On 05/06/2016 11:21, Peter Humphrey wrote:
> X-Originating-<%= hostname %>-IP: [82.69.80.10]
> 
> Hello list,
> 
> I've just about finished installing a new ~amd64 system, because it was
> getting too intricate to maintain ~arch versions of all the packages I need
> with some pretty new hardware in here. When it came to emerging libreoffice,
> though, and its last dependency app-text/libetonyek, I found a bit of a
> puzzle.
> 
> The build log of app-text/libetonyek includes this:
> 
> [...]
> checking for MDDS... no
> configure: error: Package requirements (mdds-1.0) were not met:
> 
> No package 'mdds-1.0' found
> 
> ...and on the console I see:
> 
>  * emerge --keep-going: app-office/libreoffice-5.1.3.2 dropped because it
>  * requires >=app-text/libetonyek-0.1, >=app-text/libetonyek-0.1
> 
> Now that's odd, because DEPEND in the libetonyek ebuild contains
> "dev-util/mdds:1" and see here:
> 
> $ eix mdds
> [I] dev-util/mdds
>      Available versions:  
>      (0)    0.10.3(0/0.10.3) (~)0.11.2-r1(0/0.11.2) 0.12.0(0/0.12.0) 
> 0.12.1(0/0.12.1){tbz2}
>      (1)    (~)1.2.0(1/1.2){tbz2} **9999(1/9999)
>        {doc valgrind}
>      Installed versions:  0.12.1{tbz2}(09:46:34 05/06/16) 
> 1.2.0(1){tbz2}(09:57:29 05/06/16)(-doc -valgrind)
>      Homepage:            https://gitlab.com/mdds/mdds
>      Description:         A collection of multi-dimensional data structure 
> and indexing algorithm
> 
> So, as I already have versions in slots 0 and 1, and libetonyek wants the
> one in slot 1, what is portage complaining about?
> 


Portasge is not complaining, ./configure is and you appear to be
misreading the output between what ./configure says with what eix &
ebuilds says.

This:

No package 'mdds-1.0' found

is not the same as this:

$ eix mdds
     (1)    (~)1.2.0(1/1.2){tbz2} **9999(1/9999)

You don't have mdds-1.0, you have mdds-1.2 and there is no 1.0 candidate
available for installation. So this is an ebuild bug and should be
reported at bgo.

My setup has all those packages installed since a long time, but trying
to re-install fails the same way as you. I reckon the maintainer missed
the build failure.


-- 
Alan McKinnon
alan.mckin...@gmail.com


Reply via email to