I think I need to add some tests to Travis-CI that use
PGObject::Util::DBAdmin to test the loadability of these files.

We have much better testing here on master and 1.4 than we do on 1.3....

On Mon, Sep 15, 2014 at 10:47 AM, Michael Richardson <[email protected]>
wrote:

>
> I tried 1.3.43rc2.  I got past the employee__save, although I had
> two ghost postgres accounts half-created that I should have cleaned up
> first.
> I'll try again after dropping all the roles, but see the bottom of this
> email.
>
> I then got, in the edit function of the main system, trying to clean stuff
> up, no function employee__get, so I decided to do:
>
> ostgres@money13:/usr/share/ledgersmb$ psql credilorg
> psql (9.1.13)
> Type "help" for help.
>
> credilorg=# \i sql/modules/Employee.sql
>
> and I got:
>
> psql:sql/modules/Employee.sql:126: ERROR:  syntax error at or near "."
> LINE 7:           ee.startdate, ee.enddate, ee.role, ee.ssn, ee.sale...
>                     ^
> psql:sql/modules/Employee.sql:130: ERROR:  function employee__get(integer)
> does not exist
> psql:sql/modules/Employee.sql:155: ERROR:  return type mismatch in function
> declared to return employee_result
> DETAIL:  Final statement returns date instead of boolean at column 7.
>
> looking deeper, I found the missing trailing ,
>
> mcr@money13:/usr/share/ledgersmb$ git diff !$
> git diff sql/modules/Employee.sql
> diff --git a/sql/modules/Employee.sql b/sql/modules/Employee.sql
> index ff6b454..8db9dd2 100644
> --- a/sql/modules/Employee.sql
> +++ b/sql/modules/Employee.sql
> @@ -114,7 +114,7 @@ CREATE OR REPLACE FUNCTION employee__get
>  returns employee_result as
>  $$
>     SELECT p.entity_id, p.id, s.salutation,
> -          p.first_name, p.middle_name, p.last_name, ee.is_manager
> +          p.first_name, p.middle_name, p.last_name, ee.is_manager,
>            ee.startdate, ee.enddate, ee.role, ee.ssn, ee.sales,
> ee.manager_id,
>            mp.first_name, mp.last_name, ee.employeenumber, ee.dob,
> e.country_id
>       FROM person p
>
>
> {I wonder if I should just upgrade from 1.2 to 1.4. An upgrade that skips a
> version is usually a bad idea, but maybe it's better tested... I suspect
> that
> many might still be at 1.2 due to fears about 1.3 stability...}
>
> It's also very hard to see the errors due to the verbosity, or even know
> where the errors are going during the upgrade process.  Best I have done
> is:
>   postgres@money13:/tmp/ledgersmb$ tail -f * | grep ERROR
>
> ===== this is what I get from the AddUser page:
> Error!
>
> Conflict with Existing Data. Perhaps you already entered this?
> More information has been reported in the error logs
>
> I understand the whole import user part; if I say "Yes", then it won't
> create a new database user.  But I'm rather unclear what other info
> I need to supply, and I think that if I've just dropped that database role,
> then I shouldn't hit import.
>
> And more information doesn't get reported.
> Aha, it's in error.log:
>
> ==> /var/log/apache2/error.log <==
> DETAIL:  Key (username)=(mcr) already exists.
> CONTEXT:  SQL statement "insert into users (id, username, entity_id)
> VALUES (
>                 v_user_id,
>                 in_username,
>                 in_entity_id
>             )"
> PL/pgSQL function "admin__save_user" line 49 at SQL statement
>
> After futzing around with creating a second user, I tried to edit the
> original "mcr" user and got:
>
> Error!
>
> 22P02:ERROR: invalid input syntax for type boolean: "manager"
> LINE 1: ...12-30', NULL, NULL, 'user', NULL, '1', NULL, '2', 'manager') ^
>
> from: ... well, doesn't help, it the URL hasn't changed.
>
> --
> ]               Never tell me the odds!                 | ipv6 mesh
> networks [
> ]   Michael Richardson, Sandelman Software Works        | network
> architect  [
> ]     [email protected]  http://www.sandelman.ca/        |   ruby on
> rails    [
>
>
>
>
>
> ------------------------------------------------------------------------------
> Want excitement?
> Manually upgrade your production database.
> When you want reliability, choose Perforce
> Perforce version control. Predictably reliable.
>
> http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
> _______________________________________________
> Ledger-smb-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/ledger-smb-devel
>



-- 
Best Wishes,
Chris Travers

Efficito:  Hosted Accounting and ERP.  Robust and Flexible.  No vendor
lock-in.
http://www.efficito.com/learn_more
------------------------------------------------------------------------------
Want excitement?
Manually upgrade your production database.
When you want reliability, choose Perforce
Perforce version control. Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
Ledger-smb-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ledger-smb-devel

Reply via email to