Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo


#914 dont copmlie on win7 64bit by MinGW&MSYS


I flow the step of the guide in your wiki
everything goes ok ,but the last step
could anybody help me ?

log is here:

lion@lion-HPC /d/temp/iverilog
$ make
g++ -DHAVE_CONFIG_H -I. -I. -Wall -Wshadow -g -O2 -MD -c main.cc -o main.o
In file included from netlist.h:38:0,
from pform.h:22,
from main.cc:56:
verinum.h:133:8: error: reference to 'ostream' is ambiguous
verinum.h:28:7: error: candidates are: class ostream
In file included from d:\mingw\bin\../lib/gcc/mingw32/4.7.2/include/c++/ios:39:0
from d:\mingw\bin\../lib/gcc/mingw32/4.7.2/include/c++/ostream:
from d:\mingw\bin\../lib/gcc/mingw32/4.7.2/include/c++/iostream
from main.cc:42:
d:\mingw\bin\../lib/gcc/mingw32/4.7.2/include/c++/iosfwd:138:33: error:
typedef class std::basic_ostream<char> std::ostream
In file included from netlist.h:38:0,
from pform.h:22,
from main.cc:56:
and so on ...............


  • I tried to reproduce this with a fresh checkout of development head from git, but do not see this problem (I do see a problem when compiling the FST waveform dumper which I need to investigate). I have not yet tried updating my MinGW installation. Just to check:

    1) Are you using the latest development code (head of master branch in git), or are you using a development snapshot (tarball)?

    2) What MinGW gcc version are you using?

  • Cary R.
    Cary R.


    Regarding the FST issue.

    Remember that certain dumper files come directly from GTKWave and we don't change them. If you find a problem in one of these files you'll need to propose a change to Tony and then pull the change once it is in GTKWave SVN. I can help with all the back end updating if all you want to do is propose a fix.

    FYI using the latest 32-bit MinGW-w64 under Cygwin the latest development code from git will compile fine, but does not run since there is something broken in how paths are translated. I believe this is a problem in the recent MinGW/libraries/headers, but I have not had time to verify this.

  • The FST issue can be resolved by finding the lines

    #ifdef __MINGW32__
    #include <io.h>
    /* #define ftello ftell */
    /* #define fseeko fseek */

    in vpi/fstapi.c and uncommenting the two #define lines.

    With an up-to-date MinGW installation, I still can't reproduce the problem. However, rereading your error report has jogged my memory, and I think your problem is that you have git set up to use Windows style line endings. The easiest way to check if this is the case is to open one of the source files in your git checkout directory using Notepad - if it's readable, you have Windows style line endings, if it's all one line you have Unix style line endings (which is what you want).

  • Cary R.
    Cary R.

    I have pushed an update from Tony that should fix the ftello()/fseeko() compile issue for MinGW.


    If you don't see any problems I think we can close this as invalid since the original poster has not commented in a week. Can you also please double check that the line ending issue is documented on the wiki.

  • Cary R.
    Cary R.

    Martin has verified that the latest code from git compiles and runs correctly using the procedure described on the wiki, so I'm going to close this report as invalid.

  • Cary R.
    Cary R.

    • status: open --> closed-invalid