csql-development

Re: [Csql-development] CVS repository imported into git and my fixes to CSQL

[prabakaran <pra...@ya...>]

Hi Przemyslaw,
Thanks for x86_64 porting.
Shall I know more about you and your assignments?
Regards
Prabakaran
Architect-CSQL

--- On Wed, 4/15/09, Przemysław 'Przemoc' Pawełczyk <pr...@gm...> wrote:

From: Przemysław 'Przemoc' Pawełczyk <pr...@gm...>
Subject: [Csql-development] CVS repository imported into git and my fixes to CSQL
To: csq...@li...
Date: Wednesday, April 15, 2009, 11:51 PM

Hi,

I created git repository from CVS using git-cvsimport. The result is
available at:
http://repo.or.cz/w/csql.git

Why? Git is better, but the real reason is its flexibility and that i
feel more comfortable using git rather than cvs.

I also created clone of this git repository for my changes:
http://repo.or.cz/w/csql/przemoc.git

Currently there are only few commits:

840d10 -- repository and Makefile.cvs correction
086db7 -- fix for build errors
9927d6 -- fix for x86_64-specific build errors
efbe85 -- build.ksh correction

Nevertheless CSQL without them is non-compilable on Ubuntu
32-bit/64-bit, at least on my desktop computer and laptop.

So I am upstream-compliant developer, right? :)

Regards.

P.S. I am not on the mailing list.

-- 
Przemysław 'Przemoc' Pawełczyk
http://przemoc.net/

------------------------------------------------------------------------------
This SF.net email is sponsored by:
High Quality Requirements in a Collaborative Environment.
Download a free trial of Rational Requirements Composer Now!
http://p.sf.net/sfu/www-ibm-com
_______________________________________________
Csql-development mailing list
Csq...@li...
https://lists.sourceforge.net/lists/listinfo/csql-development



      

Re: [Csql-development] CVS repository imported into git and my fixes to CSQL

[Przemysław 'P. P. <pr...@gm...>]

Hi Prabakaran,

> Thanks for x86_64 porting.

Exaggerated words. I made CSQL only compilable on x86_64.
The simplest way was using CPUTYPE=i686 for x86_64 (only one change in
configure.in) and effect would be the same.
The right way is adding new CPUTYPE and handling it properly in
Mutex.* files and I've done it.
Currently defined(x86_64) alternative uses i686 code, but maybe in
future that will change somehow.

Using unsigned int for Lock type (in Mutex.h) on x86 architectures
should be reconsidered in favour of unsigned long, because long size
equals processor word size (in general, there are exceptions but not
in x86 CPUs). This change may impact (or not -- I'm not sure, because
I haven't investigate the whole source code yet) some other areas, but
fixing will be rather trivial.

> Shall I know more about you and your assignments?

I'm member of collegiate team committed to project based on CSQL and
working on Ubuntu distribution. My current task is making CSQL
"available" for the rest of the team.

I sent two new commits:
f2cf58 -- fix for GCC 4.3
8a221e -- fix for more recent autotools

They are crucial for successful build in Ubuntu 8.10. This apply to
Ubuntu 8.04 if you remove ltmain.sh (and recreate it).

Side notes: libtool generated from ./configure script with old
ltmain.sh on Ubuntu 8.10 was causing something that looked like fork
bomb. On Ubuntu 8.04 there was no problem. Strange, really strange.

Regards.

-- 
Przemysław 'Przemoc' Pawełczyk
http://przemoc.net/