Current MinGW+MSYS versions deployed on my Wintel have  the libtool  [2.4.1-msys ?? I doubt: really only usual  libtool-2.4-1-mingw32-bin.tar.lzma ],
GCC 4.8.1.
Before a build of the 2.4.2 [for MinGW - not for MSYS] (i.e. the one for this report): I already decided to use (even for a build and hence for this build) 
the libtool file (in /mingw/bin) with my manual correction: I saw the strange [for the presence of GCC 4.8.1 deployed] script - working, 
but inside it there were 4.5.0 paths instead of 4.8.1 (and this is not strange for the  libtool-2.4-1-mingw32-bin.tar.lzma , but only for the  current MinGW+MSYS).
I almost ignore this hand-made file,  but here it is among the attachment.
The  build of the 2.4.2 that  I mention above  was successful, but here the .log - more important file, of cource - is among  the attachment.  

(One can see 
sys_lib_search_path_spec="c:/tingw/lib/gcc/mingw32/4.8.1 c:/tingw/lib/gcc c:/tingw/mingw32/lib c:/tingw/lib "
and  
compiler_lib_search_dirs="c:/tingw/bin/../lib/gcc/mingw32/4.8.1 c:/tingw/bin/../lib/gcc c:/tingw/bin/../lib/gcc/mingw32/4.8.1/../../../../mingw32/lib c:/tingw/bin/../lib/gcc/mingw32/4.8.1/../../.." 
in the 312.96KB file of the 2.4.2 build 
- the "TinGW" substrings are due to my allusion to tin plague under a frost:/)

Best regards, while you can be giving an advice what a libtool without any hand-made file must become a mean for my case.