Nasty Bug in MinGW

35 views
Skip to first unread message

Gabor Grothendieck

unread,
May 7, 2014, 10:57:01 AM5/7/14
to r-an...@googlegroups.com
Just thought I would ponit out this nasty bug involving std::string with g++ on MinGW under Windows to save anyone else who runs into it.  One does not usually expect such widely used tools to be the problem but that's the case here:



Kevin Ushey

unread,
May 7, 2014, 11:43:49 AM5/7/14
to Gabor Grothendieck, r-an...@googlegroups.com
Thanks. Hopefully, the compiler toolchain will eventually get updated
for Windows, but it seems like R Core is comfortable with the current
state of affairs for now.
> --
> You received this message because you are subscribed to the Google Groups "R
> and C++" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to r-and-cpp+...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

Romain Francois

unread,
May 7, 2014, 1:51:03 PM5/7/14
to Kevin Ushey, Gabor Grothendieck, r-an...@googlegroups.com
Conceptually they are not opposed to modernize the toolchain. The first step is for us to prove that we can build R from source with a more recent gcc build.

Romain

Gabor Grothendieck

unread,
May 7, 2014, 3:16:59 PM5/7/14
to r-an...@googlegroups.com, Gabor Grothendieck
Note that there is a hack mentioned at the link I sent that will fix this problem even in the version of gcc that is being used for R with Windows now.  Applying that could be a first step and that would not involve upgrading gcc itself..
Reply all
Reply to author
Forward
0 new messages