> On Feb. 14, 2011, 3:24 p.m., Wolfgang Rohdewald wrote: > > now I actually found a note in the wiki about a short path option in > > kdesettings. > > > > but I lost too much time with this - I do want to make sure this does never > > happen again. > > Ralf Habacker wrote: > emerge provides two strategies in thia area: > > 1. path substition by drive letters, which has to be configured in > kdesettings.bat (you already refered to this) > 2. short path usage - replace pathes by the short form. This feature is > currently package based and could be activated by setting > self.subinfo.options.useShortPathes - see EmergeBase.__adjustPath() and > where is is used for more details. > both strategies could be used together > > There are several limits in this area - nmake, mingw32-make, cmake, > command line, each one has its own different limit > > I found some related thread: > > http://lists.kde.org/?l=kde-windows&m=125717156605667&w=2 > http://www.cmake.org/pipermail/cmake/2008-August/023279.html > http://lists.kde.org/?l=kde-windows&m=124222975100566&w=2 > > From my personal experience does mingw builds have more problems as msvc > builds and the most annoying package is qt > > Ralf > >
So the most transparent solution for qt would be to useShortPathes for portage/qt if KDEROOT exceeds a length of 10 (c:\kderoot did work for me with mingw4). Would that be OK? - Wolfgang ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://svn.reviewboard.kde.org/r/6483/#review9838 ----------------------------------------------------------- On Feb. 14, 2011, 3:18 p.m., Wolfgang Rohdewald wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://svn.reviewboard.kde.org/r/6483/ > ----------------------------------------------------------- > > (Updated Feb. 14, 2011, 3:18 p.m.) > > > Review request for kde-windows. > > > Summary > ------- > > On windows, gcc/g++ only take a maximum of 8192 chars for command line > arguments. If they get more they will not complain but strange things will > happen like compiler segfaulting or gcc misinterpreting the arguments at > around position 8192. I did not report this as a gcc bug, maybe somebody else > already is registered wherever that can be done. I was using mingw4. > > So emerge.py now first checks - it allows a maximum of 10 chars like in > c:\kderoot. This results in command line lengths up to around 7700 (maybe > more, this is what I saw). > > I was using kderootsvn which makes gcc die when compiling > c:\kderootsvn\download\svn-src\qt\src\3rdparty\sqlite\sqlite3.c, the command > line having a length of 8214 > > When this review is closed, I will update the emerge wiki > > > Diffs > ----- > > /trunk/kdesupport/emerge/bin/emerge.py 1220611 > > Diff: http://svn.reviewboard.kde.org/r/6483/diff > > > Testing > ------- > > > Thanks, > > Wolfgang > >
_______________________________________________ Kde-windows mailing list [email protected] https://mail.kde.org/mailman/listinfo/kde-windows
