Hi,

I've been spending some time with studying a bunch of SlackBuild scripts. I
don't quite understand what that "stripping" is exactly about? What is it?
What do I need it for? Is there a different stripping process for binaries
(well, things that go in /usr/bin) and libraries (things that go
in /usr/lib)?

Any newbie-friendly doc about this?

Niki Kovacs
--
I'm not as think as you stoned I am.

Niki Kovacs <mickey@mouse.com> wrote:
> I've been spending some time with studying a bunch of SlackBuild scripts. I
> don't quite understand what that "stripping" is exactly about? What is it?

After compilation of an executable the file contains text strings which
describe the names of the functions and variables. Having such information
in executable and libraries makes it easier to reverse-enginer or debug
them. On the other hand, this information also takes some space.

> What do I need it for? Is there a different stripping process for binaries
> (well, things that go in /usr/bin) and libraries (things that go
> in /usr/lib)?

The command to strip the binaries are simply called "strip". There is a
man-page for that command.

To save some space it is usually a good idea to strip binaries and
libraries at installation. However, there are some exceptions. X is one
such exception where stripped drivers will not work.

How are things going with your transcode SlackBuild script? I have started
working on a Makefile framework for installing software with complex
dependencies. It is being published at http://makepack.sourceforge.net/
however, at the time of this writing it is only a framework. I haven't
populated it with rules for so many applications yet and documentation
remains.

regards Henrik
--
The address in the header is only to prevent spam. My real address is:
hc7(at)uthyres.com Examples of addresses which go to spammers:
root@variousus.net root@localhost

Henrik Carlqvist wrote :

> To save some space it is usually a good idea to strip binaries and
> libraries at installation.

When it comes to libraries its a good idea to use the '--strip-unneeded'
option or else you might break shared libraries. This option will leave
the relocation symbols untouched.
--
Thomas O.

This area is designed to become quite warm during normal operation.

Henrik Carlqvist wrote:

> How are things going with your transcode SlackBuild script? I have started
> working on a Makefile framework for installing software with complex
> dependencies. It is being published at http://makepack.sourceforge.net/
> however, at the time of this writing it is only a framework. I haven't
> populated it with rules for so many applications yet and documentation
> remains.

I've started recently to make my first SlackBuilds. While searching for
documentation, I stumbled over this: http://hitek.monespace.net. A french
Slackware fanatic who writes loads of SlackBuild scripts. I contacted the
guy, and we chatted a lot, so I could have a good grasp of the question.

Essentially, he explained me why it is a lot better to build packages as a
normal user, eventually using Debian's fakeroot. So, when you do 'make
install DESTDIR=/tmp/build' or whatever, if it doesn't work out, you don't
end up with a messy install. You simply get a 'Permission denied', and you
try something else. I tried this with about a dozen packages, and I'm
slowly getting accustomed to it.

Maybe in a few week's time, I'll move on to the more complex packages like
mplayer, ffmpeg and transcode. But I prefer to handle dependencies
manually.

Check out http://slackware.tuxfamily.org, I created some space for a
SlackBuild repository. I like the idea behind slacky.it, though too many of
their SlackBuild scripts are full of errors and simply don't work.

You'll hear from all this soon.

Cheers,

Niki Kovacs
--
I'm not as think as you stoned I am.

Niki Kovacs <mickey@mouse.com> wrote:
> Check out http://slackware.tuxfamily.org, I created some space for a
> SlackBuild repository.

It looks like a very good start. In the README you say that you don't put
the sources on your web-site as you only have a modem connection and only
have limited space on the server. You also say that you don't have any
automatic routine to use wget.

If you would like you could borrow my automatic wget routine. It does the
following:

1) It pings all the servers.

2) It sorts the servers so the quickest to reply to ping comes first.

3) It downloads until wget ruturns true.

4) (optional) it checks an md5sum of the downloaded file and redo the
download from the next server if md5sum fails.

The above functionality is written in a Makefile, but it shouldn't be hard
to translate to a shell script as it basically consists of shell commands.

If you don't want to add the functionality to your script but still want
to borrow the functionality from my Makefile you could get the
functionality by writing a rule file which contains the name of the
package and the directory part of the URLs. With such a rule file you will
be able to use my Makefile to do "make all download" to get all the
sources or "make speex download" to only get the source of speex.

regards Henrik
--
The address in the header is only to prevent spam. My real address is:
hc7(at)uthyres.com Examples of addresses which go to spammers:
root@variousus.net root@localhost

Henrik Carlqvist wrote:

> It looks like a very good start.

Thanks. See my post "looking for contributors". Are you interested? I think
it's more fun to work in a team on this...

Cheers,

Niki
--
I'm not as think as you stoned I am.

Henrik Carlqvist wrote:

> If you would like you could borrow my automatic wget routine. It does the
> following:
>
> 1) It pings all the servers.
>
> 2) It sorts the servers so the quickest to reply to ping comes first.
>
> 3) It downloads until wget ruturns true.
>
> 4) (optional) it checks an md5sum of the downloaded file and redo the
> download from the next server if md5sum fails.

A first version of the scripts had a wget routine in it (not as
sophisticated as yours, though, only with one URL). I erased it... but
thinking of it, since other people will maybe use the scripts, I guess I'll
put the wget routine back. Yours looks more sophisticated and powerful, so
I'd prefer that.

Niki
--
I'm not as think as you stoned I am.

On Mon, 9 Jan 2006, Niki Kovacs wrote:

> Henrik Carlqvist wrote:
>
>> If you would like you could borrow my automatic wget routine. It does the
>> following:
>>
>> 1) It pings all the servers.
>>
>> 2) It sorts the servers so the quickest to reply to ping comes first.
>>
>> 3) It downloads until wget ruturns true.
>>
>> 4) (optional) it checks an md5sum of the downloaded file and redo the
>> download from the next server if md5sum fails.
>
> A first version of the scripts had a wget routine in it (not as
> sophisticated as yours, though, only with one URL). I erased it... but
> thinking of it, since other people will maybe use the scripts, I guess I'll
> put the wget routine back. Yours looks more sophisticated and powerful, so
> I'd prefer that.......

heh i just add near x.SlackBuild script x.download which downloading
sources and no more
be easy =)

Niki Kovacs <mickey@mouse.com> wrote:
> See my post "looking for contributors". Are you interested? I think
> it's more fun to work in a team on this...

At this moment I don't think that I will start writing complete SlackBuild
scripts also as I instead will be working on adding rules to my Makefile
concept. However, I would be happy to share ideas and script code as we
basically have the same goal with our projects.

The SlackBuild approach is a little bit more "serious" way to make
Slackware packages. My Makefile concept use the checkinstall approach
instead which could be considered the "amateur" way of creating packages.
Both approaches have their advantages and disadvantages.

regards Henrik
--
The address in the header is only to prevent spam. My real address is:
hc7(at)uthyres.com Examples of addresses which go to spammers:
root@variousus.net root@localhost

Henrik Carlqvist wrote:

> At this moment I don't think that I will start writing complete SlackBuild
> scripts also as I instead will be working on adding rules to my Makefile
> concept. However, I would be happy to share ideas and script code as we
> basically have the same goal with our projects.
>
> The SlackBuild approach is a little bit more "serious" way to make
> Slackware packages. My Makefile concept use the checkinstall approach
> instead which could be considered the "amateur" way of creating packages.
> Both approaches have their advantages and disadvantages.

Writing a SlackBuild script is far less work than expected. Once you built a
template, you can copy/paste most of it to any new script in a matter of
minutes.

I'm very much interested by your wget code snippet for retrieving the
fastest download location. Could you send one to me please? I'll include it
in the scripts, because I think it's basically a good idea, and makes
package building from the scripts much easier.

Thanks,

Niki
--
I'm not as think as you stoned I am.