[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#419570: webcalendar: Package dependencies must allow php5 instead of php4

On (17/04/07 02:27), Rafael Laboissiere wrote:
> I see a problem with the above dependencies.  Imagine the following
> combination of packages in a given system:
>     apache2 installed
>     apache, apache-ssl, and apache-perl NOT installed
>     libapache2-mod-php4 installed
>     libapache2-mod-php5 NOT installed
>     php4-mysql, php4-pgsql, and php5-pgsql NOT installed
>     php5-mysql installed
> The above satisfy the dependencies as you wrote them.  The question is:
> would php5-mysql work well with libapache2-mod-php4?

That's not your job, the php maintainers have to ensure their packages
work, you just have to make sure enough php/mysql/apache stuff is
installed to work.

If you look at php5-mysql you will see that it depends on
phpapi-20060613, which is a virtual package that is provided by


So a php5 package must be installed.

> At any rate, the problem comes from the orthogonality of the three
> "dimensions":
>    * PHP 4 vs. 5
>    * mysql vs. pgsql
>    * apache 1 vs. 2
> The problem exists already in the the current version of the package.
> Indeed, the following combination satisfy the dependencies:
>    apache2 installed
>    other apaches NOT installed
>    libapache-mod-php4 installed 
>    libapache2-mod-php4 NOT installed

libapache-mod-php4 depends on apache-common, which in turn depends on
apache, so that works ok.

You must remember that the dependencies you specify are not all of them,
there is a whole tree of dependencies being considered. I think if you
consider the tree in all cases you should find that everything that is
required is installed, if not then it is probably a bug. Some
dependencies are hard to specify in a pleasant way though, so it might
get a little ugly.

I hope this helps,



  James Westby   --    GPG Key ID: B577FE13    --     http://jameswestby.net/
  seccure key - (3+)k7|M*edCX/.A:n*N!>|&7U.L#9E)Tu)T0>AM - secp256r1/nistp256

Reply to: