This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Other format: | [Raw text] |
I have been working with the experimental WIN64 support for MinGW, and I have run into something that I cannot seem to solve. Earlier this week, I updated to Binutils 2.18 after successfully using a snapshot built using the CVS HEAD sources from June 29, 2007. Since then, I have been unable to create static libraries using x86_64-pc-mingw32-ar and x86_64-pc-mingw32-dlltool. Regrettably, I did not back up my old binutils installation. In general, I get errors such as this when trying to do anything with the static libraries created by these tools: C:\msys\1.0\local\bin\x86_64-pc-mingw32-nm.exe: libwhatever.a: File format not recognized I have seen x86_64-pc-mingw32-ar complain about truncated object files when trying to add to an existing archive, but that only occurred when I tried recreating bugs using simple test cases. I have tried building binutils using the MinGW versions of GCC 3.4.5 and 4.2.1 as well as the Cygwin version of GCC 3.4.4. I also tried backing off to the June 29 sources that were working, and they fail, too. Because of that, I suspect very strongly that there is something wrong in my environment, but I have had no luck in figuring out what. I tried getting my 64-bit toolchain out of the way (renamed /usr/local to /usr/local-broken) and installing a fresh binutils 2.18, but x86_64-pc-mingw32-ar and x86_64-pc-mingw32-dlltool keep giving me bad static libraries. I certainly want to be able to rule out operator error before I file any binutils bug report. Every time I have built binutils, I use the following commands from within a build directory under the binutils source tree: ../configure --target=x86_64-pc-mingw32 make make install In case it makes a difference, I can use the Visual C++ linker to make static libraries from object files compiled using GCC 4.3 (built from sources checked out three days ago). For a while, I thought maybe the object files being generated by GCC were corrupt or truncated, but that seems not to be the case. -Patrick -- Patrick L. Hartling VP Engineering, Infiscape Corp. http://www.infiscape.com/
Attachment:
signature.asc
Description: OpenPGP digital signature
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |