[Zope3-dev] Re: [Zope-Coders] Someone broke the HEAD, again :-(

[Chris Withers]

Guido van Rossum wrote:
> 
> Maybe we misunderstood each other.  I'm not saying that Jeremy isn't
> responsible for fixing it.  I'm also not saying that you should fix
> it.  I'm just saying that *if* somebody else (like me) figures out how
> to fix it, there's no rule that says I shouldn't check in the fix
> (assuming I know what I'm doing -- never commit when in doubt).

*grinz* yup, misunderstanding then... If I'd known hwo to fix it withotu buggering
anything else up, I woulda given it a shot and seen if I could make all the tests pass...

> How could I break something and not know about it?  There are unit
> tests.  (Well, some people apparently don't run them, and that's a
> separate issue that the authorities will take very seriously...)

Indeed. Either there wasn't a test ( which can be valid, if you cause or uncover a bug,
then there won't be a test... of course the first thing anyone should in that case is
write a test ;-) or the tests weren't run, in which case I favour kneecapping or tarring
and feathering ;-)

> When the unit tests don't reveal a problem, it may be a compiler or
> platform issue, and then I may be dependent on someone with access to
> the platform where it breaks for the fix.

...yeah, I'm in the unfortunately position of being able to find these problems but not
fix them :-S

> > In this case, there are unit tests, which should have been run. I'm
> > not sure how they could have run if the code wouldn't compile. If
> > the code did compile on the platform of whoever ran the tests, then
> > fine, but now they're aware, I'd expect they're in a much better
> > position to fix the code than me...
> 
> Not true.  Compilers differ vastly between platforms.  It's often hard
> to know what the compiler on a platform you don't use will complain
> about. 

We're misunderstanding each other again, I'm in agreement with this ;-)

> together.  In this particular case, it's not clear to me that Jeremy
> should have known that a redundant semicolon causes an error on
> Windows.  (I do think he should've noticed that he was inserting
> declarations after statements. :-)

I think my C must be rustier than I thought... why didn't that semicolon break the code on
all platforms? surely it's not correct C?

> > PS: I didn't make any assertions about code ownership, just
> > resonsibilities for actions people take ;-)
> 
> I think I've explained why it isn't always as easy as you seem to think.

well, the responsobilities are pretty simple, but yeah, acting on those responsibilities
is 'fun' ;-)

cheers,

Chris