[Slackbuilds-users] dependencies

[Martin Lefebvre]

In my opinion, the README file's purpose is to be read, not parsed.

It is my understanding that the Slackware devs and the community believe 
that the best dependency resolution tool is the user building the 
application, and that we at SBo, as part of the community, also believe it 
that argument. resolving dependencies automatically is just not the 
Slack way...

in addition to that, the content of the readme is what the website shows 
when you browse the package repo, so it's not like you have to go out of 
your way to read it...

also, if let's say I don't have X on my machine, many apps will still 
build properly, thanks to autotools and ./configure. But if there is a 
mechanism in place to resolve deps, I don't wanna have to start 
modifying scripts everytime I want to build something
so that it doesn't bug down because X is not there. Imagemagick is an 
example of an application that will build wether X is present or not.

On Fri, 2 Mar 2007, Deak, Ferenc wrote:

> Hi list members,
>
> I know this is FAQ no 14, and I know all the technical and political issues
> of the
> "dependecy problem", but...
>
> The current rule of this slackbuild project is that we can put the
> "informal"
> list of dependencies into the README file.
>
> Supposing that most of us have a full slackware install on our machines,
> is it a very bad idea that a package lists those dependencies formally which
> are
> solved by the slackbuilds project?
> For example we can put "SBODEPS=pkg1,pkg2" os something similar to the .info
> file.
>
> The reason is that I would like to write a script to
> 1. to compile an application automatically, building its sbo deps before
> 2. I would like to write a script which checks which packages have to be
> rebuilt, when
>  an other package releases a newer version. I know this later is harder
> because I have
>  to know that I really need a rebuild or not. (when I have a newer .so lib
> it does not
>  necessary need to rebuild of its dependant)
>
> So the first target is 1.
>
> If we do not want to put it to the .info the second choice is to put it to
> the README as now,
> but in a standard form (in a separate line by a leading token) to able to
> parse by grep or
> similar tools.
>
> Sorry if it was discussed later!
>
> thanks for your tolerance,
> feco
>
>
>