Hi
Ed,
Thanks
for this interesting feedback. What do the Java 8
annotations look like?
Annotation
of libraries sounds great, as this is usually the main issue
when trying to improve our own code (as it is for generics)
Regards,
Rémi
-------------------------------------------------------
Rémi
SCHNEKENBURGER
+33
(0)1 69 08 48 48
CEA
Saclay Nano-INNOV
Institut CARNOT CEA LIST
www.eclipse.org/papyrus
Hi
Your specific slide looks wierd. "@Nonnull". The functionality
incorporated by JDT is "@NonNull".
I've been using the Java 7 @NonNull annotations for a couple
of years for Eclipse OCL. The checking is 'native' at least
anything that JDT does, I regard as native.
The annotations are useful, but far from the hyped panacea.
The real killer is that the Java libraries / platform code /
EMF generated models have no annotations, so large parts of
your code remain unchecked. Deceptively so. You may get used
to good diagnosis and erroneously remove checks. You are
forced to use quite a few suppress null warnings, which if
accidentally applied at method or worse class level can hide
problems that would have been diagnosed without any
annotations. The flow analysis unfortunately fails in complex
control statements (more than a couple of if's/for's and you
may have to suppress the false diagnoses here too). Bugzillas
on these problems have been outstanding for quite some time.
I haven't yet successfully moved to the Java 8 annotations
which are incompatible and syntactically perverse. I have yet
to decide whether to stick with obsolete Java 7 annotations,
abandon them altogether or try to move on to Java 8
annotations that might yet evolve to something more
acceptable.
I suggest that you give them an extended try on a couple of
plugins that use a high level of autogenerated code since
without accurately annotating your autogeneration your efforts
will be fruitless.
There is now support for annotations for libraries. You might
want to look at autogenerating annotations for your external
dependencies such as EMF / GEF / ...
Regards
Ed Willink
On 31/08/2015 07:55, MAGGI Benoit wrote:
Hi
everyone,
NullPointer
seems to be a week point in Papyrus code.
I
saw this in a presentation (Eclipsecon Toulouse [1])
It
may be a solution to take down most of the NPE, but the
annotations are not native,
We
may have to add dependency to a checker (maybe also change
the jdk we use)
Any
thought on that ?
Regards,
Benoit
1
:
http://batmat.github.io/presentations/50-slides-of-ide/prez.html#__span_class_tip_count_26_51_span_null_analysis_one_step_further_with_type_annotations_jsr_308
2
:
http://types.cs.washington.edu/checker-framework/
Hi,
this
is the weekly digest for Papyrus. The status of all
problems on this list is open. Items on this list may
already be tracked in Bugzilla but were not closed yet. To
remove a item from this list, you can either (i)
set the status of its associated bug to resolved, verfied,
or closed or (ii) mark it as log message in the reporter's
web ui.
If
you have any questions about this report, please send an
email to
error-reports-inbox@xxxxxxxxxxx
Top
new problems (last 7 days):
Top
unresolved, active problems (last 7 days):
Top
reoccurring problems (last 3 months):
Links:
Thank
you for your assistance.
Your friendly error-reports-inbox.
--
You
are receiving this email because you are subscribed to
alerts of Papyrus. To alter your subscriptions, visit the
project
configuration page.
_______________________________________________
mdt-papyrus.dev mailing list
mdt-papyrus.dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/mdt-papyrus.dev
No
virus found in this message.
Checked by AVG - www.avg.com
Version: 2015.0.6125 / Virus Database: 4409/10547 - Release
Date: 08/30/15