Asier Lostalé <asier.lostale@openbravo.com> [Thu, 16 Mar 2017 15:05:03 +0100] rev 31697
fixed bug
35543: incorrect export of indexes including CASE expressions
Indexes with CASE expressions, were incorrectly and differently exported in
PG and ORA. PG transforms the expression in multi line and in both ORA and
PG extra blankspaces are added.
Fixed in by removing new lines in PG (the same as done in views) and removing
extra blankspaces in both cases.
Asier Lostalé <asier.lostale@openbravo.com> [Thu, 16 Mar 2017 10:24:21 +0100] rev 31696
fixed bug
35537: 0 concurrent users is seen in About popup for unlimited users
In this case now Unlimited is displayed instead
RM packaging bot <staff.rm@openbravo.com> [Wed, 15 Mar 2017 19:16:54 +0000] rev 31695
CI: merge back from main
RM packaging bot <staff.rm@openbravo.com> [Wed, 15 Mar 2017 19:01:32 +0000] rev 31694
CI: update AD_MODULE to version 31688
Pablo Lujan <pablo.lujan@openbravo.com> [Wed, 15 Mar 2017 17:34:41 +0100] rev 31693
Fixes issue
35535 and related to issue
29414: Updated browser supported/recommended list for PR17Q2
* Chrome. Supported 56, Recommended 57 or higher.
* Firefox. Supported 45, Recommended 52 or higher.
* Internet Explorer. Supported 9, Recommended 11 or higher.
* EdgeHTML (MS Edge). Supported 12, Recommended 13 or higher.
* Safari. Supported 8, Recommended 9 or higher.
Alvaro Ferraz <alvaro.ferraz@openbravo.com> [Wed, 15 Mar 2017 14:54:07 +0100] rev 31692
Related to issue
35486: Add some automated tests
Alvaro Ferraz <alvaro.ferraz@openbravo.com> [Wed, 15 Mar 2017 14:36:31 +0100] rev 31691
Fixes issue
35486: Net Amount not properly rounded in ERP with PostgreSQL
Problem was in following calculation: Net Amount = Gross Amount * (Gross Amount / (Gross Amount + Tax Amount)).
In case Gross Amount = 357.99 and Tax Amount = 71.598, Net Amount = 357.99 * (357.99 / (357.99 + 71.598)) = 298.325.
It was correct in Oracle as it retrieves 298.325 but it was failing in PostgreSQL as it retrieves 298.3249999999999999988067:
SELECT 357.99 * (357.99 / (357.99 + 71.598)) FROM dual;
If we do multiplication before division, we avoid this problem and both PostgreSQL and Oracle retrieves the correct result (298.325):
SELECT (357.99 * 357.99) / (357.99 + 71.598) FROM dual;
Mark <markmm82@gmail.com> [Wed, 15 Mar 2017 09:49:13 -0400] rev 31690
Backed out changeset: 498699b38a5d
Víctor Martínez Romanos <victor.martinez@openbravo.com> [Wed, 08 Mar 2017 12:50:21 +0100] rev 31689
Fixed bug
33581: slow defaults retrieval in add payment
The FIN_Payment table didn't have an index on the c_bpartner_id column, thus forcing a sequencial scan in any query run over this table.
Added an index on the c_bpartner_id column to fix that problem.
This index should fix most of the cases, however when for example a concrete business partner has an important percentage of the records in that table, the previous index would be useless, and the db might decide to run a seq. scan.
To control this scenario we have extended the previous index with a function to distinguish credit payments (generated credit != used credit). This way we could filter by the credit records for the business partner directly using the index.
Note that the function obequals is declared in the prescript because it must be defined as IMMUTABLE/DETERMINISTIC and this syntaxis is not supported yet by the DBSM.
The AdvPaymentMngtDao.getCustomerPaymentsWithCredit() has been rewritten to take advantage of the new index.
Asier Lostalé <asier.lostale@openbravo.com> [Wed, 15 Mar 2017 14:08:12 +0100] rev 31688
fixed bug
35474, fixed issue
35519: function based indexes improvements
Improved the way function definitions in indexes are read from DB, now they are
read per column, using in PG the index specific function pg_get_indexdef. This
allows to:
* properly read indexes based on arithmethic expressions ie (col1-col2)
* support more than one index column based on function
RM packaging bot <staff.rm@openbravo.com> [Wed, 15 Mar 2017 12:57:57 +0100] rev 31687
Merge back from main
RM packaging bot <staff.rm@openbravo.com> [Wed, 15 Mar 2017 11:16:29 +0000] rev 31686
Merge temporary head for 3.0PR17Q1
RM packaging bot <staff.rm@openbravo.com> [Mon, 13 Mar 2017 11:02:48 +0000] rev 31685
Added signature for changeset 3a7510a5ee0b
RM packaging bot <staff.rm@openbravo.com> [Mon, 13 Mar 2017 11:02:48 +0000] rev 31684
Added tag 3.0PR17Q1 for changeset 273c24deef8f
RM packaging bot <staff.rm@openbravo.com> [Mon, 13 Mar 2017 11:02:48 +0000] rev 31683
Update AD_MODULE version to 3.0PR17Q1
Miguel <martencobasmiguel@gmail.com> [Fri, 10 Mar 2017 11:11:58 -0500] rev 31682
Fixes issue
35491:Error creating an Order from a Requisition To Order.
In AUM project the Constraint c_orderline_check1 was modified, to only allow creation
of order lines whith QUANTITYORDER and M_PRODUCT_UOM_ID values null or not null at same time.
When they were created from the Requisition to Order process, the Order line's QUANTITYORDER
was created with 0 in the case it had a NULL value, instead of maintain the value as NULL.
Because of it the constraint was failing because a QUANTITYORDER=0 and a M_PRODUCT_UOM_ID=NULL.
To solve this problem was modified the linesToOrder query in RequisitionToOrder_data.xsql, to
retrieve the correct value of QUANTITYORDER, including NULL if there are no qty in the second
UOM.
Miguel <martencobasmiguel@gmail.com> [Thu, 09 Mar 2017 14:22:29 -0500] rev 31681
Fixes issue
35371: Wrong orders shown in "Create Invoices From Orders" window
Fixed the C_INVOICE_CANDIDATE_V view, changing the condition to not take into account
the sales orders different than 'ON' AND 'OB' AND 'WR' instead of using an incorrect
OR condition as suggested in the proposed solution.
Also were included some performances in the view replacing OR and AND conditions with they
respectively translations to IN and NOT IN clauses.
David Miguelez <david.miguelez@openbravo.com> [Fri, 10 Mar 2017 12:14:46 +0100] rev 31680
Fixes Issue
35479. Fixes problem while setting alias in a subquery in Oracle.
Inigo Sanchez <inigo.sanchez@openbravo.com> [Fri, 10 Mar 2017 10:24:11 +0100] rev 31679
Fixed issue
35464: Adaptive filtering does not work for Table references.
The problem was that adaptive filtering (filter in client when all data is locally available) did
not work for Table references when filtering by text instead of by selecting a value in the drop
down list. The issue that causes this regression was not taking into account adaptive filtering.
This problem has been fixed by taking into account this case in getDisplayProperty method. Now
getDisplayProperty method manages properly when all the data are in the client (Adaptive filtering).
Asier Lostalé <asier.lostale@openbravo.com> [Thu, 09 Mar 2017 13:08:47 +0100] rev 31678
fixed bug
35467: views with several IN or NOT IN clasues fail in PG
In PG, views that contain more than one IN or NOT IN clauses were not correctly
exported.
The regexps that parses and PG IN and NOT IN clauses was including from the begining
of 1st clause till the end of the last one.
These regexps have been fixed to be more restrictive so that they stop looking
for the IN contents on first ].
Miguel <martencobasmiguel@gmail.com> [Thu, 02 Mar 2017 14:17:25 -0500] rev 31677
Fixes issue
35419:Cannot remove an orphan line in Return to/from vendor/customer
Was not able to remove an orphan line from a return to vendor or a return from
customer document because an incorrect casting to boolean.
When is validating tha Last record is selected inside the Orphan grid lines, the
lastRecord.get("obSelected") object gets as String instead of a Boolean object,
in other case it gets as Boolean.
Now, conditions are rightly validated depending of the class of the object.
Carlos Aristu <carlos.aristu@openbravo.com> [Thu, 02 Mar 2017 09:22:31 +0100] rev 31676
fixes issue
35398: Check readable clients for child properties using DAL WS
Fix for issue #29683 was introducing a security check to avoid showing child properties which belong to a client different from the current client. But this fix was preventing to display some readable objects like master data defined at system level (client 0).
Therefore this security check has been improved and now it is verified if the child property is defined for a client which is readable for the current role.
Antonio Moreno <antonio.moreno@openbravo.com> [Tue, 28 Feb 2017 08:00:23 +0100] rev 31675
Fixed issue
35373. Reenable automatic discounts in purchase flows.
Antonio Moreno <antonio.moreno@openbravo.com> [Tue, 28 Feb 2017 08:00:11 +0100] rev 31674
Fixed issue
35373. Reenable automatic discounts in purchase flows.
Inigo Sanchez <inigo.sanchez@openbravo.com> [Mon, 27 Feb 2017 10:46:35 +0100] rev 31673
Fixed issue
35349:Callout problem when working with comma as decimal separator
The problem was raised when a default format were switched, it means, when
decimal and group separator were switched. In this situation, when
SimpleCalloutInformationProvider class was used to obtain the current element
value and this current element was a BigDecimal, the format was not managed
properly. This occurred because getCurrentElementValue method always retrieved
a String element instead of Object element, in tyhis case BigDecimal.
In this situation, managesUpdatedValuesForCallout method of FIC class compare
if element is a String or not. If a BigDecimal is retrieved as String element,
in this part of the code the format is not managed properly when decimal separator
is comma. The code manage the comma separator as group separator and removed it.
When the decimal separator is dot (.) the format is managed as expected although
this method retrieves a String object (e.g. "1.23"). When decimal separator is
comma (,) this BigDecimal (e.g. 1,24) is retrieved by getCurrentElementValue method
as String value and the format is missed.
This problem has been fixed by take into account type of current element inside
getCurrentElementValue method. Now this method retrieves an object value. In this
case, a BigDecimal is retrieved. Now the format is managed properly.
Carlos Aristu <carlos.aristu@openbravo.com> [Tue, 21 Feb 2017 14:16:33 +0100] rev 31672
fixes issue
35318: documents were not being attached after send them by e-mail
The CoreAttachImplementation class is deleting the original report file after completing the upload process. When attaching a document using the e-mail pop-up this is done before attaching the document into the e-mail. This explains why the document was not being received on the e-mail.
To avoid this problem, now a copy of the original document is sent to the AttachImplementationManager, this way the original file will never be deleted by an upload process handler.
Together with this, now when attaching a document after printing (not by e-mail) we have to delete the original file because in other case it will be kept on disk.
Armaignac <collazoandy4@gmail.com> [Thu, 16 Feb 2017 11:42:21 -0500] rev 31671
Fixes issue
35276: General Ledger filter changes with pagination in JE Report
Selected General Ledger filter value changes when pagination is executed in the
Journal Entries report
The General Ledger filter is not kept in the session to make use in the pagination.
For the solution, was saved the General Ledger Filter in the session every time
the search is executed (FIND action). By default, if there is any GL in the
session, the GL field will use the defaulted for the selected org. When paginating
always is used the account schema id stored in the session.
Armaignac <collazoandy4@gmail.com> [Wed, 15 Feb 2017 19:43:11 -0500] rev 31670
Fix for issue
35273: General Ledger filter removed when paginating in GL Report
The General Ledger filter was removed when paginating in General Ledger Report
in HTML view
Now the General Ledger filter it's save in the session by the FIND action and get in the DEFAULT action
for showing in the report page
Carlos Aristu <carlos.aristu@openbravo.com> [Wed, 08 Feb 2017 14:07:20 +0100] rev 31669
fixes bug
35197: Can't attach using manual role having modules in indevelopment
It was not possible to attach a document with a role that hasn't got access to ADTab entity having modules in indevelopment. This was happening because when there are modules in indevelopemnt status, the getTab method of the ADCS class does not store the tab information in cache and uses DAL to retrieve the tab every time it is being called.
Under this circumstances, the getTab method in AttachImplementationManager must be invoked in admin mode to ensure that the role can generate the attachment even when not having access to the ADTab entity.
Carlos Aristu <carlos.aristu@openbravo.com> [Tue, 07 Feb 2017 09:48:12 +0100] rev 31668
fixes issue
35177: Initialize callout's connection pool on init