harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Geir Magnusson Jr." <g...@pobox.com>
Subject Re: [testing] test exclude list: can't we have incremental exclusions?
Date Sat, 25 Nov 2006 21:38:44 GMT


Alexey Varlamov wrote:
> 24.11.06, Geir Magnusson Jr.<geir@pobox.com> написал(а):
>>
>>
>> Alexey Varlamov wrote:
>> > Geir,
>> > This was a bit emotional maybe... Sure, any way it will be is not
>> > lethal, and I do not mind it too much.
>>
>> What, the veto?  I didn't take it that way :)  My point was that when
>> someone puts up a "-1", it really gets people's attention as a strong
>> position.
>>
>> > My point is if you modify "official" x-list you most certainly won't
>> > lose it off track, while local svn-ignored file have a good chance to
>> > hang around for a while. OTOH, is there any difference which file to
>> > edit? I suppose no, hence this is almost useless in my POV.
>>
>> I agree you won't lose it, but when it's in a file that isn't meant for
>> purely personal use then you have problems in being sure not to commit
>> it, having to deal with merge conflicts, etc.
>>
>> I think about it in the same spirit of the drlvm.properties.example -
>> people copy to an un-svn-ed local copy for local config.  Excluding
>> tests while you are working on something is that kind of thing.

> Nope - drlvm.properties fixes stable state of things, you set it and
> forget it; most merging conflicts are just bothering fuss.

No - there is no merging conflict for drlvm.properties, as it's not in 
SVN.  And you can do wacky things with drlvm.properties - like point to 
a modified classlib.

> Excluding tests should be momentary and ideally not happening at all,
> so some minor inconveniences may be paying here. And even merging
> conflict should justly draw your attention - maybe your modification
> is not that local.

I think that we can't legislate this - if people need them, they'll use 
it or find a way around it, like explicitly excluding tests, and if they 
don't, they won't.

IOW, it's a convenience for those that need it.

> 
>>
>> > If you really want it, I've withdrawn my veto.
>>
>> No :)  I'd like to come to consensus.  You may even convince me it's not
>> a good idea.
>>
>> I think that maybe one solution that may address your concerns would be
>> to actually put the file under SVN!  Then
>>
>> a) you'll notice when there's something in it - the state of the file in
>> SVN should always be empty
>>
>> b) If you forget and commit, someone can flag it.
> 
> This would also solve that problem with non-existing file.
> But, now this adds even less value - the only difference from
> mainstream x-lists is somewhat lesser chance for conflicts :).

And therefore no harm right?  So I'm not going to waste time trying to 
convince you of the value if you can't see it.  But have we come to 
agreement that there's no real danger?

geir


Mime
View raw message