June 2009 - YaST Development - openSUSE Mailing Lists

[yast-devel] Webyast and permission to call yast dbus backend
by Josef Reidinger 07 Jul '09

07 Jul '09

Hi, I study permissions problem on 11.1 in language module. I found (with mvidner help) root of problem in yast dbus backend call from yast-webservice. I check in webservice if user has rights (in this case root) and then I call to dbus. Problem is that dbus backend check not again root but again caller which is yastws user. So first solution which work is grant yastws all rights in rpm post-script as we do for root. But I think that this is quite big security issue as this mean that anyone who crack into webservice has all rights because he can act as yastws user with all rights to yast backend. This should be somehow solved. MVidner have idea that we could run backend as logged user instead yastws. This has problem that we must somehow handle sending passwords and also multiuser process (another user need maybe another port). My idea is use ssh with keys authentication and execute dbus call via this ssh, so then we can act as logged user and not as yastws. Any other ideas or comments? thanks JR -- To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org For additional commands, e-mail: yast-devel+help(a)opensuse.org

3 6

[yast-devel] WebYaST daily status call 2009-Jun-29
by Stefan Schubert 03 Jul '09

03 Jul '09

WebYaST daily status call 2009-Jun-29 attends: schubi, mschmidkunz, bgeuken, mvidner, jreidinger, jsuchome schubi: Still working on permission clients. There are some problems concerning proxy and writing permissions. bgeuken: patch module: improvement with patch "views" mschmidkunz: concept for the login/logout frame mvidner: no progress concerning DBUS bug. As WebYaST should work on SLES11 we have to take care. Martin will send an own EMAIL. jsuchome: Port user interface to use the DBUS-API jreidinger: testcase for language and timezone modules. Time setting is difficult for "future" time. Contacts mschmidkunz concerning mockups in network module. General: General discussion about permission rights and the user "yastws" with which the webservice is running. We have come to the agreement that we discuss this on the mailing list. Greetings Stefan -- ******************************************************************************* Stefan Schubert SUSE LINUX GmbH - Maxfeldstrasse 5 - D-90409 Nuernberg, Germany e-mail: schubi(a)suse.de ------------------------------------------------------------------------------- SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg) -- To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org For additional commands, e-mail: yast-devel+help(a)opensuse.org

1 4

[yast-devel] how to config bootloader/grub to correctly write menu.lst modifications at kernel upgrades?
by PGNet Dev 30 Jun '09

30 Jun '09

starting with a grub menu.lst, cat /boot/grub/menu.lst default 0 timeout 8 gfxmenu (hd0)/boot/message title PVGRUB root (hd0) kernel /boot/vmlinuz-2.6.27.25-9-xen root=/dev/xvdc resume=/dev/xvdb kbdtype=us text nofb selinux=0 edd=off splash=silent noshell showopts initrd /boot/initrd-2.6.27.25-9-xen after a/any zypper kernel upgrade, menu.lst is modified, e.g., to, cat /boot/grub/menu.lst # Modified by YaST2. Last modification on Thu Jun 25 16:43:29 EDT 2009 default 2 timeout 8 gfxmenu (hd0)/boot/message ###Don't change this comment - YaST2 identifier: Original name: linux### title Xen -- openSUSE 11.1 - 2.6.27.25-10 (xen) root (hd0,0) kernel /vmlinuz-2.6.27.25-10-xen root=/dev/xvdc resume=/dev/xvdb kbdtype=us text nofb selinux=0 edd=off splash=silent noshell showopts initrd /initrd-2.6.27.25-10-xen ###Don't change this comment - YaST2 identifier: Original name: linux### title Xen -- openSUSE 11.1 - 2.6.27.25-10 (xen) kernel (hd0,0)/vmlinuz-2.6.27.25-10-xen root=/dev/xvdc resume=/dev/xvdb kbdtype=us text nofb selinux=0 edd=off splash=silent noshell showopts initrd (hd0,0)/initrd-2.6.27.25-10-xen title PVGRUB root (hd0) kernel /boot/vmlinuz-2.6.27.25-9-xen root=/dev/xvdc resume=/dev/xvdb kbdtype=us text nofb selinux=0 edd=off splash=silent noshell showopts initrd /boot/initrd-2.6.27.25-9-xen what specific config changes do i need to make to get the newly written configs to: (1) correctly use/write, root (hd0) instead of root (hd0,0) (2) write the kernel/initrd entries as kernel /boot/vmlinuz... initrd /boot/initrd... not, kernel /vmlinuz... initrd /initrd... fyi, config files, as default-installed are, /etc/sysconfig/bootloader ## Path: System/Bootloader ## Description: Bootloader configuration ## Type: list(grub,lilo,none) ## Default: grub LOADER_TYPE="grub" ## Path: System/Bootloader ## Description: Bootloader configuration ## Type: list(floppy,mbr,root,boot,custom,none) ## Default: mbr # # Location of boot loader. # For making the change effect run bootloader configuration tool # # LOADER_LOCATION="" /etc/grub setup --stage2=/boot/grub/stage2 (hd0,0) (hd0,0) quit thanks! -- To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org For additional commands, e-mail: yast-devel+help(a)opensuse.org

4 8

[yast-devel] WebYaST: WhichBRaNCH
by Martin Vidner 30 Jun '09

30 Jun '09

Hi, jsuchome has run across this when releasing the users module. When your yast2-webservice-foo needs a change in yast2-foo (usually adding YaPI), remember that in the end it will run on SLE11, *before SP1*. The consequence is that we must make only such modifications as are appropriate for a SLE maintenance update. I figured that using Code11 branch directly is pretty safe if you only add a YaPI file, so if you need to release a maintenance update in the meantime, nothing will break[TM]. In more complex cases, use the new branch that I have just created under tmp: http://svn.opensuse.org/svn/yast/branches/tmp/SLE-11-WebYaST For package versioning, I suggest adding a new dot component on the branch, 2.17.X.*, so that it does not collide with potential maintenance updates on the Code-11 branch. -- Martin Vidner, YaST developer http://en.opensuse.org/User:Mvidner Kuracke oddeleni v restauraci je jako fekalni oddeleni v bazenu -- To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org For additional commands, e-mail: yast-devel+help(a)opensuse.org

1 0

[yast-devel] "rake test" fails in a plugin
by Martin Vidner 30 Jun '09

30 Jun '09

Hi, we've found that this was failing: cd rest-service/plugins/language rake test The solution was to remove the test task from the local Rakefile (rest-service/plugins/language/Rakefile). The intent had been probably to run all tests, but that is done by default anyway (possibly only since we started to refer to the parent rails app?). You know that this is the same problem when rake_test_loader gets a mixture of unit and functional test cases, like here: /usr/bin/ruby -Ilib:lib:test:/home/mvidner/svn/rest-service/webservice "/usr/lib/ruby/gems/1.8/gems/rake-0.8.3/lib/rake/rake_test_loader.rb" "test/unit/language_test.rb" "test/functional/language_controller_test.rb" /home/mvidner/svn/rest-service/webservice/config/environment.rb Loaded suite /usr/lib/ruby/gems/1.8/gems/rake-0.8.3/lib/rake/rake_test_loader Started EE Finished in 0.06988 seconds. 1) Error: test_getter(LanguageTest): NameError: uninitialized constant LanguageTest::Language /test/unit/language_test.rb:33:in `setup' 2) Error: test_setter(LanguageTest): NameError: uninitialized constant LanguageTest::Language /test/unit/language_test.rb:33:in `setup' 2 tests, 0 assertions, 0 failures, 2 errors rake aborted! Command failed with status (1): [/usr/bin/ruby -Ilib:lib:test:/home/mvidner...] (See full trace by running task with --trace) -- Martin Vidner, YaST developer http://en.opensuse.org/User:Mvidner Kuracke oddeleni v restauraci je jako fekalni oddeleni v bazenu -- To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org For additional commands, e-mail: yast-devel+help(a)opensuse.org

1 0

[yast-devel] Test coverage detailed analysis
by Klaus Kaempf 26 Jun '09

26 Jun '09

Hi, running e.g. "rake test:units:rcov" gives you a nice textual overview and coverage percentage on the shell console. However, even more interesting are the generated HTML pages. At the end of the console output a line starting with View the full results at <file:///...> points you to the graphical overview. Clicking on one of the tested files in the index page gives you the detailed analysis, showing - Code reported as executed by Ruby - Lines considered as run by rcov, but not reported by Ruby (e.g comments) - Lines marked as not executed. The last group even has a red background, thus giving you clear hints on where to improve your test cases. Klaus -- To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org For additional commands, e-mail: yast-devel+help(a)opensuse.org

1 0

[yast-devel] WebYaST daily status call 2009-Jun-26
by Klaus Kaempf 26 Jun '09

26 Jun '09

WebYaST daily status call 2009-Jun-26 schubi, mschmidkunz, bgeuken, mvidner, mzugec, kkaempf schubi: jQuery tree widget working fine. Still layout problems. Next: status module mschmidkunz: mockup for module overview (start) Next: refine module layout bgeuken: Cleanup of jQuery code in patches Next: status module mvidner: fix passing empty hash in bindings added bugzilla entry and testcase Next: fix bug mzugec: - bugzilla - Next: more learning kkaempf: Closer look into testing Added rails_rcov for code coverage analysis Got decision on modules for Studio launch Next: Focus on code robustness Everyone: Focus on test cases for existing code before writing new code ! -- To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org For additional commands, e-mail: yast-devel+help(a)opensuse.org

1 0

[yast-devel] Code coverage testing extension in rest-service/webservice
by Klaus Kaempf 26 Jun '09

26 Jun '09

Hi, rest-service/webservice just got a new plugin called 'rails_rcov'. This enables the use of 'rcov' to test code coverage. Code coverage measures the quality and effectiveness of your test cases by checking if all code is actually executed during testing. Installation is pretty simple: gem install relevance-rcov --source http://gems.github.com (this is already packaged in the openSUSE buildservice as rubygem-rcov in project 'ruby' but not building for 11.1 yet) Now run rake test:units:rcov or rake test:functionals:rcov and have a look how good we're doing. Klaus --- SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg) -- To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org For additional commands, e-mail: yast-devel+help(a)opensuse.org

2 1

[yast-devel] git tagging helps you
by Martin Vidner 26 Jun '09

26 Jun '09

Hi, you may have noticed the yellow tags on http://git.opensuse.org/?p=projects/yast/rest-service.git;a=summary They are actually quite more useful than on SVN. HARVEST ------- Here's how it helps me write the changelog cd .../rest-service/plugins/services git shortlog yast2-webservice-services-0.0.1..yast2-webservice-services-0.0.3 . Ladislav Slezak (1): all REST plugins - added .gitignore files Martin Vidner (6): Lsbservice: less clutter, small refactoring. Fixed mocking a bit. Be lazier: /services no longer includes command info. Redirect the other way around Removed %config 0.0.3 Stefan Schubert (1): changed licences That's git shortlog TAG1..TAG2 PATH TAG2 will usually be HEAD, and for TAG1, tab completion works. Cool, huh? PATH limits the log to my subtree of the repository. Use "log" or "diff" instead of "shortlog" to get the obvious variations. SOWING ------ The simplest form creates the tag on HEAD: git tag yast2-webservice-language-0.0.42 For consistency, I recommend the form $RPMNAME-$VERSION It just makes a label for a commit (ls .git/refs/tags/), like on CVS and unlike on SVN. There are also another kinds of tags in git, "tag objects", but I haven't used those. A catch is that pushing normally keeps the tags private. You need to remind git to make them public: git push --tags Once you push a tag out, it is hard to get rid of it if you got it wrong. See "On Re-tagging" in man git-tag and browse our repos for some examples ;-) -- Martin Vidner, YaST developer http://en.opensuse.org/User:Mvidner Kuracke oddeleni v restauraci je jako fekalni oddeleni v bazenu -- To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org For additional commands, e-mail: yast-devel+help(a)opensuse.org

1 0

[yast-devel] caching large values
by Martin Vidner 26 Jun '09

26 Jun '09

Hi, as I mentioned yesterday, I have a problem with caching of data. The data to cache are the commands accepted by individual LSB init scripts. They used to be cached in the session map which is now stored in cookies. Originally the list of scripts was a small faked one but as soon as I switched to the real thing, the cookie started to overflow: ActionController::Session::CookieStore::CookieOverflow As a temporary workaround I turned off the cache. What options are there for caching? -- Martin Vidner, YaST developer http://en.opensuse.org/User:Mvidner Kuracke oddeleni v restauraci je jako fekalni oddeleni v bazenu -- To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org For additional commands, e-mail: yast-devel+help(a)opensuse.org

4 6