Parsing model enhancements

[GoogleCodeExporter]

This issue keeps track of changes to parsed model that did not make 2.5,but 
would be useful for RIDE or other tools.

1) Support copy of test case/user keyword

2) Support rename of test case/user keyword

Original issue reported on code.google.com by janne.t....@gmail.com on 15 Jun 2010 at 6:40

[GoogleCodeExporter]

3) fix as_list implementation for settings to handle empty values correctly 
(empty values at the end of list should be left out). Now these methods have 
been monkey patched in RIDE.

Original comment by jussi.ao...@gmail.com on 15 Jun 2010 at 1:26

[GoogleCodeExporter]

1) Should be implemented so that it is possible, but not mandatory, to give new 
name for the copied tc/uk.

2) Discussed with Janne that this isn't actually needed. Anyone who needs to 
change the name can simply reset the name attribute.

3) Was done in r3780 (and implementation slightly cleaned up in later 
revisions).

Original comment by pekka.klarck on 16 Jun 2010 at 8:56

• Added labels: Priority-High, Target-2.5.1

[GoogleCodeExporter]

4) It should be possible to create TestData objects in memory with non-existing 
path as source. Currently this does not succeed because the model object tries 
to automatically populate itself if source is given, and instance creation 
fails if the source does not exist.

Original comment by janne.t....@gmail.com on 29 Jun 2010 at 5:32

[GoogleCodeExporter]

5) Comments in multiple cells should NOT be catenated with pipes (single space 
would be better)

6) Comment values should be stripped when read.

Original comment by janne.t....@gmail.com on 30 Jun 2010 at 7:58

[GoogleCodeExporter]

No need to do this prior to 2.6.

Original comment by pekka.klarck on 3 Jul 2010 at 8:03

• Added labels: Target-2.6
• Removed labels: Target-2.5.1

[GoogleCodeExporter]

The 5) is something that should be done 2.5.1 as it adds extra stuff to the 
comments and there is no way how it could be handled in RIDE.

Original comment by jpran...@gmail.com on 5 Jul 2010 at 5:57

• Added labels: Target-2.5.1
• Removed labels: Target-2.6

[GoogleCodeExporter]

5) and 6) were done in r3817 and are included in RF 2.5.1.

Original comment by jussi.ao...@gmail.com on 22 Jul 2010 at 11:37

[GoogleCodeExporter]

5) and 6) were documented in separate issue 599. Rest of these changes are too 
big to be done in minor releases.

Original comment by jussi.ao...@gmail.com on 22 Jul 2010 at 11:40

• Added labels: Target-2.6
• Removed labels: Target-2.5.1

[GoogleCodeExporter]

This isn't necessarily related to parsing but I'll add it here to this meta 
issue anyway:

7) Expose `EmbeddedArgsTemplate` that is located in 
`robot/running/userkeyword.py`. Jussi needed to copy-paste code when he added 
embedded args support to RIDE.

Original comment by pekka.klarck on 20 Dec 2010 at 2:21

[GoogleCodeExporter]

Issue 814 has been merged into this issue.

Original comment by pekka.klarck on 13 Apr 2011 at 8:49

[GoogleCodeExporter]

8) Make _Setting non-private i.e. rename it to Setting (originally from issue 
814)

Original comment by pekka.klarck on 13 Apr 2011 at 8:53

[GoogleCodeExporter]

Settings are difficult to deal with in a test case. For example, there is no 
list of settings. I have to iterate over all attributes and call is_setting. 
IMO the settings should be a list or dictionary so I can easily iterate over 
them -- they shouldn't be mixed in with internal attributes. 

Also, settings need a unified API. Right now, if I want to modify a setting I 
have to special-case anything that is an instance of settings.Fixture because 
it uses .name and .args rather than .value like all the other settings.

ideally I should be able to change all the settings with something as simple as:

    new_setting = ["setup": ..., "teardown": ..., ...]
    for setting in testcase.settings:
        setting.value = new_setting[setting.name]

Original comment by bryan.oa...@gmail.com on 6 May 2011 at 3:22

[GoogleCodeExporter]

RF 2.6 is so far compatible with the latest RIDE releases and we don't want to 
break that compatibility with these changes. We have already decided to bundle 
Robot with RIDE in the future so that the installed Robot version wouldn't 
matter, and doing these changes after that is safe. Hopefully bundling is done 
already before inevitable 2.6.1.

Original comment by pekka.klarck on 13 Jun 2011 at 11:01

• Added labels: Target-2.6.1

[GoogleCodeExporter]

Original comment by robotframework@gmail.com on 27 Jul 2011 at 11:30

• Added labels: Target-2.6.2
• Removed labels: Target-2.6.1

[GoogleCodeExporter]

This issue is slightly backwards incompatible and thus safer to do in a major 
release.

Original comment by pekka.klarck on 6 Sep 2011 at 11:50

• Added labels: Target-2.7

[GoogleCodeExporter]

This issue was updated by revision 3fc16d4b41d3.

8) is now done.

Original comment by janne.t....@gmail.com on 11 Oct 2011 at 8:38

• Changed state: Started

[GoogleCodeExporter]

7) Can be achieved by subclassing, so no need for changes in RF itself.

Original comment by janne.t....@gmail.com on 11 Oct 2011 at 9:49

[GoogleCodeExporter]

The various enhancements have been already implemented before 2.7. Copying 
tests/keywords will be implemented later (before 2.7 is released, though)

Original comment by janne.t....@gmail.com on 11 Oct 2011 at 9:55

• Changed state: Done
• Removed labels: Target-2.7

[GoogleCodeExporter]

So, am I correct in my understanding, that item #4 (comment 3) is done, and the 
item in comment 12 is also done?

Original comment by bryan.oa...@gmail.com on 11 Oct 2011 at 10:46

[GoogleCodeExporter]

item #4 is fixed already with RF 2.6.3.

As for the item in comment 12, I plan to add #settings property to tests and 
user keywords. The settings themselves have consistent API using #as_list() and 
#populate() methods.

Original comment by janne.t....@gmail.com on 11 Oct 2011 at 5:27