Valid modules are removed in specific case

[DarkFenX]

[Confessor, Confessor fit]

[Empty Low slot]
[Empty Low slot]
[Empty Low slot]
[Empty Low slot]
[Empty Low slot]

[Empty Med slot]
[Empty Med slot]
[Empty Med slot]

Small Focused Pulse Laser II
[Empty High slot]
Small Focused Pulse Laser II
Small Focused Pulse Laser II
[Empty High slot]
Small Focused Pulse Laser II

[Empty Rig slot]
[Empty Rig slot]
[Empty Rig slot]

Import fit. Rearrange guns the way it's shown on EFT format fit. Relaunch pyfa.

2015-07-09 23:28:42,221 eos.saveddata.module     ERROR    Item (id: 0) is not a Module
2015-07-09 23:28:42,222 eos.saveddata.module     ERROR    Item (id: 0) is not a Module
2015-07-09 23:28:42,222 eos.saveddata.module     ERROR    Item (id: 0) is not a Module

3 lasers (besides 1st one) are removed.

[blitzmann]

It seems to remove any module which is moved. I did not consider testing that scenario, as that doesn't really have to do with the modules item ID. Will look into it

[blitzmann]

You know, this probably has something to do with the os x fix. In order to flag something, I just set itemID to 0. Didn't consider this, where modules are removed and readded. The itemID is probably overwritten. Hopefully it's a simple fix

[blitzmann]

Got a fix, but it's not ideal (then again, the ideal thing is not supported by our OS X lol)

def remove(self, thing):
    # We must flag it as modified, otherwise it not be removed from the database
    # @todo: flag_modified isn't in os x skel. need to rebuild to include
    #flag_modified(thing, "itemID")
    if thing.isInvalid:
        thing.itemID = 0
    list.remove(self, thing)

if thing.isInvalid is the key part. I checked, and indeed modules are removed and then inserted again when they swap positions, rather than just swapping the position. When they are removed, the itemID was reset. So we check if the item is inValid, in which case it's definitely okay to reset the itemID.