296. Building Wine 1.5.19 on Debian (Wheezy/Testing) without errors

UPDATE 16 May 2013: See here for Wine 1.5.30: http://verahill.blogspot.com.au/2013/05/416-wine-1530-in-chroot.html

UPDATE (10th Jan 2013): See here for Wine 1.5.21 using the multiarch approach: http://verahill.blogspot.com.au/2013/01/308-compiling-wine-1521-on-debian.html

This time we'll be cheating and using the carbon-dev.org debian rules to build wine 1.5.19, which works amazingly well with a tiny bit of editing. In spite of what it looks like, this build is easy.

This is also my first encounter with multiarch since libgsm was causing issues (think it's moved out of ia32). Ultimately, multiarch will be the proper way to build 32-bit wine anyway, but I'll deal with that some other day.

Interesting fact: if you try to build with ... 1>build.log 2> build.err the build will fail since files are changing in the main directory during build. Make sure that you do e.g. .. 1> ../build.log 2> ../build.err if you want to track the build.

Finally, I've also built wine 1.5.18 this way, in addition to wine 1.5.19.

The build starts here:

Here's my current best guess at dependencies (Note that multiarch on ubuntu is a little bit different from Debian.):

sudo dpkg --add-architecture i386
sudo apt-get update
sudo apt-get install ia32-libs ia32-libs-dev bison flex gcc libc6-dev libfontconfig-dev libfreetype6-dev libglu-dev libgsm1-dev libice-dev libjpeg-dev libldap-dev libmpg123-dev libncurses5-dev libopenal-dev libpng-dev libsm-dev libssl-dev libusb-dev libx11-dev libxcomposite-dev libxcursor-dev libxext-dev libxi-dev libxinerama-dev libxml2-dev libxrandr-dev libxrender-dev libxslt-dev libxt-dev libxxf86vm-dev make libcapi20-dev liblcms-dev libsane-dev libhal-dev libdbus-1-dev valgrind prelink libcups2-dev opencl-dev lib32opencl1 oss4-dev gettext lib32v4l-dev lib32ncurses5-dev lib32asound2-dev libtiff4-dev libgphoto2-2-dev libxkbfile-dev libxxf86dga-dev freeglut3-dev unixodbc-dev gcc-multilib
sudo apt-get install libgsm1:i386

Reading package lists... Done
Building dependency tree       
Reading state information... Done
The following extra packages will be installed:
  gcc-4.7-base:i386 libc6:i386 libc6-i686:i386 libgcc1:i386
Suggested packages:
  glibc-doc:i386 locales:i386
The following NEW packages will be installed:
  gcc-4.7-base:i386 libc6:i386 libc6-i686:i386 libgcc1:i386 libgsm1:i386
0 upgraded, 5 newly installed, 0 to remove and 2 not upgraded.
Need to get 5,412 kB of archives.
After this operation, 11.9 MB of additional disk space will be used.
[..]

Once you've got that sorted, it is time to download the sources:

mkdir -p ~/tmp/wine_1.5.19_carbon/
cd ~/tmp/wine_1.5.19_carbon/
wget http://mirrors.ibiblio.org/wine/source/1.5/wine-1.5.19.tar.bz2 -O wine-unstable_1.5.19.orig.tar.bz2
tar xvf wine-unstable_1.5.19.orig.tar.bz2
cd wine-1.5.19/
wget http://dev.carbon-project.org/debian/wine-unstable/wine-unstable_1.5.5-0.1.debian.tar.bz2
tar xvf wine-unstable_1.5.5-0.1.debian.tar.bz2
rm wine-unstable_1.5.5-0.1.debian.tar.bz2

We could do the editing the proper way, or the quick way. Since I'm not really that familiar with build debian packages the right way (I cheat using checkinstall) we're doing this the quick and dirty way:

sed -i 's/1.5.5/1.5.19/g' debian/changelog
sed -i 's/\-4.5//g' debian/rules
sed -i 's/\-4.5//g' debian/control
sed -i 's/\-4.5//g' debian/control.in

Next, edit debian/patches/series and change it from

  1 debian-changes-from-1.1.32-1.patch
  2 readd_xpm.patch
  3 function_grep.patch
  4 Bug#29669_proposed-fix.patch
  5 fix-winegcc-paths.patch
  6 Bug#28898_squashed-proposed-patches.patch
  7 Bug#28201_proposed-fix-modified.patch

to

  1 debian-changes-from-1.1.32-1.patch
  2 readd_xpm.patch
  3 function_grep.patch

And build:

time dpkg-buildpackage -uc -us

The entire build takes about 40 minutes and gives the following files:

../libwine-alsa-unstable_1.5.19-0.1_amd64.deb
../libwine-dbg-unstable_1.5.19-0.1_amd64.deb
../libwine-ldap-unstable_1.5.19-0.1_amd64.deb
../libwine-sane-unstable_1.5.19-0.1_amd64.deb
../libwine-bin-unstable_1.5.19-0.1_amd64.deb
../libwine-dev-unstable_1.5.19-0.1_amd64.deb
../libwine-openal-unstable_1.5.19-0.1_amd64.deb
../libwine-unstable_1.5.19-0.1_amd64.deb
../libwine-capi-unstable_1.5.19-0.1_amd64.deb
../libwine-gl-unstable_1.5.19-0.1_amd64.deb
../libwine-oss-unstable_1.5.19-0.1_amd64.deb
../wine-bin-unstable_1.5.19-0.1_amd64.deb
../libwine-cms-unstable_1.5.19-0.1_amd64.deb
../libwine-gphoto2-unstable_1.5.19-0.1_amd64.deb
../libwine-print-unstable_1.5.19-0.1_amd64.deb
../wine-unstable_1.5.19-0.1_amd64.deb

To install,

cd ../
sudo dpkg -i *.deb

And you're done.
To confirm what version you're using (and get a screenshot like above) do

winecfg

and click on 'about'.

---

Links to this post:
http://www.debian-srbija.iz.rs/p/kako-da.html

295. Patching kernel 3.7.1 to fix the azx_runtime_suspend bug

(For a general approach to compiling kernel 3.7,see http://verahill.blogspot.com.au/2012/12/compiling-kernel-37-on-debian.html )

Note that it is NOT necessary to patch version 3.7.2

If you're seeing the following error messages during boot of kernel 3.7.0:

azx_runtime_suspend
pci_pcm_runtime_suspend
snd_hda_intel returns -11

you will want to patch your kernel. Technically I should be patching 3.7, but we might as well go straight for kernel 3.7.1.

So, here's how to patch it:

wget http://www.kernel.org/pub/linux/kernel/v3.0/linux-3.7.1.tar.bz2
tar xvf linux-3.7.1.tar.bz2
cd linux-3.7.1/
cat /boot/config-`uname -r`>.config
make oldconfig
wget https://patchwork.kernel.org/patch/1865521/raw/
mv index.html hda_intel.patch
patch -i hda_intel.patch -p 1

patching file sound/pci/hda/hda_intel.c
Hunk #1 succeeded at 2557 (offset -134 lines).
Hunk #2 succeeded at 2571 (offset -134 lines).

make-kpkg clean
time fakeroot make-kpkg -j3 --initrd --revision=3.7.1 --append-to-version=-amd64 kernel_image kernel_headers

Once building is complete

mv ../*3.7.1*.deb .
sudo dpkg -i *.deb

and reboot.

Tried it and it works.

294. Bruker 1D processing using octave/matlab

I wanted a set of scripts that behaved a little bit like the commands in bruker xwin-nmr/topspin, so that I could do some quick processing for visual inspection without having to do too much coding.

I also wanted some simple modules that I can plug into automated processing routines for large numbers of spectra (e.g. when doing kinetics).

So here are a few simple octave routines which should work in matlab as well. They won't change the world, but should be good enough for some basic 1D processing.

Because of the groupdelay (GRPDLY) used in by Bruker (see e.g. here (16th of June post) and here), you need to use the bruk2ana converter. There's little science behind the values which are applied since they are hardware specific.

Example workflow:

./bin2ascii experiment_1/1 fid
getpar experiment_1/1 my.par

octave:1> [fid,pars]=loadfid('fid.ascii','my.par');
octave:2> [zfid,pars]=zf(fid,pars);
octave:3> [test,phc1]=bruk2ana(zfid,pars);
octave:4> test=em(test,0.5);
octave:5> plot(test(:,1),test(:,3));
octave:6> spectrum=ft(test,pars);
octave:7> phased=apk(spectrum,phc1);
octave:8> final=absd(spectrum);
octave:9> pltspec(final)

---

Linux shell-scripts

---

bin2ascii

#!/bin/bash
#bin2ascii dir fid
cp $1/$2 $1/$2.bak
ls $1/$2.bak | cpio -o | cpio -i --swap -u
od -An -t dI -v -w8 $1/$2.bak| gawk '{print NR,$1,$2}' > $2.ascii

getpars

 #!/bin/bash
 #getpars $1 $2
 # $1 is the location (directory or .) and $2 is the root of the output file name
 SW=`cat $1/acqus | grep 'SW_h' | sed 's/\=/\t/g' | gawk '{print $2}'| tr -d '\n'`
 TD=`cat $1/acqus | grep 'TD=' | sed 's/\=/\t/g' | gawk '{print $2}'| tr -d '\n'`
 O=`cat $1/acqus | grep '$O1=' | sed 's/\=/\t/g' | gawk '{print $2}'`
 SFO=`cat $1/acqus | grep 'SFO1=' | sed 's/\=/\t/g' | gawk '{print $2}'`
 DECIM=`cat $1/acqus | grep 'DECIM=' | sed 's/\=/\t/g' | gawk '{print $2}'`
 DSPFVS=`cat $1/acqus | grep 'DSPFVS=' | sed 's/\=/\t/g' | gawk '{print $2}'`
 echo $SW > $2
 echo $TD >> $2
 echo $O >> $2
 echo $SFO >> $2
 echo $DECIM >> $2
 echo $DSPFVS >> $2

---

Octave scripts

---

loadfid.m

function [fid,pars]=loadfid(infile,parfile)
%%Usage: [fid,pars]=loadfile(infile,parfile)
%%reads a pts, re, im ascii array 
%%generated by bin2ascii
%%and a parfile generated with genpar
 fid=load(infile);
 pars=load(parfile);
 t=linspace(0,(1/(pars(1)/(pars(2)/2))),pars(2)/2);
 fid=[t' fid(:,2) fid(:,3)];
end

zf.m

function [newfid,pars]=zf(fid,pars)
%% Usage:[newfid,updatedpars]=zf(fid,pars)
%% Doubles the number of points by zerofilling
 dims=size(fid);
 newfid=[fid' zeros(3,dims(1))]';
 pars(2)=pars(2)*2;
end

bruk2ana

function [fid,phc1]=bruk2ana(fid,pars)
%% Usage: [fid,phc1]=bruk2ana(fid,pars)
%% where phc1 is the first-order phase correction
%% Using https,//nmrglue.googlecode.com/svn-history/r44/trunk/nmrglue/fileio/bruker.py
%% and https,//ucdb.googlecode.com/hg/application/ProSpectND/html/dmx_digital_filters.html
%% The short version is: bruker fid data needs pre-processing and it's hardware dependent

%%D contains the digital filter parameters
D=[[10,2, 44.75];
[10,3, 33.5];
[10,4, 66.625];
[10,6, 59.083333333333333];
[10,8, 68.5625];
[10,12, 60.375];
[10,16, 69.53125];
[10,24, 61.020833333333333];
[10,32, 70.015625];
[10,48, 61.34375];
[10,64, 70.2578125];
[10,96, 61.505208333333333];
[10,128, 70.37890625];
[10,192, 61.5859375];
[10,256, 70.439453125];
[10,384, 61.626302083333333];
[10,512, 70.4697265625];
[10,768, 61.646484375];
[10,1024, 70.48486328125];
[10,1536, 61.656575520833333];
[10,2048,70.492431640625];
[11,2, 46.];
[11,3, 36.5];
[11,4, 48.];
[11,6, 50.166666666666667];
[11,8, 53.25];
[11,12, 69.5];
[11,16, 72.25];
[11,24, 70.166666666666667];
[11,32, 72.75];
[11,48, 70.5];
[11,64, 73.];
[11,96, 70.666666666666667];
[11,128, 72.5];
[11,192, 71.333333333333333];
[11,256, 72.25];
[11,384, 71.666666666666667];
[11,512, 72.125];
[11,768, 71.833333333333333];
[11,1024, 72.0625];
[11,1536, 71.916666666666667];
[11,2048, 72.03125];
[12,2, 46. ];
[12,3, 36.5];
[12,4, 48.];
[12,6, 50.166666666666667];
[12,8, 53.25];
[12,12, 69.5];
[12,16, 71.625];
[12,24, 70.166666666666667];
[12,32, 72.125];
[12,48, 70.5];
[12,64, 72.375];
[12,96, 70.666666666666667];
[12,128, 72.5];
[12,192, 71.333333333333333];
[12,256, 72.25];
[12,384, 71.666666666666667];
[12,512, 72.125];
[12,768, 71.833333333333333];
[12,1024, 72.0625];
[12,1536, 71.916666666666667];
[12,2048, 72.03125];
[13,2, 2.75]; 
[13,3, 2.8333333333333333];
[13,4, 2.875];
[13,6, 2.9166666666666667];
[13,8, 2.9375];
[13,12, 2.9583333333333333];
[13,16, 2.96875];
[13,24, 2.9791666666666667];
[13,32, 2.984375];
[13,48, 2.9895833333333333];
[13,64, 2.9921875];
[13,96, 2.9947916666666667];];

 h=find(D(:,2)==pars(5));
 j=find(D(h,1)==pars(6));
 magickey=D(h(j),3);
 chop=floor(magickey);

 phc1=(magickey-chop)*2*pi; %the first-order phase correction gets mangled by bruker

 tmp=size(fid); %matlab workaround. rows/columns would be more elegant
 
 newfid=[fid(chop:tmp(1),2:3)' fid(1:chop-1,2:3)']';
 fid=[fid(:,1) newfid(:,1) newfid(:,2)];
 
end

em.m

function fid=em(fid,lb)
%%Usage: fid=em(fid,lb)
%%Exponential multiplication window function
%%Increases Signal-to-noise at the expense
%%of resolution
 fid(:,2)=fid(:,2).*exp(-lb.*fid(:,1));
 fid(:,3)=fid(:,3).*exp(-lb.*fid(:,1));
end

gm.m

function fid=gm(fid,lb)
%%Usage: fid=gm(fid,lb)
%%Gaussian multiplication window function
%%Increases Signal-to-noise at the expense
%%of resolution
 fid(:,2)=fid(:,2).*exp(-(lb.*fid(:,1)).^2);
 fid(:,3)=fid(:,3).*exp(-(lb.*fid(:,1)).^2);
end

de.m

function fid=de(fid,lb,gm)
%%Usage fid=de(fid,lb,gm)
%%Double-exponential window function
%%Increases resolution at the expense of
%%signal-to-noise
 at=max(fid(:,1));
 defun= @(lb,gm,t) (exp(-(t.*lb-gm*at))).^2;
 fid(:,2)=fid(:,2).*defun(lb,gm,fid(:,1));
 fid(:,3)=fid(:,3).*defun(lb,gm,fid(:,1));
end

traf.m

function fid=traf(fid,lb)
%%Usage: fid=traf(fid,lb)
%%TRAF window function
%%Increases resolution at the expense of
%%the signal-to-noise
 at=max(fid(:,1));
 traffun= @(lb,t) (exp(-t.*lb)).^2./((exp(-t.*lb)).^3+(exp(-at*lb)).^3);
 fid(:,2)=fid(:,2).*traffun(lb,fid(:,1));
 fid(:,3)=fid(:,3).*traffun(lb,fid(:,1));
end

ft.m

function spectrum=ft(fid,pars)
%%Usage: spectrum=ft(fid,pars)
%% Spectrum is a complex array with the frequency in
%%the first column and the real and imaginary parts
%%in the second column
%%pars(3)=centrefreq, pars(1)=SW
 spectrum=fftshift(fft(fid(:,3)+i*fid(:,2)));
 tmp=size(spectrum);%matlab workaround
 freq=linspace(pars(3)+pars(1)/2,pars(3)-pars(1)/2,tmp(1));
 spectrum=[freq' spectrum];
endfunction

apk.m

function spectrum=apk(spectrum,phc1)
%%Usage spectrum=apk(spectrum,phc1)
%%Spectrum is a complex matrix with
%%the frequency in the first column
%%and the complex spectrum in the 
%%second column. phc1 is the first order
%%phase correection

 tmp=size(spectrum);
 m=720;
 ph=linspace(-2*pi,2*pi,m);
 maxsig=0;k=1;
 minsig=-inf;
 for n=1:m;
  spex=real( (spectrum(:,2)).*exp(i*(ph(n)+phc1*i/tmp(1))) );
  localmin=min(spex);
        localmax=max(spex);
        if (localmin>minsig) 
                minsig=localmin;
                k=n;
        end
 end
 ph0=ph(k);
 spectrum(:,2)=spectrum(:,2).*exp(i*(ph0+phc1*i/tmp(1)));
end

altapk.m

function [spectrum,ph]=altapk(spectrum,phc0,phc1)
%%Usage -spectrum,ph]=altapk(spectrum,phc0,phc1)
%%Spectrum is a complex matrix with the frequency in the first column
%%and the complex spectrum in the second column. phc0 and phc1 are the first order
%%phase correction parameters, respectively, and are used as initial guesses.
%%This is an implementation of Chen, Weng, Goh and Garland, J. Mag. Res., 2002, 158, 164-168 and depends on entropy.m.

    ph=[phc0;phc1];
        ph=minimize("entropy",{ph,spectrum});

        %compute spectrum with optimal phase params
        pts=linspace(1,size(spectrum(:,2),1),size(spectrum(:,2),1));
        phi=(ph(1)+ph(2).*pts./max(pts))';
        spectrum(:,2)=spectrum(:,2).*exp(i*phi);
end

entropy.m

function E=entropy(ph,spectrum)
%%Used by altapk.m
    pts=linspace(1,size(spectrum(:,2),1),size(spectrum(:,2),1));
        penalty=5.53;

    phi=(ph(1)+ph(2).*pts./max(pts))';
        size(phi);
        spectrum(:,2)=spectrum(:,2).*exp(i*phi);
        R=real(spectrum(:,2));
        size(R);
    Rm=firstderiv(R);
        size(Rm);
    h=abs(Rm)/sum(abs(Rm));
        size(h);

        negs= imag((R).^(1/2));
        negs(find(negs>1))=1;
    P= @(R) penalty.*sum((negs).*R.^2);

    E=-sum(h.*log(h))+P(R);
end

absd.m

function spectrum=absd(spectrum)
%%Usage spectrum=absd(spectrum)
%%Simple (linear) baseline correction
        bsline=@(m) sum(abs(real(spectrum(:,2))-m));
        guess=0;
        p=0;
        newm=minimize(bsline,p);
        spectrum(:,2)=spectrum(:,2)-newm;
end

pltspec.m

function pltspec(spectrum)
%%Usage: pltspec(spectrum)
%%Where spectrum is a complex matrix
%%with the frequency in the first column
%%and the complex spectrum (a+i*b) in the
%%second column
 plot(spectrum(:,1),real(spectrum(:,2)))
end