Hardcoded moneyGuru

Comments and changes to this ticket

• 

  Virgil Dupras March 21st, 2011 @ 01:42 PM

  Got another one:

  Application Identifier: com.hardcoded_software.moneyguru
  Application Version: 2.3.5
  
  Traceback (most recent call last):
   File "hscommon/cocoa/inter.pyc", line 147, in setValue_forColumn_row_
  AttributeError: 'NoneType' object has no attribute 'set_cell_value'
  
  Relevant Console logs:
  
  Mar 19 22:34:29 iMac-de-Gregory-TRAUMAT [0x0-0x1f21f2].com.hardcoded_software.moneyguru[3164]: WARNING Trying to get an out of bounds row (22 / 21)
  

• 

  Virgil Dupras March 21st, 2011 @ 05:55 PM

  • Assigned user set to “Virgil Dupras”

  Hum, the last log is not quite the same as the first, might be unrelated.

• 

  Virgil Dupras March 21st, 2011 @ 06:11 PM

  • State changed from “new” to “hold”

  I have no idea how such a crash is possible. Obviously, the first crash is in the entry table since it's the only table that has a transfer column, but still, I can't reproduce the crash. For the second one, it would help to know the table name, so I'll add that in the logged warning.

• 

  Virgil Dupras March 21st, 2011 @ 06:11 PM

  (and next time I receive such a crash report, I have the ask the user how it happened)

• 

  Virgil Dupras March 22nd, 2011 @ 09:04 AM

  • Assigned user cleared.
• 

  Virgil Dupras April 2nd, 2011 @ 11:14 AM

  • State changed from “hold” to “new”

  Got details from a user having had the same crash:

  The sequence as far as I remember was:

  I added a new entry to the end of an existing account opened at the 'March' page. When I entered the date it opened an 'April' page I finished that entry an added another The program crash when I tried to look at another account.

  Gotta try to reproduce again with this in mind.

• 

  Virgil Dupras April 18th, 2011 @ 11:54 AM

  • State changed from “new” to “hold”

  Ugh, can't reproduce :(

• 

  Virgil Dupras June 1st, 2011 @ 02:08 PM

  • State changed from “hold” to “new”

  Another similar crash, with another traceback, maybe there's a way to find something from this one:

  Application Identifier: com.hardcoded_software.moneyguru
  Application Version: 2.3.7
  
  Traceback (most recent call last):
    File "/Applications/moneyGuru.app/Contents/Resources/mg_cocoa.plugin/Contents/Resources/mg_cocoa.py", line 1142, in showAccount
    File "core/gui/main_window.pyc", line 365, in show_account
    File "core/gui/account_view.pyc", line 89, in show_account
    File "core/gui/entry_table.pyc", line 96, in show_transfer_account
    File "core/gui/main_window.pyc", line 446, in shown_account
    File "core/gui/main_window.pyc", line 64, in _add_pane
    File "core/gui/main_window.pyc", line 392, in current_pane_index
    File "core/gui/main_window.pyc", line 75, in _change_current_pane
    File "core/gui/account_view.pyc", line 38, in show
    File "core/gui/base.pyc", line 268, in show
    File "core/gui/transaction_table_base.pyc", line 43, in show
    File "core/gui/base.pyc", line 129, in show
    File "core/gui/transaction_table_base.pyc", line 40, in _revalidate
    File "hscommon/gui/table.pyc", line 251, in refresh
    File "core/gui/entry_table.pyc", line 54, in _fill
    File "core/gui/column.pyc", line 97, in set_column_visible
    File "/Applications/moneyGuru.app/Contents/Resources/mg_cocoa.plugin/Contents/Resources/mg_cocoa.py", line 302, in set_column_visible
    File "hscommon/cocoa/inter.pyc", line 148, in setValue_forColumn_row_
  AttributeError: 'NoneType' object has no attribute 'set_cell_value'
  
  Relevant Console logs:
  
  May 27 01:28:58 MBAir256-Ppc [0x0-0x111111].com.hardcoded_software.moneyguru[2317]: WARNING Trying to get an out of bounds row (20 / 2) on table EntryTable
  May 27 01:28:58 MBAir256-Ppc [0x0-0x111111].com.hardcoded_software.moneyguru[2317]: WARNING Trying to set an attribute that can't: description with value Imp\xf4t 02 at row 20 on table EntryTable
  

• 

  Virgil Dupras June 4th, 2011 @ 04:04 PM

  • State changed from “new” to “hold”

  Nah, still the same unreproducable thing. The key here is not in how the txn-save operation is triggered (pane switch, show account, whatever) but very likely what kind of item is edited.

  Note, however, that in the last traceback, the edited item seems to have been an account since the traceback starts from an account view.

• 

  Virgil Dupras June 5th, 2011 @ 09:58 PM

  • State changed from “hold” to “accepted”

  Ah... at last! I got it! After the hint of #264, I told myself that it's very possible that the bug here also comes from tab switching... and it does! Steps:

  1. Open two accounts, with one of them having more txns than the other.
     
  2. The smallest account (1) goes first, the biggest one (2) after it.
     
  3. While having account 1 selected, swap it with account 2. There, you'll see a glitch: Instead of still having account 1's contents in the table, you'll have account 2's, despite account 1 being selected in the panes.
     
  4. Start editing a txn that has an index greater the the txn count of the other account.
     
  5. Don't commit the edit, and select the pane of the other account. Nothing has changed because the other account's pane is in fact the pane of the account we're editing.
     
  6. Select the other account again: boom crash!

  Gosh am I glad to have found this!

• 

  Virgil Dupras June 7th, 2011 @ 05:12 PM

  • State changed from “accepted” to “fixed”

  This was already fixed by #264