Hi Steffen, Andreas, Thanks for your views on Steffen's solution and my taking it up. As you may have seen in the commits I went for Andreas' way (even before I read Andreas' reply, actually). Using chmod before dh_install didn't work, Lintian told me that extIDS.pl wasn't executable. I'm not sure why this is, maybe because of a similar reason as why automake doesn't install 'data files' with 755 permissions. Anyway, I used the suggested bowtie example rule and it works ok on my machines. It turned out to be much simpler than I expected. And definitely 'cleaner'/'nicer'. I'll update the d/changelog to prepare for probabel_0.4.1-2. Thanks for your help and insights. Lennart. On 01-01-14 23:07, "Steffen Möller" wrote: > Hi all, > > >> >> On Wed, Jan 01, 2014 at 08:44:20PM +0100, L.C. Karssen wrote: >>>> The easiest fix coming to my mind is possibly to >>>> chmod src/extIDS.pl >>>> before dh_install is invoked. >>>> >>> >>> Indeed, nice and simple. >> >> I agree that this is nice and simple but I personally would not do it >> in this cheap way: >> >> 1. It is changing the upstream source tree which I would generally >> avoid (even in those slight cases) > > This is an excellent point. A single chmod on the source tree basically > interferes with what we manage with git. With me mostly using svn-buildpackage, > where we store the debian folder only, this would have been acceptable. > To not infer with the upstream sources, we use the quilt patch system. > However, this also changes the sources - it just also unchanges them. > So, it is not the change of the source tree itself that cumbersome. To save > my suggestion, you would undo the chmod after running dh_install, as in > > override_dh_install: > chmod +x src/whatever.pl > dh_install > chmod -x src/whatever.pl > > Hm. I admit, not so cute, but idempotent. > >> 2. You need to remind this change for new upstream versions (except >> you can convince upstream to do it in their source distribution) > > Er, no, this is not how it was meant. The chmod would have been part of > debian/rules, so it would have been fine with updates. > >> I'll leave it to you what solution you prefer - just want to make sure >> you understand its consequences. > > Same here :) > > Steffen > > -- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* L.C. Karssen Utrecht The Netherlands lennart@karssen.org http://blog.karssen.org GPG key ID: A88F554A -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-
Attachment:
signature.asc
Description: OpenPGP digital signature