Use Advanced Search to search the entire archive.
[jsr363-experts] Re: 0.8-RC3
- From: Werner Keil <
>
- To: "
" <
>
- Subject: [jsr363-experts] Re: 0.8-RC3
- Date: Sat, 11 Jul 2015 14:43:58 +0200
Dear Experts,
Happy to announce, JSR 363 RI also is 0.8 RC3 ready after pending issues
were resolved.
The Q2/15 Sprint closed, see the report:
https://java.net/jira/secure/RapidBoard.jspa?rapidView=12&view=reporting&chart=sprintRetrospective&sprint=47
Thanks everyone who helped by either reporting or fixing stuff or
contribute to the decision process.
The JavaOne talk Leo and I proposed (modeled after his and Chris' succesful
DevoXX UK talk last month) was not accepted by the main conference track.
As a result there's a little less time pressure for the Public Draft and it
feels best to offer a Hackergarten session with those who are at JavaOne
(Leo, myself, and if they got time at least Otavio and Mohamed also should
be there) and give the community another first-hand chance to do final
polishing and tweaking to the JSR before Public Draft should be filed in
early November (the JCP.next deadline is 1 year after EDR just at Christmas
'14, we'd be close but within that threshold then)
Either at the October JCP F2F or (especially if it was to be hosted in
Austin where Leo is based now) the Q1/16 one (Jan/Feb) we also could offer
another review by the JCP EC. If it seems beneficial to do this right
before Public Review, then maybe in Fall, otherwise after that.
This brings a Final Draft or Release (unless there was a need for a second
PD) close to Java SE 9 schedule (currently at the end of Sep 2016, guess
they want to have it out at JavaOne '16;-) Whether or not JavaOne likes
our proposal then, we'll see, but more importantly we can offer it for both
Java SE (8/9) and ME (8 until 9 was released) pretty much when the new
versions are out, too.
Public Draft shall be used to improve the "Service Character", e.g. trying
to get the ServiceProvider/Bootstrap in a similar shape and place as other
JSRs (CDI, BeanValidation, JCache or JavaMoney are among good examples)
which will benefit the aim of "pluggable" modular Unit Systems, which can
then simply be called via ServiceLoader, too.
These are tasks for the Public Draft sprint as of now:
https://java.net/jira/secure/RapidBoard.jspa?rapidView=12&view=planning
Regards,
Werner
On Fri, Jul 3, 2015 at 1:47 PM, Werner Keil
<
>
wrote:
>
Dear Experts,
>
>
The API already went out as 0.8-RC3 yesterday and is now available on
>
MavenCentral, too.
>
>
For 2 formatting issues, at least one under advanced investigation (see
>
https://java.net/jira/secure/RapidBoard.jspa?rapidView=12&view=planning)
>
the RI is not currently out. I know what the problem is there, the duality
>
of KILOGRAM and GRAM in SI and other unit systems causes an ambiguity for
>
some Mass units, where it can happen that MICRO(GRAM) internally becomes
>
NANO(KILOGRAM) (so the factor is always correct, so is the numeric value)
>
but a UnitFormat hack aiming to replace KILOGRAM with GRAM does not
>
understand that, so MIGRO(GRAM) ended up "ng" rather than "µg".
>
>
RI Unit tests catch this, and some aspect of the problem is already
>
addressed.
>
The 5 or so JUnit tests failing show this on the CI server:
>
>
expected:<[n]g> but was:<[pk]g>
>
>
>
It is actually mathematically correct but ugly and usually not what people
>
would expect.
>
>
A side-effect could be, that "chaining" operations, e.g. "PICO(KILO(GRAM)"
>
would result in a formatting based on GRAM for everything but KILOGRAM. So
>
if you chained them like that in the end you'd always get "ng", not "pkg".
>
If we really needed variation for some UnitFormat implementations, then we
>
could address that. For now especially the RI (which also does not
>
translate or localize units by default) should be fine with that.
>
>
If at least the GRAM vs. KG bug can be fixed, I expect RI to be released
>
no later than a week from now. I'll check the separate HERTZ issue (also
>
mathematically correct but the "human-readable" format expects different;-)
>
and everyone who can is welcome to have a look, too. Otherwise we have to
>
shift it into the next sprint. Either Public Draft or (should there be more
>
changes worth publishing in the meantime) another RC.
>
>
Regards,
>
>
Werner
>