Restoring Windows backup

A relative of mine asked me to restore a couple of files from backup CDs created by the windows native backup utility.
It turned out more difficult than I initially expected.

First, unlike other tools, the windows backup utility zips the files using the backslash as a directory separator and none of the Linux decompression tools can handle that.

Second, the files were encoded with one of the Windows native encodings, and none of the Linux decompression tools handle that (apparently unzip on Ubuntu is patched to add support for filename encodings, but Debian, as usual, is lacking).

What a mess ...

1. unzip with bsdtar as unzip would scramble the non-ascii filenames
$ bsdtar xf ../Backup\ files\ 7.zip
2. convert filenames to utf-8
$ convmv -r -f cp852 -t utf-8 --notest .
3. recreate the directory structure
$ cat y.py
#! /usr/bin/env python

import os
import errno

# already created directories, walk works topdown, so a child dir
# never creates a directory if there is a parent dir with a file.
made_dirs = set()

for root, dir_names, file_names in os.walk('.'):
    for file_name in file_names:
        if '\\' not in file_name:
            continue
        alt_file_name = file_name.replace('\\', '/')
        if alt_file_name.startswith('/'):
            alt_file_name = alt_file_name[1:]  # cut of starting dir separator
        alt_dir_name, alt_base_name = alt_file_name.rsplit('/', 1)
        print('alt_dir', alt_dir_name)
        full_dir_name = os.path.join(root, alt_dir_name)
        if full_dir_name not in made_dirs:
            try:
                os.makedirs(full_dir_name)
            except OSError as exc:
                if exc.errno == errno.EEXIST and os.path.isdir(full_dir_name):
                    # the pass already exists and is a folder, let's just ignore it
                    pass
                else:
                    raise
            made_dirs.add(full_dir_name)
        os.rename(os.path.join(root, file_name),
                  os.path.join(root, alt_file_name))
$ python y.py

Arduino development in Solaris Studio

The unbeatable advantage of the Arduino IDE is that it just works, together with the large library of examples. However it really sucks as a source code editor.
Fortunately there's an Arduino plugin for Netbeans (and therefore for Solaris Studio, which is just thinly veiled Netbeans bundled with the Oracle compilers).

The trouble is the Netbeans plugin URL download button gets you the Windows version, but do not despair. The author apparently plays with Linux too, and there's a tmp/linux folder in the repository, containing a functional Linux version of the plugin.

I only had to alter the default Makefile slightly to adjust for Debian's packaging of the Arduino software and my version of the chip:

8,9c8,9

8,9c8,9
< COM_PORT = /dev/ttyACM0
< BAUD_RATE = 115200
---
> COM_PORT = /dev/ttyUSB0
> BAUD_RATE = 19200
11,12c11,12
< ARDUINO_BASE_DIR = /home/jaques/opt/arduino-1.6.5-r5
< ARDUINO_CORE_DIR = ${ARDUINO_BASE_DIR}/hardware/arduino/avr/cores/arduino
---
> ARDUINO_BASE_DIR = /usr/share/arduino
> ARDUINO_CORE_DIR = ${ARDUINO_BASE_DIR}/hardware/arduino/cores/arduino
20c20
< ARDUINO_MODEL = atmega328p
---
> ARDUINO_MODEL = atmega168
22c22
< ARDUINO_PINS_DIR = ${ARDUINO_BASE_DIR}/hardware/arduino/avr/variants/standard
---
> ARDUINO_PINS_DIR = ${ARDUINO_BASE_DIR}/hardware/arduino/variants/standard
38c38
< AVR_DUDE = ${ARDUINO_BASE_DIR}/hardware/tools/avr/bin/avrdude -C ${ARDUINO_BASE_DIR}/hardware/tools/avr/etc/avrdude.conf
---
> AVR_DUDE = ${ARDUINO_BASE_DIR}/hardware/tools/avrdude -C ${ARDUINO_BASE_DIR}/hardware/tools/avrdude.conf

What can go wrong with Debian upgrade

When I moved from testing to stable I thought it'd be enough to just stop updating testing a while before the release, modify the apt sources and dist-upgrade.
And indeed all seemed to be working well.

Until now.

I wanted to install wine32, but there were all sorts of errors popping up.
I narrowed it down to libxml2 - it wasn't possible to install the 32-bit part:

# apt-get install libxml2:i386
Reading package lists... Done
Building dependency tree      
Reading state information... Done
Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
 apache2-bin : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libabw-0.1-1 : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libaugeas0 : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libe-book-0.1-1 : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libepub0 : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libetonyek-0.1-1 : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libfontforge1 : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libgdraw4 : Depends: libxml2 (>= 2.6.27) but it is not going to be installed
 libgeoclue0 : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libgtk2.0-0 : Depends: shared-mime-info
               Recommends: libgtk2.0-bin
 libgupnp-dlna-2.0-3 : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libmate-window-settings1 : Depends: libxml2 (>= 2.6.27) but it is not going to be installed
 libopenconnect3 : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libplist1 : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libstoken1 : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 libxml-libxml-perl : Depends: libxml2 (>= 2.7.4) but it is not going to be installed
 marco : Depends: libmarco-private0 (= 1.8.2+dfsg1-6) but it is not going to be installed
         Depends: zenity
 mate-media : Depends: mate-media-pulse (>= 1.8.0+dfsg1-3) but it is not going to be installed or
                       mate-media-gstreamer (>= 1.8.0+dfsg1-3) but it is not going to be installed
 mate-settings-daemon : Depends: mate-settings-daemon-pulse (>= 1.8.2-4) but it is not going to be installed or
                                 mate-settings-daemon-gstreamer (>= 1.8.2-4) but it is not going to be installed
 python-libxml2 : Depends: libxml2 (>= 2.9.1) but it is not going to be installed
E: Error, pkgProblemResolver::Resolve generated breaks, this may be caused by held packages.
It took me a while to realize that my libxml2 was actually of higher version than the one in Jessie! As soon as I downgraded it, the installation of wine32 went well...

# dpkg -l |grep libxml2
ii  libxml2:amd64                         2.9.2+dfsg1-1+b1
# apt-get install libxml2=2.9.1+dfsg1-5
Reading package lists... Done
Building dependency tree      
Reading state information... Done
The following packages will be DOWNGRADED:
  libxml2
0 upgraded, 0 newly installed, 1 downgraded, 0 to remove and 0 not upgraded.
Need to get 800 kB of archives.
After this operation, 413 kB disk space will be freed.
Do you want to continue? [Y/n]
Get:1 http://ftp.cz.debian.org/debian/ jessie/main libxml2 amd64 2.9.1+dfsg1-5 [800 kB]
Fetched 800 kB in 0s (2,198 kB/s)
dpkg: warning: downgrading libxml2:amd64 from 2.9.2+dfsg1-1+b1 to 2.9.1+dfsg1-5
(Reading database ... 285067 files and directories currently installed.)
Preparing to unpack .../libxml2_2.9.1+dfsg1-5_amd64.deb ...
Unpacking libxml2:amd64 (2.9.1+dfsg1-5) over (2.9.2+dfsg1-1+b1) ...
Setting up libxml2:amd64 (2.9.1+dfsg1-5) ...
Processing triggers for libc-bin (2.19-18+deb8u1) ...

# apt-get install libxml2:i386
Reading package lists... Done
Building dependency tree      
Reading state information... Done
Recommended packages:
  xml-core:i386
The following NEW packages will be installed:
  libxml2:i386
0 upgraded, 1 newly installed, 0 to remove and 0 not upgraded.
Need to get 833 kB of archives.
After this operation, 1,910 kB of additional disk space will be used.
Get:1 http://ftp.cz.debian.org/debian/ jessie/main libxml2 i386 2.9.1+dfsg1-5 [833 kB]
Fetched 833 kB in 4s (168 kB/s)      
Selecting previously unselected package libxml2:i386.
(Reading database ... 285066 files and directories currently installed.)
Preparing to unpack .../libxml2_2.9.1+dfsg1-5_i386.deb ...
Unpacking libxml2:i386 (2.9.1+dfsg1-5) ...
Setting up libxml2:i386 (2.9.1+dfsg1-5) ...
Processing triggers for libc-bin (2.19-18+deb8u1) ...#

http://blog.hajma.cz/2015/04/fixating-testing-to-jessie.html

Heroes of Might and Magic II

The latest Humble Bundle comes with HOMM2 Gold in the pay-what-you-want group. While it's for Windows only, the installer works in Wine flawlessly.
When it's started I noticed the game actually runs in DosBox, which sounds somewhat silly - to run a game inside a DOS emulator which runs in a Windows emulator :-)
The natural instinct lead me to try to remove the unnecessary levels.
I used lnk-parser to find out what the HOMM2 Desktop link was actually doing:

$ lnk-parse-1.0/lnk-parse-1.0.pl  /home/hajma/.wine/dosdevices/c\:/users/Public/Desktop/Heroes\ of\ Might\ and\ Magic\ 2\ Gold.lnk 

Link File:  /home/hajma/.wine/dosdevices/c:/users/Public/Desktop/Heroes of Might and Magic 2 Gold.lnk
Link Flags:  HAS SHELLIDLIST | POINTS TO FILE/DIR | NO DESCRIPTION | NO RELATIVE PATH STRING | HAS WORKING DIRECTORY | HAS CMD LINE ARGS | HAS CUSTOM ICON |
File Attributes: 
Create Time: Mon Jan 01 1601 00:57:44
Last Accessed time: Mon Jan 01 1601 00:57:44
Last Modified Time: Mon Jan 01 1601 00:57:44
Target Length: 0
Icon Index: 0
ShowWnd: 1 SW_NORMAL
HotKey: 0
Target is on local volume
Volume Type: Fixed (Hard Disk)
Volume Serial: 00000000
Vol Label: 
Base Path: C:\Program Files\Ubisoft\Heroes of Might and Magic 2 Gold\dosbox.exe
(App Path:) Remaining Path: 
Working Dir: C:\Program Files\Ubisoft\Heroes of Might and Magic 2 Gold\
Command Line: -conf dosboxhomm2.conf -noconsole
Icon filename: C:\Program Files\Ubisoft\Heroes of Might and Magic 2 Gold\Icon.ico

but when I run

dosbox -conf ./dosboxhomm2.conf -noconsole

the game doe not run in full version, complaining about missing CD in the drive.
Here's the necessary change to the config file:

$ diff -u dosboxhomm2.conf.orig dosboxhomm2.conf
--- dosboxhomm2.conf.orig 2015-05-01 00:38:24.937021111 +0200
+++ dosboxhomm2.conf 2015-05-01 01:05:30.530285702 +0200
@@ -193,8 +193,8 @@
 
 @ECHO OFF
 mount C "."
-imgmount d ".\homm2.inst" -t iso -fs iso
 c:
+imgmount d "./homm2.inst" -t iso -fs iso
 cls
 heroes2.exe
-exit
\ No newline at end of file
+exit

Fixating testing to Jessie

So, Jessie is released, and I'd like to keep it on my laptop for some time.
It turns out it's reather trivial to do, just make sure all entries in one's sources.list point to jessie instead of testing:

--- /etc/apt/sources.list.bak    2015-04-27 22:23:17.664326851 +0200
+++ /etc/apt/sources.list    2015-04-27 22:27:33.559723639 +0200
@@ -1,10 +1,5 @@
-deb http://ftp.cz.debian.org/debian/ testing main  non-free
-deb-src http://ftp.cz.debian.org/debian/ testing main non-free
-
-deb http://security.debian.org/ testing/updates main non-free
-deb-src http://security.debian.org/ testing/updates main non-free
-
-# testing-updates, previously known as 'volatile'
-deb http://ftp.cz.debian.org/debian/ testing-updates main non-free
-deb-src http://ftp.cz.debian.org/debian/ testing-updates main non-free
+deb http://ftp.cz.debian.org/debian/ jessie main  non-free
+deb-src http://ftp.cz.debian.org/debian/ jessie main non-free

+deb http://security.debian.org/ jessie/updates main non-free
+deb-src http://security.debian.org/ jessie/updates main non-free

and run the usual

apt-get update

apt-get dist-upgrade

Yes, Debian stinks

# apt-get dist-upgrade
...
dpkg: dependency problems prevent processing triggers for hal:
 hal depends on libdbus-glib-1-2 (>= 0.78); however:
  Package libdbus-glib-1-2:amd64 is not configured yet.
 hal depends on pciutils; however:
  Package pciutils is not configured yet.
 hal depends on dbus (>= 0.61); however:
  Package dbus is not configured yet.

dpkg: error processing package hal (--configure):
 dependency problems - leaving triggers unprocessed
dpkg: dependency problems prevent processing triggers for hal:
 hal depends on libdbus-glib-1-2 (>= 0.78); however:
  Package libdbus-glib-1-2:amd64 is not configured yet.
 hal depends on pciutils; however:
  Package pciutils is not configured yet.
 hal depends on dbus (>= 0.61); however:
  Package dbus is not configured yet.
...
dpkg: error processing package hal (--configure):
 dependency problems - leaving triggers unprocessed
dpkg: too many errors, stopping
Processing triggers for libc-bin (2.19-13) ...
Errors were encountered while processing:
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
 hal
Processing was halted because there were too many errors.
E: Sub-process /usr/bin/dpkg returned an error code (1)


Jesus, do you guys really have no QA???

tail -f /dev/ttyUSB0 in Debian

After a recent Debian 7.4 (I hate the codenames!) update I was no longer able to read data from the Arduino's USB serial.
I found out that
cat /dev/ttyUSB0
worked, while
tail -f /dev/ttyUSB0
did nothing. Strangely enough, running the command in strace revealed that tail was reading the data correctly, it just failed to display them ...

I'm not sure whom to blame - coreutils or Debian. Anyway, here's the solution/workaround I found at debianforum.de:

$ stty -F /dev/ttyUSB0 0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0
$ stty -F /dev/ttyUSB0 9600 raw hupcl min 0
$ tail ---disable-inotify -f /dev/ttyUSB0