The next commands are for ALT Linux, maybe they are not correct for other distributions.

1) Find what a package needs:

$ apt-cache show lyx | grep 'Depends:'
... texlive-latex-recommended ...
... /user/share/texmf ...

2) Create a fake spec-file “/home/rpmer/RPM/SPECS/dummy-provides.spec“, which “provides” the required facilities:

Name: dummy-provides
Version: 0
Release: alt2

Summary: Trick apt/rpm dependencies
License: Public domain
Group: Toys

Packager: Oleg Parashchenko 
BuildArch: noarch
Provides: /usr/bin/dvips dvipng texlive-latex-recommended /user/share/texmf

%description
%summary

%prep

%build

%install
mkdir -p %buildroot

%files

%changelog
* Fri Apr 01 2011 Oleg Parashchenko 
- initial list

3) Create a rpm-package:

$ rpm -bb dummy-provides.spec
...
Wrote: /home/rpmer/RPM/RPMS/noarch/dummy-provides-0-alt2.noarch.rpm

4) Install the package (”-ihv” to install, “-Uhv” to upgrade):

$ sudo rpm -Uhv /home/rpmer/RPM/RPMS/noarch/dummy-provides-0-alt2.noarch.rpm

Based on the discussion in the mailing list, thanks everyone for help.

Download: http://uucode.com/download/2011/12/vim-keymap-de-nn.tar.gz

keymap switch

I have no idea how to press “CTRL-^“, therefore decided to use another combination: “Shift-F1“. Add to vimrc:

map! <S-F1> <C-^>

German-Russian

Put the file “russian-jcuken-z.vim” to the directory “~/.vim/keymap“. Add to vimrc:

set keymap=russian-jcuken-z
set imsearch=-1
set iminsert=0

Only the first line is important, the rest is my preferences.

German-Arabic

Put the file “arabic_utf-8.vim” to the directory “~/.vim/keymap“.

German-something

Modify Python script “keymap-de-nn.py” to your needs.

The tools are:

* diffdelta.py: Finding which diff chunk causes an error.
* latexdd.py: Finding a minimal failing example
* stydd.py: sty-bug hunting

Step 1:

Make a copy of the directory with Portfile (located somewhere in /opt/local/var/macports/sources/rsync.macports.org/release/ports)

Step 2:

Put the file into the directory

Step 3:

Modify Portfile, add:

set orig_dir [pwd]
fetch {
  set file "PyICU-0.8.1.tar.gz"
  file mkdir $distpath
  file copy -force ${orig_dir}/$file ${distpath}/${file}
}

Instead of “PyICU-0.8.1.tar.gz” write the name of your file.

Step 4:

Install

$ sudo port install -d -v

Search in internet gave only one suggestion: use the gtk libraries, which are shipped together with vmware, not the system ones. Did not help.

But what if to force to use all the shipped libraries instead of the system ones?

Yet another log file, “appLoader-NNN.log” traces loading of libraries. A Python script to process it:

import sys, re

re_so = re.compile('LOG NOT INITIALIZED \\| (.*so)[0-9.]* <SYSTEM>')

fname = sys.argv[1]
h = open(fname)
for l in h:
  m = re_so.search(l)
  if m:
    s_so = m.group(1)
    print 'mv /usr/lib/%s* /usr/lib/tmp/' % s_so
h.close()

After processing the log, I get a file “move.sh”:

mv /usr/lib/libsexy.so* /usr/lib/tmp/
mv /usr/lib/libart_lgpl_2.so* /usr/lib/tmp/
mv /usr/lib/libXrandr.so* /usr/lib/tmp/
...

(This codes only for my backup reasons, do not even try to use it unless you understand it completely.)

With all these libraries “removed”, vware started.

Using bisection, I found that in my case the bad library was “libgiomm”.

Repair script is:

#!/bin/sh

( \
mv /usr/lib/libgiomm-2.4.so.1 /usr/lib/libgiomm-2.4.so.1.3.0 /usr/lib/tmp/; \
  sleep 5; \
mv /usr/lib/tmp/libgiomm-2.4.so.1 /usr/lib/tmp/libgiomm-2.4.so.1.3.0 /usr/lib/ \
) &
sleep 1
vmware

So far I have not found a rapair, but only a workaround:

ps2pdf -dPDFSETTINGS=/prepress print.ps print.pdf

And print the PDF-file.

* boot up: “Hello, World” example
* recursion: factorial, fibonacci, ackermann
* closures: function which return function which return function etc
* continuations: basic return example, yin-yang puzzle

Scheme code and generated PHP code:
http://github.com/olpa/schemevm/tree/milestone1/step-by-step-examples

The second use of this milestone release is a demostration how to start a backend. The package contains the minimal amount of code to compile the examples. Therefore, the code is much easier to understand then the code of a complete working backend.

I coded straightforward. If I saw a way to solve a problem, I used this way without trying alternatives. In the development process “make it work, make it work correctly, make it work fast”, I’m on the step zero “experiment if it would work”.

One of the consequences is that performance is poor. For example, Ackermann function A(3,9) is 100 times slowed as by gsi-script and 30 times slower as native PHP implementation. The output from yin-yang puzzle is 30 times slower as by gsi-script and 10 times slower as by Guile. However, I’m optimistic and think there is a lot of possibilties for optimization.

The code is developed for PHP 4.4.8 and tested under PHP 3.0.

Quite fast, I found that I want to use the tool dtrace. To my disappointment, I realized that I could not just click and run this tool, but had to learn.

Therefore, at the first I skipped the document Tracing User Processes and returned to it only after making all the exercises of a tutorial: Introduction. Finally, I managed to get what I want:

$ cat getenv.d 
pid$target:libSystem.B.dylib:getenv:entry
{
    printf("getenv: '%s'", copyinstr(arg0));
}
% sudo dtrace -s getenv.d -c my_program

Note 1: Under Mac OS X, libc seems to be libSystem.B.dylib

Note 2: According to the documentation, I could use args[] array. But for some reason, I could not. Instead, I cast the first argument (int) to a string.

Somewhere in .tex-file define the command:

\newcommand\ForceNarrowText{%
\def\addtoname"##1"{\def\NarrowFontName{"##1:extend=0.8"}}%
\expandafter\addtoname\fontname\font%
%\show\NarrowFontName
\font\NarrowFont=\NarrowFontName\relax \NarrowFont}

and then use it:

hello {\ForceNarrowText hello} hello

usb 6-1: new high speed USB device using ehci_hcd and address 62
hub 6-0:1.0: unable to enumerate USB device on port 1
hub 6-0:1.0: unable to enumerate USB device on port 1

At first, I decided that the stick had broken after years of work, but out of curiousity tried to connect it to every USB port I found. It worked. The stick worked again when plugged into the keyboard, which in term was connected to the KVM.

What is funny, I did nothing, but after a few days the stick worked again correctly in its original port. I forgot about the problem. But after yet another vacations, the problem appeared again, and again disappeared in a few days.

I have no idea why.

But I finally found a solution here: [Solved] Unable to enumerate USB device (Disabling ehci_hcd). I need to disable ehci_hcd. In my case, the commands is:

# cd /sys/bus/pci/drivers/ehci_hcd
# echo -n "0000:00:1a.7" > unbind

As a side effect, I do not have USB 2.0 on this bus anymore, but it is not a problem.

24.02.2010, update

To issue the command automatically on boot, one could edit /etc/rc.local, but in my case I need to have network loaded before. The best solution (I hope) is based on advice here:

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/354832

You can disable this on boot by creating a /etc/udev/rules.d/disable-ehci.rules file containing:

ACTION=="add", SUBSYSTEM=="pci", DRIVER=="ehci_hcd", \
        RUN+="/bin/sh -c 'echo -n %k > %S%p/driver/unbind'"

After a HDD is attached, ask Linux to re-scan the bus to find the HDD:

$ sudo sh -c 'echo "- - -" >/sys/class/scsi_host/host0/scan '

or without sudo:

# echo "- - -" >/sys/class/scsi_host/host0/scan

To detach the HDD (umount it first!) (in this case, the hdd is sdb):

$ sudo sh -c "echo 1 > /sys/block/sdb/device/delete "

or without sudo:

# echo 1 > /sys/block/sdb/device/delete

Well, I actually don’t need hotswap, manual swap is ok for me. The problem is solved.

set encoding=utf-8
set guifont=Courier_New:h12
set guifontwide=Arial_Unicode_MS:h12

The built-in method to input diacritical marks is to use vim digraphs, Ctrl-K + symbol1 + symbol2. The full list of them can be displayed using :digraphs command.

This is LuaTeX, Version beta-0.40.6-2009100220 (TeX Live 2009)
(Fatal format file error; I'm stymied)

There is no README. No installation instructions. No FAQ entry. Just a binary file.

Below is my way to run a new luatex without changing the working installation of TeX. First of all, I need the textool.sh wrapper to run TeX tools from non-standard locations, decribed here.

The next step is to generate a new luatex format:

$ ./textool.sh ./luatex --ini luatex-plain.tex
$ mv luatex-plain.fmt luatex.fmt

Finally, the file compiles:

$ ./textool.sh ./luatex LuaTUGboat.tex

As for luatex, context etc, I think the approach is the same. I don’t need these formats (yet), therefore haven’t tried to find details. They are welcome as comments.

For closures, CPSCM relies on the host language. It works well for Scheme, Lisp and JavaScript backends, but no luck for PHP.

function getMyFunc($a){
  return create_function('$b', 'print "a='.$a.', b=$b";');
}
$f1 = getMyFunc(1);
$f1(2);

The code above and a few tricks to emulate lambdas do not work in PHP 4. As I was told on Stackoverflow, “…some changes have been made starting with PHP 5.3, but even there you don’t have your expected variable scopes that would allow you to do what you have in your examples.”

http://stackoverflow.com/questions/4155254/anonymous-functions-lambdas-closures-in-php-4

Anyway, despite of dropping, I like CPSCM.

18.11.2010, update

The same problem is with Chicken. The final winner is Gambit.

The first task was easy, all I needed is to carefully to change the templates and the source code to remove all the dynamic elements from the forum pages. Then mirrored:

wget -r -np -m -p --wait=5 -o log http://mysite/forum2/

I checked the “log” file to make sure that all the pages downloaded ok:

grep '^   ' log  | grep -v '     0K'

Then I manually looked at the set of the downloaded files. I made several edit-mirror-check cycles before the result was ok.

After I uploaded the static copy, I got a nasty surprise. The names of files are:

viewforum.php?f=1
viewtopic.php?t=378&start=15
...

The web server (apache) can’t deliver such files because of the question sigh, which is reserved to separate the file name and the query. How to stop this correct behaviour?

The task is challenging. My answer is:

* Use mod_rewrite to change the question sign to the underscore sign
* Correspondingly, rename the files: “viewforum.php?f=1″ to “viewforum.php_f=1″, “viewtopic.php?t=378&start=15″ to “viewtopic.php?t=378&start=15″ etc.

The .htaccess is:

RewriteEngine On
RewriteCond %{ENV:REDIRECT_STATUS} =""
RewriteCond %{QUERY_STRING} !=""
RewriteRule ^(.*)$ $1_%{QUERY_STRING} [L]

ForceType "text/html; charset=iso-8859-1"

Everything is obvious except the line 2. After applying the rewrite rule, the sever repeats the rewriting process and can fall into a loop. The REDIRECT_STATUS-line detects the second pass and prevents the mod_rewrite from looping.

It was hard to construct this line. The documentation on mod_rewrite does not describe “REDIRECT_STATUS”. Thanks log level 9 for explaining the problem, Google for relevant search and Internet for storing knowledge.

As for rewriting the file names, different approaches are possible. I used Perl.

find forum2/ -name '*\?*' >flist
cat flist | perl -nle '$a=$_; s/\?/_/; rename $a,$_;'