3.0.2

[Lubomir I. Ivanov]

sergey,

can i ask you to give this binary a try:
http://dl.dropbox.com/u/1627980/subsurface/subsurface_x86_debug.zip

it includes a possible fix (by a distant guess) for the bug you are
reporting, but also some debugging output options.
if the bug still exists you can try gathering some useful info in the
following way:

- extract the exe to a "subsurface" folder
- you need my _dependencies.zip from here in the same folder:
https://github.com/neolit123/subsurface_win32
- select a cyrillic locale, either a bg_BG or ru_RU locale. i read
Russian, so no worries.
- open "start menu" -> run -> cmd
- navigate to the subsurface folder and start subsurface as:
subsurface.exe > output_save.txt 2>&1
- create a planner dive, reproduce all the steps for the save stage, etc...exit
- call subsurface again:
subsurface.exe > output_parse.txt 2>&1
- exit subsurface

you can then send me the output*.txt files and examine them yourself,
but perhaps you should zip them,
since i've noticed that your previous email attachment had corrupted
encoding(?).

this build should dump hopefully all relevant information to the
parsing/planner/save stages in those files.
you can look for a place where something doesn't look right in there.

thanks
lubomir
--