marisa-trie does not build under mingw32 about marisa-trie

[GoogleCodeExporter]

Thanks for your report.

This problem may be fixed in r94.

http://code.google.com/p/marisa-trie/source/diff?spec=svn94&r=94&format=side&pat
h=/trunk/lib/marisa/grimoire/io/mapper.cc

Original comment by [email protected] on 29 Aug 2012 at 2:09

• Changed state: Accepted

[GoogleCodeExporter]

Original comment by [email protected] on 29 Aug 2012 at 5:00

[GoogleCodeExporter]

Thanks for a quick fix! 

Original comment by [email protected] on 29 Aug 2012 at 10:53

[GoogleCodeExporter]

The original issue is confirmed to be fixed, thanks again! 

It is reported that saving doesn't work under mingw32, but I don't know if this 
a bug in my Python wrapper or not.

Original comment by [email protected] on 29 Aug 2012 at 7:04

[GoogleCodeExporter]

Thanks for your valuable report.

I see the comment about the problem 
(https://github.com/kmike/marisa-trie/issues/1#issuecomment-8135066).

It seems that marisa-trie threw an exception, but I'm not sure where the 
exception was thrown.
So, I'd like to ask you what error message was thrown with the exception.

Could you check it?

Original comment by [email protected] on 30 Aug 2012 at 12:52

[GoogleCodeExporter]

I haven't figured out how to properly wrap C++ exceptions (preserving the error 
messages) using Cython so the exception raised is just a generic RuntimeError. 

I also don't have an access to a Windows machine to try to debug this issue 
myself (probably using your SWIG Python wrapper). 

So unfortunately there are no more details available and I probably won't be 
able to provide them soon.

Original comment by [email protected] on 30 Aug 2012 at 1:02

[GoogleCodeExporter]

I've submitted a new issue for the saving problem, because it is another 
problem.

http://code.google.com/p/marisa-trie/issues/detail?id=13

Original comment by [email protected] on 31 Aug 2012 at 1:23

• Changed state: Fixed