gmp refuses to build c++ headers

John Perry John.Perry at
Wed Aug 5 02:31:23 UTC 2020

I am sorry my message came across as sarcastic. That was not my intent, though upon rereading I can see how my wording could make it seem that way.

I also apologize for not stating my question clearly. I wanted to know if I was using the configure command correctly to build gmp’s C++ headers with clang++. vk answered my question in the affirmative (and did more than I had wanted).

I did find a solution: I deleted the gmp source directory, re-extracted from the archive, then configured and compiled in this new directory. This time, it compiled perfectly, with no errors. ‘make check’ passed all tests. ‘make install’ worked: I was able to compile, link, and execute two client programs.

I have since compiled successfully with g++, then again with clang++ (after a ‘make distclean’ each time). Both worked.

In the end, I have no idea what went wrong, but Torbjörn and vk did help.

john perry

> On Aug 4, 2020, at 5:28 PM, Torbjörn Granlund <tg at> wrote:
> John Perry <John.Perry at> writes:
>  I downloaded the file from GMP’s website, unpacked it, changed to the directory, then:
>     ./configure CC=clang CXX=clang++ --enable-cxx
>     make
>  Configure seemed to work fine; make failed with the given error message.
>  That’s it. I’m not sure what more I could tell you. (I thought “make” was obvious, sorry.)
> A note of sarcasm there.  How nice.
> If you know anything about programming you ought to know that
> reporoducing other people's errors is often hard.  One really needs to
> know every little detail.
> You chose to send 3 sloppy reports in quick succession, clearly
> expecting us to do some trial and error over GMP versions, operating
> systems, exact commands used.  That's a pretty arrogant attitude, even
> without the sarcasm.
> There are several helpful people on this list.  They tend to help people
> who provide information to work on.  Being sloppy and rude is not going
> to make anybody here particularly helpful.
> -- 
> Torbjörn
> Please encrypt, key id 0xC8601622

More information about the gmp-discuss mailing list