[Dbix-class] update on DBIC / SQLA1.50 tests

[Dami Laurent (PJ)]

>-----Message d'origine-----
>De : Peter Rabbitson [mailto:rabbit+list at rabbit.us] 
>Envoyé : jeudi, 6. novembre 2008 09:31
>À : Class user and developer list
>Objet : Re: [Dbix-class] update on DBIC / SQLA1.50 tests
>
>Dami Laurent (PJ) wrote:
>> For info : the latest revision of SQLA1.50
>> 
>(http://dev.catalyst.perl.org/svnweb/bast/browse/SQL-Abstract/1
.x/branch
>> es/1.50_RC/) 
>> together with the DBIC patch from mendel 
>(http://scsys.co.uk:8001/20236)
>> fixes most problems.
>> 
>> 
>> Still have to look at failures in 
>>   t\95sql_maker_quote
>> and
>>   t\96file_column
>> 
>> 
>
>t\96file_column does not fail for me, please provide your test output.


Seem to be Win32 problems about files :

not ok 5 - old storage does not exist # TODO design change required
#   Failed (TODO) test 'old storage does not exist'
#   at t/96file_column.t line 54.

ok 8 - new content matches
Can't remove file D:\DOCUME~1\dami\LOCALS~1\Temp\Ty9g5V4LhY\1/File.pm (Permission denied) at lib/DBIx/Class/InflateColumn/File.pm line 51
Can't remove directory D:\DOCUME~1\dami\LOCALS~1\Temp\Ty9g5V4LhY\1 (Directory not empty) at lib/DBIx/Class/InflateColumn/File.pm line 51
ok 9 - storage deleted
not ok 10 - storage does not exist after $rs->delete # TODO need resultset delete override to delete_all
#   Failed (TODO) test 'storage does not exist after $rs->delete'
#   at t/96file_column.t line 84.


>Other than that I see the same output. Apart from bracketed 
>mismatches I
>have order_by DESC is quoted wrong in 95sql_maker_quote.t
>

That one is quite delicate, there are several problems :

a) after long discussions with MST, the agreement is that "column ASC/DESC"
   will NOT be automatically quoted as `column` ASC/DESC; it will be
   quted as `column ASC/DESC`. The proper syntax for separating the ASC/DESC 
   modifier is {-asc => "column"} or {-desc => "column"}

b) current _order_by in DBIx::Class::Storage::DBI is not adapted for this
   syntax change, because a hashref {-desc => "column"} passed as argument
   is silently ignored.

c) there is a problem in SQL::Abstract::Test, because "is_same_sql_bind" does
   not detect when it is inside a TODO block