abijah
7181bc7f9b
Minor logic and formatting changes. Considering making Linkable operate recursively.
...
git-svn-id: file:///svn-source/pmgr/branches/yafr_20090716/site@354 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-07-20 19:00:54 +00:00
abijah
3d262fd4db
This rework is nowhere near complete, but there are certain things that are falling in place, and worth capturing. I started a branch for just this purpose of being able to check in intermediate work.
...
git-svn-id: file:///svn-source/pmgr/branches/yafr_20090716/site@352 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-07-19 23:35:25 +00:00
abijah
3550bf775c
Merge in the bug fixes made to Linkable, as well as the better logging mechanism.
...
git-svn-id: file:///svn-source/pmgr/branches/invoice_receipt_20090629/site@194 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-06-30 18:11:40 +00:00
abijah
cb02b9e1e0
Further tweaking to the linkable behavior, first by fixing the $efields addition from the last checkin of this file, and second by fixing a problem with linking onto a HABTM relationship, since the alias was coming out screwy (although this issue was hacked by using 'table' to determine the class and alias).
...
git-svn-id: file:///svn-source/pmgr/branches/ledger_transactions_20090605/site@151 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-06-16 21:39:27 +00:00
abijah
76e9b7590f
Working version of the Linkable behavior that allows for a HABTM joining table to have an alias specified (thus allowing more than one instance in a query). I don't like the $efields bit though, and will change that on the next checkin. Don't want to lose a working version though!
...
git-svn-id: file:///svn-source/pmgr/branches/ledger_transactions_20090605/site@148 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-06-16 17:19:07 +00:00
abijah
e9dd1366bc
Further work on the linkable behavior, since it wasn't working right with aliases. It doesn't help matters that models are instantiated from the model associations with their own alias. Thus, the priority of a tables alias is 1) 'alias' as defined in the 'link' structure'; 2) alias member of the model instance; 3) class name of the model instance. I may need to revisit this work later, since while typing these comments I managed to enlighten myself on the alias scheme. But it does seem to work for now.
...
git-svn-id: file:///svn-source/pmgr/branches/ledger_transactions_20090605/site@140 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-06-15 23:06:27 +00:00
abijah
c3139220ee
Moved the ledger_entries controller/view to jqGrid.
...
Changed the app controller jqGrid virtual functions for getting the tables for both count and the full query. The ExtraTables bit was too confusing, and was only intended to allow the user to specify a subset for count and augment it for full query. This is easily accomplished by having DataTables just call DataCountTables first... one line solution ends the confusion.
Modified linkable behavior to support a %{MODEL_ALIAS} tag, which is replaced in the relationship conditions at runtime with the actual model alias. This is experimental at best, and certainly will create a problem for any model that ends up using the conditions in 'contain' instead of 'link'.
Broke the LedgerEntry->findInLedgerContext function out into multiple pieces, so those same pieces could be used in the LedgerEntries controller to populate jqGrid.
Changed the ledger_entries element, as a special case, to also allow listing control from ledger_id and account_type, instead of idlist as a mechanism for populating the grid.
Changed the infobox summary css, which will break several pages until I rework them.
Some next steps include fixing the broken infoboxes, and changing the transactions view to use the ledger_entries element. This also means the ledger_entries element will need to add support for idlist, or if we have any credit/debit issues, perhaps another custom transaction_id instead.
git-svn-id: file:///svn-source/pmgr/branches/ledger_transactions_20090605/site@134 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-06-15 19:41:20 +00:00
abijah
3dcd83229b
Fixed problem with relationships that are defined soley on the 'conditions' part of the association, and not on 'foreignKey' (which is allowed to be 'false').
...
git-svn-id: file:///svn-source/pmgr/branches/ledger_transactions_20090605/site@80 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-06-10 02:31:49 +00:00
abijah
677458e942
I'm still in the middle of moving onto a ledger based system. However, I'm am now changing how transactions and entries relate back to the customer. I'll be using a ledger for each lease (for rent, late charges, security deposits, etc), and a ledger for each customer (for POS, non-specific deposits such as reservations or covering mulitple units, bad debt writeoff, and possibly customer credits, when not obviously lease specific). This coming change might not be in the right direction, so I want to capture the work as is right now. This change set is not fully functional. Many operations do work, but there are obviously transaction problems with units and customers.
...
git-svn-id: file:///svn-source/pmgr/branches/ledger_transactions_20090605/site@71 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-06-06 20:18:56 +00:00
abijah
991b5a3317
Performed cleanup, and started the thought process of figuring out how to use one link table multiple times in the same query. My changes worked actually, but I want to wait on this to see if I can figure out if there is already a core solution (since it seems likely that things behaving like a tree would have one table used multiple times in the same query... although they probably get around that by running 100 different queries come to think of it).
...
git-svn-id: file:///svn-source/pmgr/branches/initial_20090526/site@50 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-06-01 04:27:34 +00:00
abijah
68585a0b30
I've been around and around with paging. I started using the Containable behavior, which is fantastic for something like 'view', but is highly problematic for listings, especially with pagination. I discovered the Linkable behavior, and it's much more database efficient, and I actually can get it to work with pagination (except when using GROUP BY). I did have to tweak it though, to handle some of the more complex queries that I intend. This checkin includes a bunch of garbage, but given the amount of time I've spent on this crap, I'm afraid to lose anything that might later be useful information. I'll clean up on the next checkin.
...
git-svn-id: file:///svn-source/pmgr/branches/initial_20090526/site@49 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-06-01 02:36:26 +00:00
abijah
ce252f2e70
Changed how fields are handled, so that they are all included by default and the user is not forced to specify them. This works with paginate, but the additional fields are requested in the count query. Would like to resolve this, but it doesn't seem to be confusing paginate.
...
git-svn-id: file:///svn-source/pmgr/branches/initial_20090526/site@46 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-05-31 18:27:21 +00:00
abijah
4b857ff11f
Testing out a possible model behavior called linkable. I may not go this route, but I'll just delete it if not.
...
git-svn-id: file:///svn-source/pmgr/branches/initial_20090526/site@45 97e9348a-65ac-dc4b-aefc-98561f571b83
2009-05-31 17:02:41 +00:00
