Lance,
KERN-42 looks great, I have merged it. Looking at it got me tho thinking. A while back you asked me DOM or SAX and I said SAX. However we have used StaX in other places, not least because it removes most of the horrid behaviour present with the DOM and SAX parser factories, which appear to have been written at a time when there was only 1 classloader that mattered.
How would you feel about porting that part of the code to StaX ?
Ian
Which raises a more general question for the new year: What should our
standard way of handling cross-project dependencies be? We're in kind of
an unusual situation what with the gatekeeper of K2 also being a
committer to Sling, with K2 and Sling both under such active
development, and with K2's server-side code and 3akai's client-side code
being tightly linked functionally but controlled by different
individuals. (This is the first time I've needed to coordinate K2 and
Sling fixes, but it's fairly common to need to coordinate K2 and 3akai
changes.)
On the more specific question, just let me if you'd like me to merge
into master before the Sling fix is merged.
Best,
Ray
> My last comment http://jira.sakaiproject.org/browse/KERN-456 tries to
> describe the situation. The fix on my branch is tested, but it's
> dependent on a bug fix in Sling.
Ok, we are currently depending on contentloader-x.x.x-SNAPSHOT, so that should come through when the bug is fixed. I havent been heavilly involved in content loader in Sling, so I probably cant just apply the fix.
>
> Which raises a more general question for the new year: What should our
> standard way of handling cross-project dependencies be? We're in kind of
> an unusual situation what with the gatekeeper of K2 also being a
> committer to Sling, with K2 and Sling both under such active
> development, and with K2's server-side code and 3akai's client-side code
> being tightly linked functionally but controlled by different
> individuals. (This is the first time I've needed to coordinate K2 and
> Sling fixes, but it's fairly common to need to coordinate K2 and 3akai
> changes.)
I am thinking we should start using mvn version:lock-snapshot to fix the snapshots to a timed version.
>
> On the more specific question, just let me if you'd like me to merge
> into master before the Sling fix is merged.
Yes please, in general I merge from your master without being asked with a pull request.
If you have something in a branch then you will need to give me the Ok to merge.
Ian
>
> Best,
> Ray
>
> On 12/20/09 11:14 PM, Ian Boston wrote:
>> Ray,
>> Should I merge your KERN-456 branch or wait for it to me merged into your master.
>>
>> Lance,
>> KERN-42 looks great, I have merged it. Looking at it got me tho thinking. A while back you asked me DOM or SAX and I said SAX. However we have used StaX in other places, not least because it removes most of the horrid behaviour present with the DOM and SAX parser factories, which appear to have been written at a time when there was only 1 classloader that mattered.
>>
>> How would you feel about porting that part of the code to StaX ?
>>
>> Ian
>
> --
>
> You received this message because you are subscribed to the Google Groups "Sakai Kernel" group.
> To post to this group, send email to sakai-...@googlegroups.com.
> To unsubscribe from this group, send email to sakai-kernel...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/sakai-kernel?hl=en.
>
>
Yes, I understand -- the reason for the branch was that I did *not*
intend an immediate merge. If the K2 side of the fix goes in before the
Sling side of the fix, then K2 will fail during bundle activation. And I
didn't want to begin our mandated holidays by leaving a broken build
under your Xmas tree. :)
Best,
Ray
>>> On the more specific question, just let me if you'd like me to merge
>>> into master before the Sling fix is merged.
>>
>> Yes please, in general I merge from your master without being asked
> with a pull request.
>> If you have something in a branch then you will need to give me the
> Ok to merge.
>
> Yes, I understand -- the reason for the branch was that I did *not*
> intend an immediate merge. If the K2 side of the fix goes in before the
> Sling side of the fix, then K2 will fail during bundle activation. And I
> didn't want to begin our mandated holidays by leaving a broken build
> under your Xmas tree. :)
cool, thanks
Ian
Lance Speelmon
Scholarly Technologist