> The implementation looks very clean. I poked around a bit and ran several
> random test sets.
> The feeling I got is that it's not quite ready yet for merging. I would like
> to avoid waisting time when more people start using it and all run into the
> same issues.
> Before we continue, I would like to understand why the following two problems
> exist. Especially the first one seems to be a candidate to fix many more
> tests.
> - assert-serialization doesn't not seem to be implemented. Should it? I poked
> around and found many tests failing because of this (e.g. CondExpr002)
It is implemented, but the name is 'assert-xml' please check the latest version 
of catalog-schema.xsd:
look for 'assert-xml'

> - It seems that all of the prod-OrderByClause tests fail. It's not clear
> whether this is a bug in my setting, the driver, or Zorba itself.
Already reported as lp:956460.

> Some minor issues are:
> - I find it very hard to read the output of a test run but maybe this will
> change if we have the stylesheet
> - How do I know whether my test run actually ran any tests? Sometimes I had a
> typo in the prefix and it didn't run anything but it's hard to see.
Note: the FOTS driver should be used together with the -o output.xml and 
--indent of the Zorba CLI.
The user will then get trace output in the text window and the output of the 
tests in the output.xml file.
Whenever a test was run, there is a trace output for it's name on the command 
Whenever a test that was run failed, there is a trace message for that pointing 
also the name of the test case it belongs to for easy identification.
I have updated the README.txt to reflect these issues, see 

> - Rename fots_driver to fots
> - Rename zorba-fots-driver.xqy to fots-driver.xq
People must get used to the terminology:
- FOTS means the testsuite from W3C: hence when a test is wrong => file a bug 
to W3C
- FOTS driver is the thing that runs the testsuite: hence when an error is 
detected in the FOTS driver = > file a bug in Launchpad with a "fots-driver" TAG
- test case is a single test
- test set is unit (a bunch) of tests grouped together because they test a 
certain feature/function/functionality

The renaming you suggest would confuse people.

> - We generally use xq instead of xqy for xquery files. Maybe we should do this
> here to.
If you really insist then I can rename the files, sure...

> - Does the dependency output have to appear in every run?
Do not know what you mean by "dependency output": can you please be more 
Your team Zorba Coders is subscribed to branch lp:zorba.

Mailing list: https://launchpad.net/~zorba-coders
Post to     : zorba-coders@lists.launchpad.net
Unsubscribe : https://launchpad.net/~zorba-coders
More help   : https://help.launchpad.net/ListHelp

Reply via email to