[Libreoffice-qa] db test VM

drew drew at baseanswers.com
Tue Nov 1 08:41:32 PDT 2011


On Mon, 2011-10-31 at 10:19 +0100, Petr Mladek wrote:
> Hi Stephan,
> 
> Stephan Bergmann píše v Pá 28. 10. 2011 v 08:28 +0200:
> > (The smoke test is kind of a mixture,  driving soffice via remote UNO
> >  from a C++ test, but the main test  logic is included in a BASIC
> >  script that is run within soffice.   That's more for historical
> >  reasons than a clean design, and I would  not necessarily recommend
> >  modelling new tests after that pattern.)
> 
> Good to know.
> 
> > So, it remains to be seen whether subsequenttests can replace testtool; 
> > migrating the existing testtool tests would definitely not be trivial.
> 
> What is your opinion, please?
> 
> My feeling is that many people do not like testtool because it is:
> 
> 	+ gives random failures:
> 		+ sleeps and waits everywhere in the test code
> 	+ hard to debug:
> 		+ unclear error messaged
> 		+ complex mapping between the elements in the testtool
>                   code and LO code (icons, dialogs)
> 		+ need to run the test several times to understand what
>                   happens there => quite time consuming
> 	+ hard to maintain:
> 		+ menu entries are accessed by the position in menu =>
>                   new item breaks tests
> 		+ ...
> 
> 
> Do you think that the subsequenttests might solve some of the above
> problems?
> 
> IMHO, the best solution would be to involve developers into maintaining
> the tests. We need some technology that they would like and accept.
> 
> What do you think?

Howdy folks,

hmm - likely I should start a new thread, but for now here is good,
perhaps.

On the question of TestTool vs subsequenttests I'm not really qualified
to add much to a discussion, at the moment. (I'll try to fix that over
the next few weeks)

Regarding the VM I've been assembling at SuseStudio:

Published an update yesterday[1].

The size of the download has almost doubled as I've gone ahead and added
a full build environment to the image. (hopefully [2]) To go with this
also doubled the disc image max size for the VM from 16 to 32 gig. (note
this is set to grow the disc as needed, as it stands the VM requires 3.3
gig minimum to install and start)

Added a simple (simplistic) script in the home directory which uses git
to clone a copy of the core source repository into the VM after it is
running at $home/git/libo. Did not set this to auto run as the VM out of
the box already has a working copy of libreoffice at /opt/lo-dev.

Added a web browser to the VM, went with Fireofx 7 and included the SQL
Manager add on[3]. The .xpi file is stored in the $home directory, after
starting the VM just double click on the xpi file to install it.

Debuggers - seemed like a good idea, huh? GDB and Valgrind are included.

Expanded the set of drivers included (pre-installed):
  - ODBC drivers for SQLLite, MySQL and Postgres.
  - JDBC drivers for SQLite, MySQL, Postgres and HSQLdb 2.5
[Note this release does NOT include the actual MySQL and Postgres
servers packages, only client side tools]

For the moment then I'm done adding stuff the image 
- have gone ahead and put together a couple of small test scenarios.

Alright - so next steps as I see them:
Ensure the build environment is correct - without error.
Put together a few scripts that deal with logging..more on that in the
next emails.

and speaking of next emails, I think for this a new thread makes sense.

See you there.

//drew

[1] http://susegallery.com/a/EvUJ20/lo-client-daily-libreoffice_master

[2] using this command "sudo zypper si -d libreoffice-bootstrap" from a
running instance of VM resulted in an error, so ended up adding required
packages by hand.

[3]  https://addons.mozilla.org/en-US/firefox/addon/sqlite-manager





More information about the Libreoffice-qa mailing list