Re: [LinPrj] slm

2001-12-27 Thread Kohn Emil Dan

Is this a side efect of the Turkish patch !??
;-
Emil


On Fri, 28 Dec 2001, Ahmet Olmez wrote:

 ###
 BU MESAJ BIR SPAM DEGILDIR#
 ##
 http://alisverissehri.kolayweb.com
 http://alisverissehri.kolayweb.com
 http://alisverissehri.kolayweb.com

 Degerli ziyaretcimiz,

 Asagidaki avantajlardan yararlanmak istiyorsaniz lutfen sitemizi en kisa zamanda 
tekrar
 ziyaret ediniz.

 0. IOn (10) Oyun siparis eden ilk iki [EMAIL PROTECTED] , 
[EMAIL PROTECTED]
 adresleri hadiye!
 1. Bes (5) Oyunalana   1 Adet Fifa 2002 Hediye!
 2. On (10) Oyun alana  1 Adet Melekler Þehri 1 Adet Dikey Limit Hediye!
 3. Bes (5) Mp3 alana   1 Adet Pop þarkýlarý Mp3 Hediye!
 4. On (10) Mp3 alana   1 Adet Fifa 2002 Hediye!
 5. Bes (5) Film Alana bir adet yuzuklerin efendisi filmi hediye!
 6. CD-Rw   110 $   = zel siparis  = 40x okuma 30 x yazma  10x tekrar yazma ..



 Not:Avantajlar 1 ay icin gecerlidir. Yeni avantajlari gormek icin sitemizi sýk sýk 
ziyaret ediniz.


 Saygilar,
 Serdar Kahya
 UYARI: BU BIR SPAM DEGILDIR. Siz veya arkadasiniz bu email adresini
 http://alisverissehri.kolayweb.com adresine kayit yaptirmistir.

 -
 Haifa Linux Club Projects Mailing List (http://linuxclub.il.eu.org)
 To unsub send an empty message to [EMAIL PROTECTED]





-
Haifa Linux Club Projects Mailing List (http://linuxclub.il.eu.org)
To unsub send an empty message to [EMAIL PROTECTED]





Re: [LinPrj] r2l 0.9.1

2001-12-27 Thread Kohn Emil Dan


Hi Tzafrir
On Fri, 28 Dec 2001, Kohn Emil Dan wrote:

 Hi Tzafrir,

 There should be no problem regarding its compilation, however if you
 compile it don't remove that macro from refreshd_params.h refreshd will
 not run correctly.

What I wanted to say here was that if you compile with the macro
ABSTRACT_UNIX_SOCKET_ADDRESSES defined on platforms other than Linux 2.2+
then refreshd will compile, but it will not work.  You *must* undefine
that macro on platforms other than Linux 2.2+ !! You *may* also compile on
Linux 2.2+ with the macro undefined, but in this case it won't use the
abstract UNIX domain socket addresses. I don't think that anyone
could understand anything out of my previous reply. I hope this makes it
clearer.



 On Fri, 28 Dec 2001, Tzafrir Cohen wrote:

  On Fri, 28 Dec 2001, Kohn Emil Dan wrote:
 
   Just an aside, make sure that on other (i.e. non-Linux) platforms the
   ABSTRACT_UNIX_SOCKET_ADDRESSES macro is not defined in refreshd_params.h
  
  
 Regards,
  
 Emil
 
  I think it compiled without it.
 
  But I needed to add -lsocket to the link line .

Welcome to Solaris ;-)). Solaris being a SysV derived system uses streams
as the primary kernel communication mechanism. Sockets are implemented in
user space using the streams system calls. Any program using sockets on
Solaris must link with -lsocket. In addition almost any program using
sockets over TCP/IP must link with -lnsl. But this is not the case with
refreshd.


 
  It is on csd under /tmp/tzafrir (it should be world-readable)
 

/tmp/tzafrir is world-readable. However r2l-0.9.1 is not (no exec bit set
to the world). Anyway I looked around in the r2l-0.9.1.tar.gz file. The
macro in refreshd_params.h is defined. For Solaris, this is clearly wrong.
Thinking better, probably the default should be with that macro undefined
and adding -DABSTRACT_UNIX_SOCKET_ADDRESSES to gcc when compiling under
Linux 2.2+

Regards,
Emil



  
  
   On Fri, 28 Dec 2001, Kohn Emil Dan wrote:
  
Hi Tzafrir,
  
What problems did you experience with refreshd?
Can I help you with that?
   
   
Emil
   
   
On Fri, 28 Dec 2001, Tzafrir Cohen wrote:
   
 Another release

 This is something which is much more of a decent release. It should build
 on most standard machines now, I believe. You can get it, as usual, from:

 http://www.technion.ac.il/~tzafrir/R2L/r2l-tarball

 I removed the four separatesubpackages. There is only one package (r2l-*
 itself). THus there is much less junk there.

 One problem with building it is that the option '--prefix' of the
 configure script has no effect. If you want to build it under, say,
 /home/tzafrir , then you should run:

 make PREFIX=/home/tzafrir
   
 There are also configure options (that do work) for disabling the build of
 refreshd (--disable-refreshd) or for disabling the build of the
 r2l-plugins (--disable-plugins).

 I'm generally quite satisfied with this one(even though I have some
 problems when trying to build it on other platforms, mainly with
 refreshd, I can live with that)

 So I'd like to hear some success stories, and then get 1.0 out.

 --
 Tzafrir Cohen
mailto:[EMAIL PROTECTED]
 http://www.technion.ac.il/~tzafrir


 -
 Haifa Linux Club Projects Mailing List (http://linuxclub.il.eu.org)
 To unsub send an empty message to [EMAIL PROTECTED]



   
   
-
Haifa Linux Club Projects Mailing List (http://linuxclub.il.eu.org)
To unsub send an empty message to [EMAIL PROTECTED]
   
   
   
  
 
  --
  Tzafrir Cohen
  mailto:[EMAIL PROTECTED]
  http://www.technion.ac.il/~tzafrir
 
 
 


 -
 Haifa Linux Club Projects Mailing List (http://linuxclub.il.eu.org)
 To unsub send an empty message to [EMAIL PROTECTED]






-
Haifa Linux Club Projects Mailing List (http://linuxclub.il.eu.org)
To unsub send an empty message to [EMAIL PROTECTED]





Re: [LinPrj] r2l 0.9.4

2002-01-02 Thread Kohn Emil Dan

Hi Tzafrir,

It seems that things got messed up somehow with the tarball. I downloaded
r2l-0.9.4.tar.gz from

http://www.technion.ac.il/~tzafrir/R2L/r2l-tarball/

and there are problems compiling it under Solaris.

It seems that there is a problem with the following line in the configure
script:

  fribidi_config_version=`$FRIBIDI_CONFIG $fribidi_config_args \
 --version | sed 's/[^0-9]*\([0-9]*\).\([0-9]*\).\([0-9]*\)\..*/\1.\2.\3/'`

For some reason the fribidi_config_version variable gets messed up (it's
something like 0..)
 After I remove the sed stuff, the configuration script proceeds normally.
However when trying to compile the package

gmake PREFIX=/tmp/emild/r2l-install

the compilation aborts because r2l-config does not exist. I think that the
problem is caused by the fact that install-sh does not allow multiple
files to be copied such as:

install-sh -c file1 file2 dir

Each file must be installed separately. Therefore you should probably
replace the line in r2llib's Makefile with a for loop that installs each
file separately, i.e. replace

$(LIBTOOL) $(INSTALL) -c $(PROGS) $(CONFIG_BIN) $(BINDIR)

by

for file in $(PROGS) $(CONFIG_BIN); \
  do $(LIBTOOL) $(INSTALL) -c $$file $(BINDIR) ; \
done

The compilation proceeds after this but aborts further with
../.././install-sh command not found.

I think that this error is due to some messed up TOPDIR, but I think that
the main reason is that the tarball is somehow not up-to-date.

Also looking at the biditext.in file it seems to be the same as the old
one (i.e. no support for --auto-refresh). This ensures me more that the
problem is with the tarball version.


BTW, the version of biditext I provided will not work on Solaris because
Solaris's sh does not support the syntax

export variable=value

This means that you will have to change such lines to:

variable=value
export variable



Regards,

Emil


On Wed, 2 Jan 2002, Tzafrir Cohen wrote:

 r2l 0.9.4 is available. Main changes from 0.9.2:

 * the binary r2l is now installed (unless using a system r2llib, because
 it may have a different r2l binary)

 * refreshd_hook.so is optionally loaded when running biditext (using
 --auto-refresh)

 * Everything compiles, also on solaris.

 Things to verify:

 * refreshd on solaris: does it run?
 * Any problems left with tha system version of r2llib ?

 On Fri, 28 Dec 2001, guy keren wrote:

 
  On Fri, 28 Dec 2001, Tzafrir Cohen wrote:
 
   I'm generally quite satisfied with this one (even though I have some
   problems when trying to build it on other platforms, mainly with
   refreshd, I can live with that)
  
   So I'd like to hear some success stories, and then get 1.0 out.
 
  i am trying it out here now (redhat 6.2 with various manual changes).
 
  first, some output i see in the configure script's output on the screen:
 
  checking for gnome - version = 1.0... ./configure: test: integer
  expression expected after-lt
  yes
 
  there's some problem with the test there. i see this again later when
  checking for another component. running with '-x' reveals the problem is
  with 'gnome_wanted_micro_version' being set to an empty string, rather
  then a number,at line 835, and thus the 'test' on line 2581 yields a
  syntactic error.
 

 This should be fixed. I modified aclocal/ac_check_generic . I'm not sure
 my solution is optimal, but it seems to work. (I ignore everything after
 the third number)

  the other component with a problem is 'r2llib'. on line 1014. what happend
  is that 'r2llib-config --version' here returns '0.32', while the script
  seems to expect '0.3.2', and this the script sets
  r2llib_config_micro_version to an empty string. please note that i have a
  copy of r2llib installed in the system, and i also get a warning about it:
 
  configure: warning: Using system version of r2llib from
  /usr/local/bin/r2llib-config
 

 This may not be solved.

  another thing - when i ran 'make install', then after the installation, an
  'install' script is being built - either it should be built _before_
  installing the files, or it shouldn't be built at all - but certainly not
  be built as the last operation of 'make install'.

 Solved. I needed to use install-sh instead of install.sh . This is a weird
 interaction between autoconf and make.

 
  now, running biditext causes applications to crash during startup (tried
  with 'netscape' [which crashed] and with'gnome-terminal'. gnome-terminal
  itself didn't crash, but a lynx ran under it crashed with a bus error).
 
  i decided perhaps my local copy of r2llib is at fault, so i manually
  removed it (removing all files from /usr/local/lib/libr2l* and
  /usr/loacl/bin/r2l*, including the insalled r2l applets). when i did that,
  the configure script barfed that i don't have a version of 'r2llib', and
  then barfed that i don't have the right version - it 

Re: [LinPrj] r2l-0.9.5

2002-01-10 Thread Kohn Emil Dan



On Thu, 10 Jan 2002, Tzafrir Cohen wrote:

 On Thu, 10 Jan 2002, Kohn Emil Dan wrote:

  Hi Tzafrir,
 
  Played a little bit with the Solaris version.
 
  The old problem with the path to install-sh is still there. A short
  reminder of it:
 
  Aftyer running
  ./configure
 
  followedby:
 
  gmake PREFIX=/path/to/install

 What do you mean by path to install?

Yeah, I see the confusion I caused. Of course /path/to/install has nothing
to do with the install script/program. It's the path where the package is
supposed to be installed.

I actually run

gmake PREFIX=/tmp/emild/r2l-package

followed by

gmake PREFIX=/tmp/emild/r2l-package install



 If I want it to be installed under my home directory, I run:

 gmake PREFIX=$HOME install

 (the default is to install under /usr/local)

 
  the build process aborts when installing lib refreshd_cnt.a because the
  path to install-sh is wrong.
 

 On my account I have ginstall on the path, and it gets used. Therefore
 install-sh is unnecessary. I have no problems.

I see.
snipped
Emil


-
Haifa Linux Club Projects Mailing List (http://linuxclub.il.eu.org)
To unsub send an empty message to [EMAIL PROTECTED]