Daniel Tufvesson wrote:I'm really glad you guys find my work useful! It has taken fair bit of time and effort to create and maintain makeresolvedeb. I'm a big fan of Resolve and Debian so this was a good match for me.
i also like your endeavor to improve the situation for debian users. and your script looks indeed much more acceptable to me, them most of this stupid ad hock hacks, which simply advise to place silly links in the system directories, without considering their debian specific layout and its reasons, don't think about ld.so.conf requirements etc.
Daniel Tufvesson wrote:To me a .deb is the only proper way to install Resolve on a Debian based system.
yes -- that's something, i can 100% agree!
Daniel Tufvesson wrote:I'm actually surprised BMD hasn't released their own .deb yet but the current situation is actually quite good. The openness of the current Resolve installer is what makes makeresolvedeb possible. I was trying to do something similar with the Fusion installer but found no elegant way.
that's again point, which a can't agree. the usual Qt installer framework based alternatives, as for example used by nuke or sgo mistika, can be manually decomposed in a quite similar manner as the actual resolve installer (here is a practical
example). but in practice this other kind of installer looks much more user friendly. and as much as i prefer native debian packages, i still have emphasize the fact, that Qt installers are at least compatible to automatic package generation via
CheckInstall, which unfortunately isn't the case with resolves present installer.
although i really like your script, i'm still not satisfied by the actual situation. putting everything into a .deb-package and make it proper uninstallable is indeed a very important first step, but it's on the other hand just a part of the whole game. reproducible CI builds and tests, automatically generated dependency graphs, which declare all necessary other installation requirements, and comfortable updates from external repositories, without any need of manual workarounds, are very useful too.
and what's the main issue, all those really trivial fixes to the installer and naming of required libraries, which wouldn't lessen the compatibility in the case of CentOS and RHEL, but would make live much easier for debian/ubuntu/mint users, are still ignored by BMD. that's a really a very frustrating state of things, because you simply can not improve the situation in a significant manner, as long as the upstream source does their best to ignore/boycott any relevant attempts.
nevertheless i'm quite optimistic, that the situation will improve over time. it's very encouraging, how much interest among the linux users 15b already attracted. i think, BMD will have to react sooner or later and improve the software in regard to this little share of customers too.