Rustybear BlogFri, 11 Aug 2006
vi based encrypted password manager
I've played around with quite a few different password managers, most of them
being GUI based. They all worked fine, but I always wanted something I could
use easily from the console since I use SSH quite a bit from remote terminals.
The other reason I was never satisifed with a GUI solution was that the program
always had to be running. I know that some of them can sit unobtrusively in
your taskbar, but I'm not much on having lots of things running on my desktop
when I don't need them. On the other hand, I always have an xterm up, so using
the command line for password management seemed a logical approach.
So for a while I simply used vi to edit a text file that I encrypted with
bcrypt. On top of this I wrote a small shell script that would combine the
bcrypt decryption and grep'ing for a password into one comand line operation.
This worked too, but was still lacked something.
Eventually I stumbled on bits and pieces of other people's work that allowed me
to put together an console based solution that I'm now quite pleased with. It
involves the following:
* modications to your .vimrc file specifically for .gpg files
* active use of folding in your password file
* a nice function for your .bashrc config file
* and a powerful little shell script using nawk to get your password info easily
To create you own vi based encrypted password manager, do the following:
1. Create a public/private key pair on the machine where the password file
will reside. To create this key, make sure you have the package xxxxx. Logged
in as yourself, issue the command
ssh-keygen -t rsa
It will prompt you for a passphrase. Enter something you can remember but
is not too obvious. Accept the default (/home/username/.ssh) with regard to
where the keys will be saved.
2. Edit your .vimrc file and include the following:
----------------------------------------
augroup encrypted
au!
" First make sure nothing is written to ~/.viminfo while editing
" an encrypted file.
autocmd BufReadPre,FileReadPre *.gpg set viminfo=
" We don't want a swap file, as it writes unencrypted data to disk
autocmd BufReadPre,FileReadPre *.gpg set noswapfile
" Switch to binary mode to read the encrypted file
autocmd BufReadPre,FileReadPre *.gpg set bin
autocmd BufReadPre,FileReadPre *.gpg let ch_save = &ch|set ch=2
autocmd BufReadPost,FileReadPost *.gpg '[,']!gpg --decrypt 2> /dev/null
" Switch to normal mode for editing
autocmd BufReadPost,FileReadPost *.gpg set nobin
autocmd BufReadPost,FileReadPost *.gpg let &ch = ch_save|unlet ch_save
autocmd BufReadPost,FileReadPost *.gpg execute ":doautocmd BufReadPost " . expand("%:r")
" Convert all text to encrypted text before writing
autocmd BufWritePre,FileWritePre *.gpg '[,']!gpg --default-recipient-self -ae 2>/dev/null
" Undo the encryption so we are back in the normal text, directly
" after the file has been written.
autocmd BufWritePost,FileWritePost *.gpg u
augroup END
----------------------------------------
3. Using vi, create a new file called passwords.gpg.
Set up your entries something like this:
Forums
UserID: myID_1
Password: secret_1
URL: http://www.domain_1.com
More stuff in any format
UserID: myID_2
Password: secret_2
URL: http://www.domain_2.com
More stuff in any format
Memberships
UserID: myID_3
Password: secret_3
URL: http://www.domain_3.com
More stuff in any format
UserID: myID_4
Password: secret_4
URL: http://www.domain_4.com
More stuff in any format
You can experiment with the layout later, but start with this format of
indenting. It's what makes the folding in vi look nice, which we'll set up in
the next step. Save the file and close it.
4. Open the file again with vi. This time, automatically, you should be
prompted for the passphrase you used when you created your key. Enter the
passphrase and you should be in your password.gpg file again.
5. In this new password file, add the following line:
vi: noswapfile bufhidden=wipe tw=0 fdm=indent nobackup nowritebackup foldclose=all
Leave this line in place at all times and never edit it unless you know
what you want to achieve. It sets up vi formatting, specifically the ability to
fold lines. To learn more about folding type :help folding in a vi session
6. Some important folding commands:
zo open fold
zc close fold
zr open all folds
zm close all folds
7. You are now at a point where you're file is encrypted and you should be
able to edit it at will with vi after entering your passphrase. The next step
will give you quick command line access to query your password file.
8. Somewhere in your path (echo $PATH), create a new shell script called searchWord.sh. Cut and paste the following:
---------------------------------
#!/bin/sh
usage()
{
[ $# -ne 0 ] && {
echo "$0: $@" 1>&2
}
echo "usage: $0 pattern [file ...]" 1>&2
exit 1
}
if [ $# -lt 1 ] ;then
usage "missing pattern"
fi
pattern=$1
shift
nawk '
BEGIN {IGNORECASE=1}
NF==0 {
if (paragraph ~ "'$pattern'") {
print paragraph
}
paragraph=""
}
NF != 0 {
paragraph=paragraph $0 "\n"
}' "$@"
---------------------------------
Save it and chmod 755 the file.
9. In your .bashrc file add the following:
---------------------------------
### used for password program
function qpass () {
/usr/bin/gpg -d < ~/docs/passwords.gpg | ~/bin/searchWord.sh "$@"
}
---------------------------------
10. That's pretty much it. Test it out at the command line by typing a query, like ...
$> qpass domain.com
You should be prompted for your passphrase. Enter it, and the
searchWord.sh script will search your password file for the term
"domain.com". It will return multiple finds if such exist.
I hope this makes sense and works as well for others as it has for me.
Credits: Thanks to Serge Roux and Wouter Hanegraaff for portions of the code.
Posted at: 09:06 | category: /vi | Comments ()
Thu, 10 Aug 2006
Using sqlobject in cherryPy
from sqlobject import *
conn = 'mysql://dbuser:dbpass@host/dbname'
class NameOfTable(SQLObject):
_connection = conn
_fromDatabase = True
_style = MixedCaseStyle(longID=False)
class Newspage(Page):
def index(self):
article = NameOfTable.select(NameOfTable.q.field=='Joe', orderBy=-NameOfTable.q.dateAdded)
Notes:
_fromDatabase = True means sqlobject will simply take column names from
the DB. It won't create any, which it is capable of doing.
MixedCaseStyle - sqlobject uses pythonic names like my_name. This allows
the use of names like myName.
longID=False - prevents sqlobject from automatically creating an ID column
- in front of NameOfTable.q.dateAdded means reverse sort
Posted at: 21:38 | category: /database | Comments ()
Install Debian Source Packages
Debian can be a bit tricky if you're not familiar with its method of building
source packages. Although this is covered in their faq somewhere, here is the
information.
apt-get build-dep fluxbox
apt-get source fluxbox
apt-get install fakeroot
apt-get install dpkg-dev
apt-get install automake
(You may already have fakeroot and dpkg-dev installed). Although apt-get
build-dep fluxbox should also install all necessary tools, I have found that I
also have to get automake.
This will bring you three files (version number correct as of January 2004)
fluxbox_0.1.14-2.orig.tar.gz, fluxbox_0.1.14-2.diff.gz and
fluxbox_0.1.14-2.dsc. You will also have a fluxbox-0.1.14 directory.
The next step is
dpkg-source -x fluxbox_0.1.14-2.dsc
I'm assuming you have wget or something similar, if not do a quick
apt-get install wget
Now, we patch
cd fluxbox-0.1.14/src
wget kitaj.no-ip.com/files/fluxbox/fluxbox-0.1.14-menukey.patch
patch < fluxbox-0.1.14-menukey.patch
Now to build and install the package. You are in the fluxbox-0.1.14/src directory
cd ../
dpkg-buildpackage -rfakeroot -b
This will take some time. When it's done
dpkg -i ../fluxbox_0.1.14-2_i386.deb
Posted at: 21:36 | category: /configure | Comments ()
Compile a kernel the Debian way
apt-get install gcc kernel-package libc6-dev tk8.3 libncurses5-dev fakeroot
adduser kevin src
cd /usr/src
apt-get source kernel-soource-2.6.9
[dpkg will automatically unpack it for you]
ln -s kernel-source-2.6.9_2.6.9 linux
cd /usr/src/linux
rm -rf /debian
cd /usr/src/linux/scripts
./ver_linux
[to see if you've got everything you need]
cd /usr/src/linux
cp /boot/config-2.4.23 .config
make menuconfig
[do your config selection stuff]
make-kpkg clean
make-kpkg --revision $(date +'%Y%m%d') --append-to-version $(hostname) --rootcmd fakeroot clean
make-kpkg --revision $(date +'%Y%m%d') --append-to-version $(hostname) --rootcmd fakeroot kernel_image modules_image;
cd /usr/src
ls -l
...
kernel-image-2.6.9chilkoot_20041220_i386.deb
...
dpkg -i kernel-image-2.6.9chilkoot_20041220_i386.deb
[ if using initrd do the following ]
cd /boot
mkinitrd -o initrd.img-2.6.9chilkoot 2.6.9chilkoot
update-grub
check /boot/grub/menu.lst for correct entries of initrd and vmlinuz
reboot
Posted at: 21:36 | category: /configure | Comments ()
Create a new SSL certificate for Apache
Go to /etc/apache-ssl and backup apache.pem before you do anything else. The
location of this file may be different depending upon your Apache setup.
Install the Debian package ssl-cert.
Check to make sure you have the standard template file:
/usr/share/ssl-cert/ssleay.cnf
If you don't, make sure you have openssl installed.
cd over to a temporary working directory that you have write access to.
Run the command:
make-ssl-cert /usr/share/ssl-cert/ssleay.cnf privkey.pem
This will create a private key for the server named privkey.pem.
When complete with this process, store this file off site.
Next run the command:
openssl req -new -x509 -key privkey.pem -out cacert.pem -days 1095
This will create a server certificate called cacert.pem, good for 3 years.
Next run the command:
cat privkey.pem cacert.pem > apache.pem
This will create the complete certificate needed for apache. Copy it over to
the directory where the pre-existing apache.pem was that you backed up
eariler.
Don't forget to move the privkey.pem off site.
Posted at: 21:36 | category: /configure | Comments ()
To set up CUPS based printing
apt-get install hplip hpijs
This will install quite a few other things as well, including CUPS
Then go to:
http://localhost:631
and add a printer.
Present configurations include:
HP_Laserjet_2100TN
Device: AppSocket/HP JetDirect
Device URI: socket://10.10.10.99:9100
Make: HP (HPLIP)
Model: HP Laserjet 2100 Foomatic/hpijs
or
HP Laserjet 2100 Series Postscript
HP_Photosmart_7960
Device: Internet Printing Protocal (http)
Device URL: http://10.10.10.98:631/ipp/lpt2
Make: HP (HPLIP)
Model: HP Photosmart 7960 Foomatic/hpijs
Check by going to Printers/Manage Printers/Print Test Page.
Posted at: 21:36 | category: /configure | Comments ()
Configure xterm
edit ~/.Xdefaults as appropriate look at /etc/X11/app-defaults/XTerm for ideas about what and how to change run 'xrdb -merge .Xdefaults' after editing to implement the changes
Posted at: 21:36 | category: /configure | Comments ()
Settings for uploads using PHP/Apache
In Apache httpd.conf, add the following:..... other directives ..... php_value post_max_size "71M" php_value upload_max_filesize "70M" php_value memory_limit "70M"
Posted at: 21:36 | category: /configure | Comments ()
Mailman using Postfix on Debian
The following is a description on how to set up the list manager Mailman
on a Debian system that uses Postfix-MySQL-Apache for virtual websites and
email. It uses a Python script that makes it unnecessary to make entries
into /etc/aliases. But for this Python script to work, you must be
willing to prefix the FQDN portion of your listname with "lists"...
correct: mylist@lists.virtdomain1.org
incorrect: mylist@virtdomain1.org
While this may seem a sacrifice, it can easily be overcome later by simply
creating an alias in your MySQL table from mylist@virtdomain1.org to
mylist@list.virtdomain1.org. This is far easier than adding in all of the
Mailman aliases that would normally have to be added to /etc/aliases if
you were not using the Python script.
1. Make sure your virtual websites and virtual email setups are all
working before doing anything with Mailman.
2. Install Mailman via apt-get using normal defaults for the base website
that runs on the server.
3. Edit /etc/mailman/mm_cfg.py
add:
VIRTUAL_HOSTS.clear()
add_virtualhost('virtdomain1.org','lists.virtdomain1.org')
add_virtualhost('virtdomain2.org','lists.virtdomain2.org')
change DEFAULT_EMAIL_HOST to include the prefix 'lists' to the FQDN
as in:
DEFAULT_EMAIL_HOST = 'lists.rustybear.com'
change IMAGE_LOGOS from whatever default is to:
IMAGE_LOGOS = /'icons'/
4. copy Mailman images over to default /icons/ directory for Apache
cp /usr/share/images/mailman/* /usr/share/apache/icons/.
5. Edit /etc/apache/httpd.conf
for any you have configured, add in:
ScriptAlias /mailman/ /usr/lib/cgi-bin/mailman/
in the section.
and in an appropriate place (not within a VirtualHost directive) in
httpd.conf add in:
Options +FollowSymLinks
you'll need the above to get pipermail (archives) working.
6. Get this python script:
If you are installing via Debian apt-get this section is probably
mostly done for you, but check anyway.
http://www.gurulabs.com/files/postfix-to-mailman-2.1.py
and drop it in /etc/mailman/ and rename it without the version number
mv postfix-to-mailman-x.xx.py postfix-to-mailman.py
edit it by changing the entry for MailmanHome
MailmanHome = "/var/lib/mailman";
the Mailman home directory is where the lists are kept.
Then cd to /usr/lib/mailman/bin and symlink back to
postfix-to-mailman.py
cd /usr/lib/mailman/bin
ln -s /etc/mailman/postfix-to-mailman.py
7. Next go to /etc/mailman/postfix-to-mailman.py and read all of the
comments/instructions.
Note that in the recommended changes to /etc/postfix/master.cf, that
you must set the uid:gid as list:list, not as mailman:mailman.
This is probably specific to Debian only as the deb package has
Mailman using 'list' as UID/GID where source versions of Mailman
probably have 'mailman' as a default UID.
Note also that in the changes to master.cf, you need to change the
argv argument to point to /etc/mailman/postfix-to-mailman.py.
This is also a good time to:
chmod 755 /etc/mailman/postfix-to-mailman.py
chown root:list /etc/mailman/postfix-to-mailman.py
The instructions in postfix-to-mailman.py point out how to set up
/etc/postfix/tranport and mail.cf. You can do it that way, or more
simply, just make an entry in your existing transport setup for MySQL.
In my case I have a table called transport, where normally I put a
domain name (i.e. xyz.com) and transport type (usually either virtual:
or local: ). However, in this case, the entry I make is:
domain: lists.virtdomain1.org
transport: mailman:
Alternatively, you could follow the instructions and make the changes
to main.cf and transport as that works too. If you do it that way,
don't forget to execute
postmap /etc/postfix/transport
after making changes to transport to regenerate its data.
8. Make sure you have a DNS entry in your DNS server for
lists.virtdomain1.org, or at least a wildcard entry, that resolves to
virtdomain1.org.
9. Restart Apache and Mailman (/etc/init.d/mailman restart).
10. To create a new list for one of the virtual websites, at the prompt
execute:
newlist myNewList@lists.virtdomain1.org
It will prompt you for a couple things that are obvious. Make sure
you include the 'lists' prefix before the FQDN, as having 'lists'
there is what makes this whole thing work via the python script we put
in -- /etc/mailman/postfix-to-mailman.py.
You should get an email regarding the newly created list you
just made.
Note also that simply executing 'newlist --help' gives some help too.
Also, reading the docs at /usr/share/docs/mailman is a good idea too.
You should now be able to see your admin page at:
http://lists.virtdomain1.org/cgi-bin/mailman/admin/myNewList
11. Thing should be working and you should be able to subscribe new users
and send emails to be posted, etc. Try sending test posts and
watching /var/log/syslog or wherever you have you email logs set up
and watching for errors as the test posts go through Postfix. It's
best to watch this live using 'tail -f /var/log/syslog'.
12. Good luck!
Posted at: 21:36 | category: /configure | Comments ()
samba config
# Global parameters
[global]
workgroup = HAWKWOOD
server string = %h server (Samba %v)
obey pam restrictions = Yes
passdb backend = tdbsam, guest
passwd program = /usr/bin/passwd %u
passwd chat = *Enter\snew\sUNIX\spassword:* %n\n *Retype\snew\sUNIX\spassword:* %n\n .
log level = 5
syslog = 0
log file = /var/log/samba/log.%m
max log size = 1000
socket options = TCP_NODELAY SO_RCVBUF=8192 SO_SNDBUF=819
domain logons = Yes
os level = 65
domain master = Yes
dns proxy = No
panic action = /usr/share/samba/panic-action %d
printing = lprng
printer name = HP_Laserjet
[homes]
comment = Home Directories
read only = No
create mask = 0775
directory mask = 0775
guest ok = Yes
[printers]
comment = All Printers
path = /tmp
create mask = 0700
printable = Yes
browseable = No
[cdrom]
comment = Samba sakura's CD-ROM
path = /cdrom
guest ok = Yes
locking = No
exec = /bin/mount /cdrom
postexec = /bin/umount /cdrom
[kevin]
path = /home/kevin
write list = kosuke,akiko,kevin
read only = No
create mask = 0777
directory mask = 0777
[lp]
comment = HP_Laserjet
path = /usr/bin/lp
read only = No
create mask = 0700
guest ok = Yes
printable = Yes
printing = bsd
printer name = lp
share modes = No
Posted at: 21:36 | category: /configure | Comments ()
cygwin + ssh
Download and install cygwin to c:\cygwin. Select the packages that make sense, but make sure to include cygrunsrv and openssh packages. Right click 'My Computer' and under Environment Variables of the Advanced tab, in the section System variables, add 'CYGWIN' for variable, and 'ntsec tty' for value. Also add on to the end of the PATH variable the following: 'c:\cygwin\bin'. Edit c:\cygwin\cygwin.bat. Make sure it contains these lines - you will need to add the line setting the CYGWIN environment variable. @echo off set CYGWIN=binmode tty ntsec C: chdir \cygwin\bin bash --login -i Test cygwin to make sure it works. ls /bin // to see the cygwin bin directory dir c: // to see the contents of the C: directory Type "control d" or 'logout' to exit the shell. If you get a message saying 'cannot create /home/userid', run this command from the cygwin window "mkpasswd -l >/etc/passwd". Run ssh-host-config to set up the ssh host keys and create the sshd_config file in /etc/. You should see output like this: $ ssh-host-config Generating /etc/ssh_host_key Generating /etc/ssh_host_rsa_key Generating /etc/ssh_host_dsa_key Generating /etc/ssh_config file Generating /etc/sshd_config file Added ssh to /cygdrive/c/WINNT/system32/drivers/etc/services Do you want to install sshd as service? (Say "no" if it's already installed as service) (yes/no) Answer 'yes' to the prompt. Press 'Return' to accept the default at the CYGWIN environment question (default = binmode tty ntsec). The service name is CYGWIN sshd. Type 'cd' to go to your account's home directory. Run ssh-user-config to setup your ssh keys. Create only an SSH2 RSA identity (use a null passphrase - just press return). Output should be similar to this : $ ssh-user-config Shall I create an SSH1 RSA identity file for you? (yes/no) no Shall I create an SSH2 RSA identity file for you? (yes/no) (yes/no) yes Generating /home/pswander/.ssh/id_rsa Enter passphrase (empty for no passphrase): Enter same passphrase again: Do you want to use this identity to login to this machine? (yes/no) yes Shall I create an SSH2 DSA identity file for you? (yes/no) (yes/no) no You will probably want to change permissions and ownership on some files: cd /; chmod -R og-w . // note the period chmod og+w /tmp touch /var/log/sshd.log // may already be there chown SYSTEM:SYSTEM /var/log/sshd.log chown SYSTEM:SYSTEM /var/empty chown SYSTEM:SYSTEM /etc/ssh_h* chmod 755 /var/empty At the prompt type the following to install sshd as a service: cygrunsrv -I sshd -p /usr/sbin/sshd -a '-D' -d 'CYGWIN sshd' And then the following to start the service: cygrunsrv -S sshd And then test it with: ssh localhost And then test it from another machine. And if necessary, the following to remove the service cygrunsrv -R sshd There is also more documentation at /usr/doc/Cygwin.
Posted at: 21:36 | category: /configure | Comments ()
Mounting USB devices
lsmod to see what is loaded. You potentially need: usb-storage usbcore scsi_mod sd_mod usb-uhci input sr_mod If using Debian, apt-get install sg3-utils and then use: sg_scan -i // to see scsi type devices sg_map // to see device associations Mount the drive found (should be a SCSI drive): mount -t vfat /dev/sdc1 /home/kevin/mnt/usbdevice Alternatively, make an /etc/fstab entry: /dev/sdc1 /home/kevin/mnt/usbdevice vfat noauto,user 0 0 Diagnose by checking: /var/log/messages /var/log/syslog /proc/bus/usb // shows what's been seen /proc/scsi/scsi // to see what scsi devices are there Notes: Like any other hard drive, you can use other commands on your mounted USB storage device: fdisk /dev/sdc1 // play with partitions mkfs -t vfat /dev/sdc1 // format the partition(s) in FAT mkfs -t ext3 /dev/sdc1 // format in Linux ext3 but not a good // idea for digital camera disks
Posted at: 21:36 | category: /configure | Comments ()
Nagios configuration
To add a host:
1. add the new host to hosts.cfg
2. add the new host to a group in hostgroups.cfg
3. make sure the host group has a contact group
4. make sure contacts are separate from the site
being checked
5. check contactgroups.cfg
6. make selections in services.cfg
7. add the new host to dependencies.cfg if it
has any dependencies
To check config settings run
nagios -v /etc/nagios/nagios.cfg
Posted at: 21:36 | category: /configure | Comments ()
Remap that CapsLock key to Control
Hate the Windows keyboard? Like the Unix keyboard? Then remap that CapsLock key. If using Xfree86 add the following to the InputDevice section of /etc/X11/XF86Config-4 Option "XkbOptions" "ctrl:nocaps" Also note that Ctrl-[ is the same as hitting the Esc key.
Posted at: 21:28 | category: /vi | Comments ()
Embedding vim option settings in a file
You can use modelines to add Vim option settings to the contents of a file.
For example, in a C file, you can add the following line to the top or the
bottom of the file:
/* vim: set textwidth=76 nopaste: */
This will set the 'textwidth' option to 76 and not allow pasting, when editing
that C file. For this to work, the 'modeline' option should be set. By
default, the 'modeline' option is set.
Posted at: 21:28 | category: /vi | Comments ()
Create a thumbnail tack sheet from a directory of pictures
convert 'vid:*.jpg' thumbnail.png
Posted at: 21:28 | category: /graphics | Comments ()
convert
convert -size 600x480 orig-pic.jpg -resize 600x480 new-pic.jpg
Posted at: 21:28 | category: /graphics | Comments ()
qiv
qiv -fm somepicture.jpg // view a single picture
// -f fullscreen on/off
// -m scale to screen on/off
qiv -fmsd 3 *.jpg // slideshow
// -s slideshow on/off
// -d 3 three second delay
// while viewing, hitting 'd' will send the pic
// to the .qiv-trash directory
Posted at: 21:28 | category: /graphics | Comments ()
import
// to take a screenshot using imagemagick import myfilename.jpg // then with cursor either click a window or select a region
Posted at: 21:28 | category: /graphics | Comments ()
mogrify
mogrify -format xpm -geometry 32x32 -map /usr/share/pixmaps/cmap.xpm fileToConvert // creates a standardized icon from a pre-existing image ** To create a directory of email-able pics or thumbnails: cd /path/to/picfolder cd .. // back up one level cp -R picfolder smallpicfolder // create a duplicate folder cd smallpicfolder mogrify -size 800x600 -resize 800x600 *.jpg // mogrify to email size Note: the mogrify operation will overwrite the existing imgages, hence it is important to make a duplicate folder with all of the images.
Posted at: 21:28 | category: /graphics | Comments ()
Cartoonize A Photo With Gimp
Layer->Duplicate Layer Select the new layer, the Filters->Edge-Detect->Edge, select Sobel, Amount should be 2.0, and Black should be checked. Invert the edge-mask layer ... it's in Layers->Colors->Invert Back in the Layers Dialog, change the Layer Mode to Divide Play around with Layer mode settings. Desaturate the top layer, blue or otherwise mess up the bottom layer.
Posted at: 21:28 | category: /gimp | Comments ()
Add different fonts to the Gimp
Gimp uses fonts from the fontconfig program. So simply put the new fonts in your ~/.fonts directory and then run, as root, dpkg-reconfigure fontconfig.
Posted at: 21:28 | category: /gimp | Comments ()
Mouse double-click in Apple Gimp
If you find that you need to click a window twice to activate it when
moving from window to window in Apple Gimp, then open a terminal and
enter the following:
defaults write com.apple.x11 wm_ffm true
This will enable "focus follows mouse. X11 must be restarted for the
change to take effect.
Posted at: 21:28 | category: /gimp | Comments ()
Upgrade a Debian Package with new source using svn-upgrade
download a copy of the new source tarball and place it in /debian-dev/tarballs
cd /debian-dev/calcurse (just above /debian. this is the WC [working copy])
svn-upgrade ../tarballs/calcurse-1.2.tar.gz
You'll now have a copy of calcurse in /debian-dev/calcurse that is updated to
the latest source.
You can check changes with
svn diff
or specific files with
svn diff TODO
Check the status of the svn repository for calcurse with
svn status
Go ahead and make changes to things like the debian/changelog file. Once
you've made all your changes you can commit the changes to the svn repository
with
svn commit
You can also do this by specific file
svn commit src/util.c
Once everything is committed, you can build the package with
svn-buildpackage -k8CE11941 -rfakeroot
which is executed in the WC (just above /debian).
The packages that are built will show up in /debian-dev/build-area. You
should linda and lintian check them.
You should also run pbuilder on the WC. Start this process by executing
sudo pbuilder update
sudo pbuilder clean
Then, from the WC, run
pdebuild
You can then go to /var/cache/pbuilder/results to linda/lintian check the
files that were created by pdebuild.
Once you are satisfied that everything is o.k. and have tested it by
installing, upgrading, etc, then from the WC, run
svn-buildpackage -k8CE11941 -rfakeroot --svn-tag
or
svn-buildpackage -k8CE11941 -rfakeroot --svn-tag-only
This will prepare for the next source release by making an entry in
debian/changelog.
Send the following files to the uploader:
calcurse_1.2-1.diff.gz
calcurse_1.2-1.dsc
calcurse_1.2-1_i386.changes
calcurse_1.2.orig.tar.gz
That's it.
Posted at: 21:27 | category: /dev | Comments ()
Working with svn-buildpackage
Start by creating a svn repository. I'm presuming you'll be working from your
home directory --
svnadmin create svn-repo
... where svn-repo can be any name you want.
Note that this repository [svn-repo or whatever you name it] will be where
svn manages all of its work. It is not an area that you will edit.
Next, create a directory where you will do your work. For me it was --
mkdir debian-dev
cd debian-dev
Download the appropriate files: package.dsc, package.orig.tar.gz, and
package.diff.gz
Now we'll inject an existing source file into the repository --
svn-inject bbtime_0.1.5-9.dsc file:///home/kevin/svn-repo
You'll see a lot of stuff scroll by, the last line of which should say
something like
"Your working directory is /home/kevin/debian-dev/bbtime - have fun!"
The svn-inject command actually makes your working directory for you.
For most purposes you'll want to work out of the directory, although you can
create other working directories if you like with the command --
svn co file:///home/kevin/svn-repo/bbtime
Wherever you execute this command is where the working directory will be
built. When it is built, you'll see the following tree built:
bbtime
|----- branches
|----- tags
|----- trunk
You'll want to descend into trunk, where you'll see all of the source files
and a debian/ directory with all of the rules/control files. Later, when you
execute the svn-buildpackage command, you'll want to do so from this trunk
directory.
Return back to /home/kevin/debian-dev (or your equivalent). Make a directory
in here as follows --
mkdir tarballs
... so you should now have:
home/kevin/debian-dev
|------------------- bbtime
|------------------- tarballs
Into the tarballs directory copy the original source tarballs, which should
have a name like bbtime_0.1.5.orig.tar.gz. You should have been able to
download a copy of this file from the same place you downloaded the .dsc file
(most likely the Package area on Debian.org).
At this point you are ready to do some work. Remember that svn-inject did
the original checking out of the bbtime working files for you, so you can
simply go to them and start your editing --
cd /home/kevin/debian-dev/bbtime
... and do your work, most of which will probably be in the debian/ directory.
Once you've done your editing, you'll need to execute various svn commands,
all of which should be executed from the working directory, which is
/home/kevin/debian-dev/bbtime. Examples of some of these commands:
svn add debian/newfile
svn delete debian/oldfile
svn status
svn log
svn diff
svn commit
This last command - svn commit - must be executed before you can do any
building. As you execute it, you'll be given a chance to write notes about
the changes you made.
The next step is the actual building. The command is --
svn-buildpackage -k8CE11941 -rfakeroot
where -k is your GnuPG key ID and -r sets up fakeroot, which you should have
installed.
You'll see plenty of screen action, which should end with a few lines that
tell you where everything was built. If you move up a level in the directory,
you'll now find a new sub-directory called build-area. This is where all of
the built files are put. So you should now have a tree like this:
home/kevin/debian-dev
|------------------- bbtime
|--------------------build-area
|------------------- tarballs
In the build-area you'll find all of your newly generated files:
bbtime_0.8.3-3.diff.gz
bbtime_0.8.3-3.dsc
bbtime_0.8.3-3_i386.changes
bbtime_0.8.3-3_i386.deb
bbtime_0.8.3.orig.tar.gz
Obviously you'll want to test everything, including running lintian and linda.
You'll probably want to build a few times as you work out your bugs. Each
time you change something in bbtime/debian, you'll have to svn commit it to
the repository before you can execute svn-buildpackage again. And each time
you build, svn-buildpackage will automatically overwrite the files in the
build-area directory, so don't worrying about cleaning that out.
When you are finally happy with your results, you need to tag it as a final
build. You can do this with one of two commands --
svn-buildpackage -k8CE11941 -rfakeroot --svn-tag
or
svn-buildpackage -k8CE11941 -rfakeroot --svn-tag-only
The first command will build the package and tag it, while the latter command
will just tag it (presumes your earlier builds were satisfactory).
When you tag it, you'll also be making a new debian/changelog entry, which
essentially is preparation for the next time you rebuild the package and
release a new version.
At this point you are basically done. Make sure you don't delete the working
copy of your package that you have in /home/kevin/debian-dev. Leave it there
for future versions.
Before ending this, here's a couple more things to think about that may hang
you up.
First, let's say you've just finished building, are happy with the package,
and you've tagged it, which puts a copy into the repository trunk. You
send it to your upload sponsor, but he spots a bug and wants you to correct
it. But you've already finalized and tagged your build. No problem. Go to
your working copy - /home/kevin/debian-dev/bbtime - and execute the command --
svn update
This will refresh your working copy from the repository. It probably already
is the latest copy, but just in case, update it anyway. Next go into
debian/changelog and remove the next version entry that the tagging process
created for you. Make any other bug corrections. And then execute your svn
commit and then you can build again and eventually tag it. Your versioning
will still be preserved.
The second hang up that may get you is if you checkout a copy of a package
into a different working area. Say for instance that you are in /tmp and you
execute --
svn co file:///home/kevin/svn-repo/bbtime
This will create the following tree for you:
/tmp/bbtime
|------- branches
|------- tags
|------- trunk
You can descend into trunk, where you'll find all the source files and the
debian/ directory. You can do all the editing you need and commit all of the
changes back to the repository. And you can even build a deb by executing the
svn-buildpackage command from the trunk/ directory. It will create a
build-area directory for like this:
/tmp/bbtime
|------- branches
|------- build-area
|------- tags
|------- trunk
... and in this build-area directory you'll find the results of your build.
What you won't find is a .diff.gz file. That file is only generated if you
are working from the originally created working directory from when you
executed svn-inject.
If you, or someone else, works on a copy of the package checked out from the
repository to a working area not in the original /home/kevin/debian-dev and
commits changes, then make sure next time you work in the original work area -
/home/kevin/debian-dev/bbtime - that you execute svn update to incorporate
their changes into your working copy. This is basic svn stuff but may be
helpful to those new to all of this.
Good luck!
Posted at: 21:27 | category: /dev | Comments ()
Subversion: new project/repository configuration and some basic commands
As root:
svn mkdir /var/svn/repo/newproject
chown -R myUser:svnusers /var/svn/repo/newproject
(presumes you have a svnusers group)
As a user/owner (myUser) of the newproject:
svnadmin create /var/svn/repo/newproject
As root:
chown -R myUser:svnusers /var/svn/repo/newproject
chmod 664 /var/svn/repo/newproject/db/*
As a user/owner (myUser) of the newproject:
cd /tmp
mkdir -p newproject/trunk
mkdir -p newproject/branches
mkdir -p newproject/tags
cd /tmp/newproject/trunk
cp -pvr /home/newproject/www .
svn import /tmp/newproject file:///var/svn/repo/newproject -m "initial import"
cd /tmp
rm -rf newproject/
To verify results of the import:
svn list --verbose file:///var/svn/repo/newproject
To export the project without .svn files
cd [to target export directory]
svn export file:///var/svn/repo/newproject/trunk/www [option name]
To checkout the project with .svn files
cd [to target working copy directory]
svn checkout file:///var/svn/repo/newproject/trunk/www [option name]
svn status [file:///var/svn/repo/newproject]
prints the status of working copy files and directories
svn diff
shows changes to the working copy
svn update [file:///var/svn/repo/newproject]
brings changes from the repository to your working copy
A == added
D == deleted
U == updated
C == conflict
G == merged
svn add PATH
adds files/directories to your working copy and schedules them for
addition to the repository. will be uploaded and added to the repository
during the next commit. when adding a directory, the default behavior is
to recurse
svn delete PATH
schedules PATH for deletion on the next commit. if PATH is a URL, it will
immediately delete and a log message must be supplied.
svn copy SRC DST
copy a file/directory in your working copy or repository.
svn move SRC DST
move a file/directory in your working copy or repository. equivalent to
svn copy/svn delete.
example: svn move foo.c bar.c
svn commit [--message] [PATH]
sends changes from working copy to the repository. alias == ci.
example: svn commit -F message foo.c
this only commits the foo.c file and a message
svn commit without any args will commit everything
svn log [PATH]
display all of the commit messages
Posted at: 21:27 | category: /dev | Comments ()
qmake - compiling
To compile a C++ application you need to create a makefile. The easiest way to create a makefile is to use the qmake build tool supplied with Qt. If you've saved main.cpp in its own directory, all you have to do is: qmake -project -o main.pro // this creates a project file main.pro qmake // this creates the Makefile make // this creates the executable The first command tells qmake to create a .pro (project) file. The second command tells it to create a (platform-specific) makefile based on the project file. You should now be able to type make (or nmake if you're using Visual Studio).
Posted at: 21:27 | category: /dev | Comments ()
Important pbuilder commands
sudo pbuilder create --distribution sarge
sudo pbuilder update --override-config --distribution sid
pdebuild
// run this in the directory with source files (above debian/)
// it builds a deb in /var/cache/pbuilder/results
note that presently pbuilder (sid) uses gcc-4.0 and g++-4.0, which may not be
appropriate for your package.
sudo pbuilder clean
------
piuparts - to check for proper dependencies and whether package install
properly in unstable after completing the build process
cd to the source build directory (above debian/)
piuparts ../bbmail-0.8.3_i386.deb
------
Building the final package:
// check which complier is default
cd to the source build directory (above debian/)
dpkg-buildpackage -rfakeroot -k8CE11941
Posted at: 21:27 | category: /dev | Comments ()
ezcounter.php -- counter for webpages
The above page should be added to the target webpage via an include. The include needs to be before any echo/print statements as headers cannot be sent twice. Print out the counter value later on the page with:
Posted at: 21:27 | category: /dev | Comments ()
