manual nor googling. (And maybe the answer is the python plugin.)

(I originally tagged this openoffice, as there was no gnumeric tag. I suspect there's a way to do this in ooo iff there's a way in gnumeric. Thanks to the kind soul who created a gnumeric tag and retagged!)

I enabled emacs keybindings as an experiment using gconftool-2 and setting /desktop/gnome/interface/gtk_key_theme to Emacs. I wasn't happy with it and set it back to Default.

jeff@london:~ $ gconftool-2 --get /desktop/gnome/interface/gtk_key_theme
Default
jeff@london:~ $ 

But chrome continues to interpret keybindings as emacs. C-a goes to beginning of line rather than selecting all, C-k kills current line rather than beginning a web search, etc.

I find plenty of references online on how to set emacs bindings, but removing them again is proving harder. Any suggestions on what to look for? (It is possible, of course, that I did something or that something happened beyond the gconftool-2 setting.)