[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gnash-dev] Dictatorship again ? (was: revno.h)

From: strk
Subject: [Gnash-dev] Dictatorship again ? (was: revno.h)
Date: Wed, 8 Dec 2010 23:50:37 +0100

For some obscure reason, rob reverted this change
I've made (that took me a long time) w/out explaining
what was wrong with it, except

 "it was better before"

It's over my head. Once again.

Last time this happened we almost forked.
Came out of that with help of others that care about
the project and arrived at a set of rules for everyone 
to follow:

Rule number 3 is:
 - commits shall not be reverted except as a last resort.

the Notes section says:
 A last resort means that a commit that breaks the other rules is not
 fixed or reverted by the committer in a reasonable time.
 Code reversion not done by the original committer should generally
 only be done by the maintainer after being proposed on the mailing list.

Did anyone read anything in thi mailing list about what was wrong
with my commit except from pere's observation ?

Of course I'm pretty much pissed off ... 


On Wed, Dec 08, 2010 at 01:57:14PM +0100, strk wrote:
> I've cleaned up the way revno.h is generated
> and it's contents.
> (1) revno.h is updated/generated in source tree, not build tree
> (2) revno.h is never cleaned (should be in distributed tarballs)
> (3) revno.h is rebuilt at _every_ make invocation (fast, with git)
> (4) revno.h is not touched if its content don't change
> (5) BRANCH_REVNO is now abbreviated commit-id
> That's pretty much it.
> I've updated the debian/rules script to go look in the
> correct place.
> Only concern is about rob mentioning he wanted a sequential
> number for the debian package versions. The new BRANCH_REVNO
> doesn't give you that. Once I tought we could count commits
> in the repo, but that's also depnedent on the kind of repo you
> have (if full or rooted at some commit). 
> In any case we surely DO WANT the commit id too, as that gives
> you exact information about the source tree from which a binary
> is derived, so if we really need a sequencial number too, we should
> add something to revno.h rather than condending BRANCH_REVNO.
> --strk;
>   ()   Free GIS & Flash consultant/developer
>   /\
> _______________________________________________
> Gnash-dev mailing list
> address@hidden


  ()   Free GIS & Flash consultant/developer

reply via email to

[Prev in Thread] Current Thread [Next in Thread]