|
proposing a compromise:
Create a new RFE for general implementation of a new correlation-aware flow analysis.
(It should be clear we speak of an entirely new feature, not a little tweak of an existing feature).
Mark all existing correlation bugs as duplicates of the new one.
Stephan
PS: some time ago I was JDT's joker/sweeper (whatever the English term may be), who could implement new features at will...
Is it a law of OSS that every joker will be absorbed into day-to-day business sooner or later?
PPS: while comparing null-analyses, has anybody recently checked, whether IDEA is able to leverage JSR 308 annotations after all?
IMHO this makes a much more significant difference than correlation analysis. Someone in @EclipseJavaIDE may want to advertise this ...
Am 2018-05-30 18:13, schrieb Daniel Megert:
> Reopening makes it harder to focus on the things we are able to achieve without external help.
That's exactly why we have the 'helpwanted' keyword. You can use/ignore that to keep the focus :-).
Dani
From: stephan.herrmann@xxxxxxxxx
To: "Eclipse JDT general developers list." <jdt-dev@xxxxxxxxxxx>
Date: 30.05.2018 18:09
Subject: Re: [jdt-dev] Null dataflow analysis & community
Sent by: jdt-dev-bounces@xxxxxxxxxxx
Reopening makes it harder to focus on the things we are able to achieve without external help.
just saying.
Stephan
Am 2018-05-30 18:05, schrieb Daniel Megert:
Hi Mickael
I agree that WONTFIX is not ideal. I suggest you reopen the corresponding bugs and add a 'helpwanted'.
Dani
From: Mickael Istria <mistria@xxxxxxxxxx>
To: "Eclipse JDT general developers list." <jdt-dev@xxxxxxxxxxx>
Date: 30.05.2018 09:28
Subject: [jdt-dev] Null dataflow analysis & community
Sent by: jdt-dev-bounces@xxxxxxxxxxx
Hi all,
With this interesting tweet for a JetBrains developer https://twitter.com/tagir_valeev/status/1001662720259903488, I had a look at the state of dataflow analysis in JDT, and saw https://bugs.eclipse.org/bugs/show_bug.cgi?id=494897, which was closed as WONTFIX. And there are a family of them: https://bugs.eclipse.org/bugs/buglist.cgi?classification=Eclipse&component=Core&list_id=17578767&product=JDT&query_format=advanced&short_desc=%5Bnull%5D%5Bcorrelation%5D&short_desc_type=allwordssubstr
While I agree with the technical and organizational reasons for this not being implement in the current system, I've got 2 comments to make about it:
* First, I don't think closing a bug as WONTFIX just because someone thinks that at a given point of time, this is not likely to be implemented by current contributors or because of manpower is a good practice. More, I strongly think it's a very bad one that gives the impression of a project that's not open (ie considering current manpower as immutable which is similar to considering the current team as closed and not open to growth). It's important to realize that the lifecycle of the project, the amount of contributors, their priorities can evolve. Closing as WONTFIX sends the opposite message. Keeping such bugs open for new contributors sends a better message of openness.
* Second, it gives the impression of a project that is not likely to innovate which cascades to the impression of a dying project.
A better practice IMO is just to put a comment saying "sure, it's interesting but [technical reasons, challenges] and current contributors currently have other tasks as high priority. But if you think you can contribute something in that direction, it would be welcome".
Now, getting into the technical details, I think that the constraint of immediate null analysis report is indeed making things hard but is worth it. But it may not prevent from allowing JDT to have different strategies for null analysis: none, as-you-type (less reliable), extra-builder (more reliable).
So if someone comes with dataflow analysis for Eclipse JDT one day, it should be possible to plug it into JDT properly.
JetBrains code for dataflow analysis is OSS https://github.com/JetBrains/intellij-community/tree/master/java/java-analysis-impl/src/com/intellij/codeInspection/dataFlow. Maybe it could even be reused.
Cheers,
--
Mickael Istria
Eclipse IDE developer, for Red Hat Developers_______________________________________________
jdt-dev mailing list
jdt-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jdt-dev
_______________________________________________
jdt-dev mailing list
jdt-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jdt-dev
_______________________________________________
jdt-dev mailing list
jdt-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jdt-dev
|
