Oh, I am testing slicing the Duplex model and I notice it crashes with OCCT 7.0.0. Did you happen to test slicing Duplex by any change on Linux with OCCT 7.0.0?

Great! Many thanks. I am having a look now.

Did you happen to test slicing Duplex by any change on Linux with OCCT 7.0.0?

No, I will. It doesn't crash on Windows or you just didn't happen to check?

Is there a reason why you copied over the entire files you want to patch rather than submit patches and git apply them like you did before?

Frankly, I was simply tired after hours of OCCT & OCE building to research into if/how git apply could be used for non version controlled files. I tried quickly creating patch files with WinMerge but those turned out not to be in format git apply would accept. If this is doable, I'm happy to provide a cleaner PR (some unwanted merge commits already sneaked in).

Well, one of the reasons I asked is that I get unexplainable %ERRORLEVEL%=1 from the copy command. Even though the script prints 1 file(s) copied. and the file indeed seems updated.

Then later on I get a CMake error:

CMake Error at adm/cmake/occt_macros.cmake:43 (math):
  math cannot parse the expression: "32 + 32*(/8)": syntax error, unexpected
  exp_DIVIDE, expecting exp_OPENPARENT or exp_NUMBER (10)

meaning that ${CMAKE_SIZEOF_VOID_P} is empty.

# COMPILER_BITNESS variable
macro (OCCT_MAKE_COMPILER_BITNESS)
  math (EXPR COMPILER_BITNESS "32 + 32*(${CMAKE_SIZEOF_VOID_P}/8)")
endmacro()

Strange errors. Not sure what all is happening here.

OK, sounds weird, I'll retry rebuilding a clean OCCT in a sec. The CMake problems sound odd - what's your version BTW? I'm using 64-bit 3.6.1.

OK, I'm getting the same copy error.

Fix committed. I'll look into cleaning up this PR tomorrow.

This is fine for me. Little reason to assume the copy would fail. Perhaps using xcopy or robocopy would also have been an option.

EDIT: scrap the following, it might be possible.

A swift research sheds no light on using git diff/apply for non version controlled files. Would you like be to use PowerShell for patching or are the file copy patches OK?

I'll take that back: git diff works OK but I cannot figure out how to use git apply for untracked files: the command runs successfully but it doesn't alter the file at all.

So, if this satisfies you as is, it's ready for merging. I can squash the commits if wanted.

Thanks, works like a charm, as always.

I wonder why on earth OCCT has changed usage of bool to unsigned int, oh well...

Yeah, this is a bit peculiar. Maybe a better way to go about this would have been to use !!b rather than b!=0. As the first one would be less awkward when OCCT decides to typedef Standard_Boolean back to bool. Or doesn't it get rid of the warning? Well, doesn't matter.

Just have one small thing relating to this ever annoying PCRE library dir. I will commit it after merging.

Did you happen to test slicing Duplex by any change on Linux with OCCT 7.0.0?

No, I will. It doesn't crash on Windows or you just didn't happen to check?

I can confirm the same happens on Windows. Will have a look.