On Fri, Oct 16, 2015 at 05:10:07PM +0000, Guenter Milde wrote:
> On 2015-10-14, Georg Baum wrote:
> > Scott Kostyshak wrote:
> 
> >> On Tue, Oct 13, 2015 at 06:31:04PM +0000, Guenter Milde wrote:
> >>> On 2015-10-12, Jean-Marc Lasgouttes wrote:
> 
> ...
> 
> >>> I believe that any commit message should go to lyx-cvs, just check for
> >>> some sign that this is an auto-generated message from the versioning
> >>> system. Authorisation is checked by the VS already.
> >>> Actgually, if there were unauthnorized activity on the repository, it
> >>> would be even more important to get this known on lyx-cvs.
> 
> > lyx-cvs is not a security device.
> 
> Of course not. However, there is no reason at all to have auto-generated
> commit messages from the repository returned from the lyx-cvs list.

The reason is simply that we have no one to fix this.

> I tried to play nice. It actually took months (if not years) until the known
> preconditions for comitting to the LyX repo were up and working (of course
> I did also something else inbetween).
> I spend hours setting up, activating, reactivating this lyx.org mail
> account and setting up a forwarding, because I was told this is required
> to get commit rights.
> 
> After my first commit, I was told: you must also subscribe to lyx-cvs (a
> list I don't want to get). Reluctantly, I tried also to hop through this
> next loop but failed. Also, my fixes were incomplete, because some tests did
> not run any more. There was no documentation about tests, no documentation
> about the need to subscribe to lyx-cvs.

This does indeed sound like a painful road. I think you are almost at
the end of the annoyances (I agree with Georg it should take 30
minutes), but after a lot of frustrations, even one more annoyance can
seem like a lot so I do kind of understand.

> Pausing and reconsidering

>From what I understand, by "reconsidering" you mean you changed your
mind that you wanted to subscribe to lyx-cvs. If I am wrong on this and
you still view subscription to lyx-cvs as the problem, I think I can
actually help with this. Please let me know.

> >> 1. You post your patches and then other LyX developers commit your
> >> patches.
> 
> >> 2. You commit your patches and you send an email with the exact same
> >> patch you just committed. This notifies everyone of your commit and also
> >> allows us to easily comment on the code if we choose to.
> 
> ...
> 
> > 1) is a waste of time IMHO and effectively revoking commit privileges 
> > (which 
> > is not a good idea either).
> 
> > 2) is a good temporary workaround, but not a long term solution IMHO.
> 
> I don't know the complete picture of the LyX code, so someone has to go
> through and audit my patches anyway, so 1) is the "safe bet".

Actually I think you have a great understanding of the LyX code that you
edit. In fact, I think Georg is the only one with overlapping
understanding of the code you touch. So if your lack of confidence is
the main reason, then please reconsider option 2.

> The patches are on the LyX bug-tracker ready for examination and
> application.

If I misunderstood the above (e.g. you just don't want to commit
yourself because you are frustrated) and we decide to go with option 1,
in my opinion the way to proceed is that you send an email to lyx-devel
and state which patches (e.g. attachment or direct link to the patch on
trac) you view as complete and ready for the master branch. Then, I or
anyone else would just apply the patches blindly. I trust your knowledge
so if you think a patch should go in then I will commit it. Often,
actually, the second-stage "audit" process happens because of lyx-cvs
*after* the commit; that is why lyx-cvs is so important to us.

In any case the important thing is that we find something that works for
everyone. We need your fixes!

Scott

Reply via email to