On Sunday 20 February 2005 7:36 pm, Ales Hvezda wrote: > Something I have been toying (both in terms of thinking about it > and trying it out) with is doing a binary distribution of gEDA/gaf for > Linux (for older and current boxes). Assuming I could pull this off > without statically linking the dependencies (which btw, does not seem to > work too well, IME), would this make people happy? Anybody interested in > such a thing? Would people even run it? I'm not sure of the technical > details yet, but my initial experiments look quite promising. > > I really dislike to keep holding off on current library versions, just > because we have users who are on older boxes. But, I don't want leave > those users out in the cold either. I'm looking for a solution which > allows us to move forward and get past these build/dependancy issues.
Hello Ales, I understand where you are coming from. At the same time, on a more atomic level, this clearly a simple dependency issue as well as a statement detailing the percieved and actual abilities of users to deal with dependency issues. Dependencies exist on all platforms. A reasonable example is a missing DLL on a MS Wintendo box. On the other hand, many users understand a missing DLL, but on the other hand, these same users do not understand a clearly stated message on a UNI* platform that says Foobar.so is not in the path or can't be located. The next thing you know, the user is posting a message to a list complaining they cannot compile and/or install a package or library Looking at the great work that Stuart Brorson is doing regarding the gEDA installation CD typifies this situation. It seems to me that the work Stuart is doing should/would take care of this issue (That is, if he can get the various distro's to play nicely with his python scripts and visa-versa). FWIW, I've used Stuarts installation CD and although it's not quite prime time, he is close - Real close. What is Stuart's position on this subject? Why should a project like gEDA and the various programs and utilities that compose it suffer under the weight of old, outdated libs when it is so unnecessary? Other OSS projects don't and they are very successful. Best regards Marvin Dickens
pgp5Zh2uHqUlr.pgp
Description: PGP signature
