I've now spent a full day investigating why Cygwin64 segfaults when calling mpn_lshift1 or mpn_lshift2.
The function mpn_lshift1 is assembled yasm. It is linked in and the symbol exists in the library. I can link a small C program against it, which I can debug. With the gnu debugger I was able to determine that the when the call to the mpn_lshift1 function is reached, the call jumps off into an invalid address for mpn_lshift1. In other words, the function is never actually called. We may have to not support Cygwin64. I don't see how we can make any further progress with the problem. In all probability it is a linker bug in Cygwin. What is confusing me a lot is why it only happens with mpn_lshift1 and mpn_rshift1. That seems like a mighty coincidence to me. But I just can't see what else could be going wrong. Bill. On 7 March 2015 at 00:12, Bill Hart <goodwillh...@googlemail.com> wrote: > Hi Simon, > > You can keep track of the ticket here: > > https://github.com/wbhart/mpir/issues/151 > > But it won't get looked at by me for at least a month, since I don't have > internet access up to the task for at least that long, as I am on leave in > the Australia outback (hours of outages, hours of 4kb/min every day). > > Brian is familiar with the Windows assembly code. But he doesn't develop > for Windows. I don't know if he can assist you or not. I have added him in > CC, just in case he can spot something. > > Sorry I couldn't be of more immediate help. > > Best Wishes, > > Bill. > > On 6 March 2015 at 15:49, Simon Sobisch <simonsobi...@gmail.com> wrote: > >> Hi Bill, >> >> deleting these two files lead to the testuite PASS completly --> >> Workaround accepted. >> Please reply and attach the fixed asm files as soon as they're ready, I'd >> like to retest it. Just because of curiosity: please post a link to the bug >> ticket you've created. >> >> For now I'm going on testing by compiling generic and with assembler >> support with VC11 and using all of these with GnuCOBOL. I'll open a new >> thread if anything strange happens :-) >> >> Simon >> >> -- >> You received this message because you are subscribed to the Google Groups >> "mpir-devel" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to mpir-devel+unsubscr...@googlegroups.com. >> To post to this group, send email to mpir-devel@googlegroups.com. >> Visit this group at http://groups.google.com/group/mpir-devel. >> For more options, visit https://groups.google.com/d/optout. >> > > -- You received this message because you are subscribed to the Google Groups "mpir-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to mpir-devel+unsubscr...@googlegroups.com. To post to this group, send email to mpir-devel@googlegroups.com. Visit this group at http://groups.google.com/group/mpir-devel. For more options, visit https://groups.google.com/d/optout.
