Cross compiling
[email protected] Tue, 29 Jan 2013 12:12:06 +0100
| Newsgroups | gmane.comp.gnu.gnucap.general |
|---|---|
| Message-ID | <[email protected]> |
Hey guys, I've been cross compiling gnucap from Linux to Windows using mingw32. I had to make some changes to the source code (adding few #include <ctime> for instance) and some other changes to the Make2.mingw32 to make it work. I'm going to document how i did this for a "personal" use. Would you like me to send this back to you so you can consider what i had to do for a future change ? Or should i make a wiki page about how I cross compiled it ? Anyway I have a wierd result which will be kinda hard to explain, so please bear with me. #################### I was using a version of gnucap I don't know anything about (compilation wise). I just happened to have the .exe file (GNUCAP 2009.12.07 RCS 26.136). When I use this version and displaying result on a big circuit i have this behaviour : assume i'm doing a tran 1 1 and print all nodes : the "#time ..." line is printed really fast then it poses a bit. the " 0. ..." line is printed really fast then it poses a bit. the " 1. ..." line is also printed really fast then i go back to the gnucap prompt. tran time : 1.83 status says : queue : 0.01 lu : 1.67 back : 0.05 output : 0.05 overhead : 0.05 others are at 0 #################### So now I'm using my cross-compiled version (GNUCAP 2009.12.07 RCS 26.136) (anyhow i still have the same result, even using the native gnucap launcher on linux or gnucap.exe in wine or a version I compiled directly into windows.) still doing tran 1 1 and print all nodes : all the lines get printed in a single row and it takes time to get displayed. total time : 7.76 status says : lu = 1.54 output = 3.63 overhead = 2.53 other are 0 or close to it. #################### Would anyone happen to have an idea about why this happens and maybe how to solve this ! And as always : Thank you for your answer !