a quick status update on uv3 experiment

[Marshall Schor <ms...@schor.com>]

Hi,

The version in the branch for this is beginning to come to life; many test cases
are now running.

The major items yet to be done include binary (plain and compressed) and vinci
serialization, and working out a good strategy and set of tools for PEARs and
JCasGen.  I'll probably work on binary serialization next.

-Marshall

Re: a quick status update on uv3 experiment

[Marshall Schor <ms...@schor.com>]

Another caution - because of ongoing major development in this branch, the
snapshot you get probably has a large number of non-compiling files (e.g. the
binary serialization ones, not done yet), and may be in a "broken" state from
time to time as things get tried and checked in...

-Marshall

On 12/2/2015 9:23 AM, Marshall Schor wrote:
> possibly, but be aware that little "tuning" has been done so far - most of the
> work is just getting things done and past test cases.
>
> -Marshall
>
> On 12/1/2015 5:23 PM, Eddie Epstein wrote:
>> Sounds promising. Is enough running to compare performance on any
>> applications?
>>
>> Eddie
>>
>> On Tue, Dec 1, 2015 at 4:31 PM, Marshall Schor <ms...@schor.com> wrote:
>>
>>> Hi,
>>>
>>> The version in the branch for this is beginning to come to life; many test
>>> cases
>>> are now running.
>>>
>>> The major items yet to be done include binary (plain and compressed) and
>>> vinci
>>> serialization, and working out a good strategy and set of tools for PEARs
>>> and
>>> JCasGen.  I'll probably work on binary serialization next.
>>>
>>> -Marshall
>>>
>

Re: a quick status update on uv3 experiment

[Marshall Schor <ms...@schor.com>]

possibly, but be aware that little "tuning" has been done so far - most of the
work is just getting things done and past test cases.

-Marshall

On 12/1/2015 5:23 PM, Eddie Epstein wrote:
> Sounds promising. Is enough running to compare performance on any
> applications?
>
> Eddie
>
> On Tue, Dec 1, 2015 at 4:31 PM, Marshall Schor <ms...@schor.com> wrote:
>
>> Hi,
>>
>> The version in the branch for this is beginning to come to life; many test
>> cases
>> are now running.
>>
>> The major items yet to be done include binary (plain and compressed) and
>> vinci
>> serialization, and working out a good strategy and set of tools for PEARs
>> and
>> JCasGen.  I'll probably work on binary serialization next.
>>
>> -Marshall
>>

Re: a quick status update on uv3 experiment

[Marshall Schor <ms...@schor.com>]

Other things to be aware of:

v3 requires a re-generate of the JCas cover classes.  The regenerate may not
work properly with previously created JCas cover classes that have customization
(hasn't been tested).

v3 also changes (slightly) a bunch of the public APIs.  Some of these changes
may require changes in the using code.  Compatibility "packages" to mitigate
this haven't been done (yet).

-Marshall

On 12/1/2015 5:23 PM, Eddie Epstein wrote:
> Sounds promising. Is enough running to compare performance on any
> applications?
>
> Eddie
>
> On Tue, Dec 1, 2015 at 4:31 PM, Marshall Schor <ms...@schor.com> wrote:
>
>> Hi,
>>
>> The version in the branch for this is beginning to come to life; many test
>> cases
>> are now running.
>>
>> The major items yet to be done include binary (plain and compressed) and
>> vinci
>> serialization, and working out a good strategy and set of tools for PEARs
>> and
>> JCasGen.  I'll probably work on binary serialization next.
>>
>> -Marshall
>>

Re: a quick status update on uv3 experiment

[Eddie Epstein <ea...@gmail.com>]

Sounds promising. Is enough running to compare performance on any
applications?

Eddie

On Tue, Dec 1, 2015 at 4:31 PM, Marshall Schor <ms...@schor.com> wrote:

> Hi,
>
> The version in the branch for this is beginning to come to life; many test
> cases
> are now running.
>
> The major items yet to be done include binary (plain and compressed) and
> vinci
> serialization, and working out a good strategy and set of tools for PEARs
> and
> JCasGen.  I'll probably work on binary serialization next.
>
> -Marshall
>