[Freeciv-Dev] Re: Please vote!

[Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>]

On Mon, Dec 03, 2001 at 01:44:44AM -0500, Ross W. Wetmore wrote:
> 
> Unfortunately, you are heading to a language police state with the
> effort to reach the ultimate precision of definition and subtleties
> of nuance far more time consuming and important than the value of the
> code change - plus demonization of those that slip up or don't
> subscribe to the purity of the "one-true-way" as preached by the
> Ayatollahs of Infinite Coding Wisdom :-).
> 
> None of these code differences were sufficiently important to worry 
> about, none in the class of the server code everyone is trying so
> hard to preserve, none worth spending time on that could be used to
> develop worthwhile improvements.
> 
> But on patch submit, if the submitter isn't doing something really 
> gross, and doesn't want to implement some suggestion, then check it in
> as submitted. 

> Let someone submit a followup patch to fix the cosmetic thing if
> there is sufficient angst.

I said in another mail that this isn't acceptable. We already have a
part where such a policy was used: the AI. If your proposal will be
implemented we will have many other parts of the code with the same
problems. 

> And if the language patch bogs down in endless debate, at least it
> won't impact real code changes.

> Note, most people will follow the spirit of general coding guidelines
> but really dislike petty rules, fixups or someone else breaking their 
> code. Over the long haul, the codebase will move to the norm if you just
> loosen up and let people do their own thing with the review suggestions
> giving a chance to redirect, and the occasional rejected patch for real
> abuse teaching people not to be sloppy or stray too far.

I general: the style guide is not primary for "normal" patches but for
"style cleanup" patches. I want to avoid a situation where author A
submits a patch which changes "return(a)" to "return a" (the patch
will be applied) and 2 months later another author B submitts a patch
which reverts this change. All because of "it looks better this
way". There has to be a rule for this. Either there is freedom (like
now) where no form is prefered (and no patch which changes the form
will be accepted) or there is exactly one form (this doesn't have to
be the best one (there are always objectations to this one choosen
form)).

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
  reality.sys corrupt. Reboot Universe? (y,n,q)