emacs-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: MS-Windows build using Posix configury


From: Dmitry Gutov
Subject: Re: MS-Windows build using Posix configury
Date: Fri, 17 May 2013 01:02:53 +0400
User-agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:17.0) Gecko/20130509 Thunderbird/17.0.6

On 16.05.2013 21:54, Eli Zaretskii wrote:
Date: Thu, 16 May 2013 19:17:04 +0200
From: Dani Moncayo <address@hidden>
Cc: Eli Zaretskii <address@hidden>, Emacs development discussions 
<address@hidden>

"/usr/local" is a valid directory specification in MSYS, so I don't
see why the installation should fail.   Omitting the --prefix in "make
install" should simply install Emacs under that directory.

Except that /usr/local is mapped by MSYS to something like
C:\MSYS\local or some such, not something Windows users who aren't
used to MSYS would expect.  After "make install", a naive user might
not be able to find the installation, especially if the sources are
not on drive C:.  That's why I wrote in INSTALL.MSYS that using the
default is not recommended.

I see. That's okay, then, thanks for the explanation.

Now that I got to compilation, "make -j2" freezes the process soon after start (I've tried it several times), here's the whole log:

$ make -j 2
[ -r "/c/Users/gutov/vc/emacs-bzr/trunk/src/config.in" ] || ( cd /c/Users/gutov/vc/emacs-b
zr/trunk && autoheader )
cd nt && make all - --jobserver-fds=4,5 -j                         \
CC='gcc -std=gnu99' CFLAGS='-g3 -O2 -gdwarf-2' CPPFLAGS='-mtune=pentium4 -DUS
E_CRT_DLL=1 -I /c/Users/gutov/vc/emacs-bzr/trunk/nt/inc' \
          LDFLAGS='' MAKE='make'
cd lib && make all - --jobserver-fds=4,5 -j                         \
CC='gcc -std=gnu99' CFLAGS='-g3 -O2 -gdwarf-2' CPPFLAGS='-mtune=pentium4 -DUS
E_CRT_DLL=1 -I /c/Users/gutov/vc/emacs-bzr/trunk/nt/inc' \
          LDFLAGS='' MAKE='make'
make[1]: Entering directory `/c/Users/gutov/vc/emacs-bzr/trunk/nt'
gcc -std=gnu99 -mtune=pentium4 -I. -I/c/Users/gutov/vc/emacs-bzr/trunk/nt -mtune=pen tium4 -DUSE_CRT_DLL=1 -I /c/Users/gutov/vc/emacs-bzr/trunk/nt/inc -g3 -O2 -gdwarf-2 /c/U
sers/gutov/vc/emacs-bzr/trunk/nt/runemacs.c emacs.res -mwindows \
         -o runemacs.exe
make[1]: Entering directory `/c/Users/gutov/vc/emacs-bzr/trunk/lib'
rm -f alloca.h-t alloca.h && \
        { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \
          cat /c/Users/gutov/vc/emacs-bzr/trunk/lib/alloca.in.h; \
        } > alloca.h-t && \
        mv -f alloca.h-t alloca.h

Is that what is meant by not "supporting parallel execution"?

Non-parallel "make" starts differently (!) and completes the build fine.

I think someone mentioned that this way of building would be faster, but configuration takes longer than with configure.bat, and compilation seems to take about the same amount of time.

configure.bat took ~15 seconds, msysconfig.sh takes a couple of minutes, the following compilation is about the same, I think.
autogen.sh takes some time, too.

Speaking of stability, I've seen it crash a few times during exit, around the time it complains about inability to write ACL. Haven't been able to reproduce this under gdb, so far.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]