Hi Thorsten, your help is much welcome, and will be especially useful to the community at large if you submit your patches upstream and get them accepted.
There is one commited change that I disagree with, and it is the disabling of the test suite. Either the architecture has no users in our field, and we can stop distributing the package on it, or it has users, and we must not knowingly distribute a package that will give them wrong results, and potentially cause them very embarassing retraction of their findings. This is very important. A lot of bioinformatics software is developed with only amd64 in mind, and it is already risky to run it on i386. We already disclaim warranty to the extent that the law allows us, but I would like to re-state that personally, I recommend to not use our bioinformatics packages on anyting else than amd64 in a professional environment. Using another architecture (such as arm) requires some extensive testing that neither us nor the upstream authors have done. I also consider that work to be outside of the scope of my duty as a maintainer. If I had to ensure seriously, that is, put my reputation as a person and as a scientist into play, that our packages are equally usable on all our release architectures, I would have the time for at maximum 10 % of the packages that I regularly upload. I also would lose the possibility to do some of my contributions from work. When I update a package on my computer for work, I share that with the world on Alioth. But there is no way that I can justify doing some work in order to build the packages on mips or sparc. Debian lives in its illusions in regard with the universality of its architectures, but in my professional enviroment, these illusions will not survive the crude examination of the facts. This is not to say that nobody should be porting these packages for fun or as part of a pioneering effort, but the momentum has to come from these people. If I request help for the porting, this sends the misleading signal that I see a value in having these binary packages built, which is not the case. It is a burden for me, and I take it by compromise, not by pleasure. Cheers, -- Charles Plessy Tsurumi, Kanagawa, Japan -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
