[OpenLayers-Dev] Request for help from Opera users
tschaub at opengeo.org
Wed Jun 10 12:15:59 EDT 2009
Christopher Schmidt wrote:
> On Wed, Jun 10, 2009 at 02:34:53PM +1000, Roald de Wit wrote:
>> Hi Chris,
>> Thank you for your reply.
>> Christopher Schmidt wrote:
>>> Any functionality that is in a released version of OpenLayers should
>>> continue to work the same in all future versions of OpenLayers of the
>>> same major revision. That is, there should never be any reason that
>>> upgrading from 2.7 to 2.8 should cause anything to break.
>> I agree in general but not when somebody (1 person) uses an old
>> version of Opera that should have been upgraded anyway (security).
> IE6 stopped receiving security updates 6 months ago. If something breaks
> in IE6, do we say "Sorry, we don't support that browser -- upgrade."?
I understand and respect your decision to not release with something
that breaks what worked previously.
However, I'd say a closer parallel would be if we introduced something
that broke apps in Firefox 3.0.3 but not in 3.0.10. Or if we introduced
something that caused exceptions in Yahoo Maps v3.6 but not in v3.8.
I wouldn't be surprised if we have created issues like those above.
However, if nobody complains, we can release with a clear conscience.
When someone does point out a regression, we work to fix it. I also
think it makes sense to get an idea of the impact of the regression.
>>> The report from the user is not clear. If -- as I am assuming --
>>> "This worked in 2.7, and 2.8 breaks it" -- that is, simply switching to
>>> 2.8 causes a regression in behavior -- then this is a bug.
>> Unless we introduce new functionality that hits a bug in Opera that they
>> fixed in a later stable version.
> This is fine if the issue is 'new functionality'. "The build doesn't
> work at all" is not new functionality. (If this was something that could
> be fixed in a build profile, I'd say remove it from the full build, and
> call everything good, but that's simply not going to cut it here, since
> it's in BaseTypes.js.)
>> It is a fact that the 2.8 branch does not work with this particular
>> version of this browser (I verified that today).
>> Plus, the 'new RegExp('...')' notation makes regular expressions harder
>> to read with the extra backslashes.
>> Anyway, the patch in #2103 does not solve the issue yet. It looks like
>> another regular expression needs to be converted.
> Thanks for looking into it. I'll try to make time to do some digging
> myself tomorrorw.
OpenGeo - http://opengeo.org
Expert service straight from the developers.
More information about the Dev