Data corruption when saving dives in win daily 4.2.-457

Lubomir I. Ivanov neolit123 at gmail.com
Fri Nov 14 09:24:17 PST 2014


On 11 November 2014 23:30, Salvador Cuñat <salvador.cunat at gmail.com> wrote:
> Good night.
>
> Yesterday I tried the last daily buid for win. Everything went fine, but
> after saving the log it couldn't be reopened.
> Tried on linux I've got this:
>
> boret at imladris:~/subsurface$ ./subsurface
> /home/boret/boret.xml:80819: parser error : Input is not proper UTF-8,
> indicate encoding !
> Bytes: 0xE4 0x27 0x20 0x2F
>   <cylinder size='12.0 l' workpressure='200.0 bar' description='�' />
>
> and with a little grep on the log:
>
>   <cylinder size='15.0 l' workpressure='200.0 bar' description='15ℓ 200 bar'
> start='195.0 bar' end='40.0 bar' />
>   <cylinder size='12.0 l' workpressure='232.0 bar' description='12ℓ 232 bar'
> start='200.0 bar' end='60.0 bar' />
>   <cylinder size='12.0 l' workpressure='232.0 bar' description='12ℓ 232 bar'
> start='215.0 bar' end='65.0 bar' />
>   <cylinder size='12.0 l' workpressure='200.0 bar' description='�       L
> 200 bar' start='190.0 bar' end='35.0 bar' />
>   <cylinder size='12.0 l' workpressure='200.0 bar' description='�' />
>   <cylinder size='12.0 l' workpressure='200.0 bar' description='4' />
>   <cylinder size='12.0 l' workpressure='200.0 bar' description='<' />
>   <cylinder size='12.0 l' workpressure='200.0 bar' description='�?L 200 bar'
> />
>   <cylinder size='12.0 l' workpressure='200.0 bar' description='�?L 200 bar'
> />
>
> The last 6 tanks are from the added dive, only the first one was used, but
> OSTC reports all of them, althoug inactive.
>
> Tried latest git in linux worked fine, it seems a bad encoding for the "l"
> thing in the windows side.
>

just built master on Win32 and i was able to reproduce the bug.

- load test0.xml
- if you add one "15ℓ 200 bar" cylinder its OK.
- but if you add a second "15ℓ 200 bar", the beginning of the
description of the first cylinder gets corrupted (e.g. has �).
(it can be seen on the UI)

i think it happens after adding any second type of cylinder and
regardless of the type of the first one.

console shows:
QAbstractItemModel::endInsertRows:  Invalid index ( 57 , 0 ) in model
TankInfoModel(0x3200ebc8)

i can try looking into that later tonight or tomorrow morning.

lubomir
--


More information about the subsurface mailing list