On Mon, Feb 15, 2016 at 08:44:03AM +0100, Michal Kubecek wrote:
On pátek 12. února 2016 10:10 Tomáš Čech wrote:
People are frustrated about the bureaucracy of the process because they see package submission as their goal. But from the distribution POV it is just beginning and all that rules and quality requirements are set to ease package maintenance - which is something that is important for every other user.
No. Definitely not in general.
Yes, sure, it was meant as a one of the reasons I see. Sorry for making it sound too general.
Sure, there are people who just want to get a package in Factory and that's it for them. But please don't just presume everyone is like that. When I submit a package or take maintainership of one, I take it as a commitment to take care of the package in the long term (which is also one of the reasons I don't do so too often). And I'm certainly not the only one.
OK.
What does the frustration come from is the fact that while I took this commitment and take it seriously, I can't have the specfile formated in a way that would make my work easier. I can't have tags ordered in a way I find logical, I can't have two lines between sections etc. because there is some supersmart script whose author thought otherwise. Is that script's author going to submit updates, is he going to fix bugs or build breakages? Most likely not, in most cases it would be me. But for some reason the specfile must be formated the way _he_ prefers (at the given moment), not the way that would make it easier to read and maintain for me. And this is exactly why I started this thread (not for some "bashing" as some chose to believe).
I thought that we all agreed already that the problem was elsewhere - that changes got accepted without your consent. Why do you blame the script again? Blame people accepting requests.
The paragraph quoted above sounds like you simply presume submitters and maintainers are irresponsible and careless people who just want to "throw the package over the fence" and are not going to take care of it anymore. (Which would also explain all this general policy of imposing rules on people's packages even when there are no objective reasons so I'm afraid it's not just you and it's not just a momentary lapse.)
But did it occur to you that by this policy of assuming the worst, you may get the exact opposite of what you want to achieve, i.e. drive away the other kind? That while a "throw over the fence" submitter would have no problem with spec cleaner completely reshuffling his specfile, someone who intends to maintain that specfile for years may much more likely be frustrated by having it reformatted against his logic and his preferences? If not, you probably should give it some thought.
In the end, it's not going to matter if you convince yourself that all these style rules like "no two empty lines between sections" are necessary to keep "high quality standards". In the end, it's always the submitter/maintainer decission who decides if his motivation to get the package in Factory is strong enough to cope with it or if he would rather keep maintaining the package in his home project.
http://38.media.tumblr.com/e8fcca6de5326bf438c4cb0fbaf4abf7/tumblr_inline_o1... First, it may sounded to general and yet it was meant as one of possible reasons. That change shouldn't have been accepted without your consent. If the situation is like you are describing, Base:System/ncurses/ncurses.spec wouldn't look like it does :) I can't say much about obs-service-format_spec_file. It is there, it worked in many cases well. spec-cleaner did a lot of good already when people used that manually. Now it has some 'minimal' mode which is meant as replacement for original obs-service-format_spec_file and that is about it. 1] noone forces you to use spec-cleaner 2] if you find problem with spec-cleaner in its minimal mode, it would be nice to speak up now and report it as a bug Best regards, S_W