Eclipse Community Forums: Java Development Tools (JDT)

Help

Home

Home » Language IDEs » Java Development Tools (JDT) » Old bugs

Show: Today's Messages :: Show Polls :: Message Navigator

Old bugs [message #1731255]

Tue, 03 May 2016 20:25

Colin Sharples

Messages: 96
Registered: July 2009
Location: Wellington, New Zealand

Member

I count 4405 open defects against JDT that were created more than 3 years ago [1]. Around half of those are still in NEW state and don't look like they've even been investigated.

Is there a resourcing issue with JDT? Has there been a plan put together to try and address this backlog? It seems strange that there is a lot of emphasis on developing new features when there are so many unaddressed defects.

JDT in many ways is the flagship of Eclipse - it's the first component that most users will encounter, so it needs to be working as smoothly and efficiently as possible. I have been an Eclipse user since 2001, and there are defects in JDT that have been there as long as I have been using it. I'm so accustomed to dealing with the bugs that I automatically type the fix without thinking about it. It just seems amazing that after all this time, Eclipse can't properly format code when you wrap it with an if statement. I mean, seriously?

The standard response seems to be "if it upsets you that much, learn the code and submit a patch", but I am an Eclipse user, not an Eclipse developer, so why should I have to do that?

Please, please, please can we have some effort put into cleaning up JDT?

[1] https://bugs.eclipse.org/bugs/buglist.cgi?bug_status=UNCONFIRMED&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&classification=Eclipse&email1=ctg%40ctg.co.nz&emailtype1=exact&f1=creation_ts&limit=0&list_id=14324568&o1=lessthan&order=priority%2Cbug_severity&product=JDT&query_format=advanced&v1=2013-05-01

Colin Sharples
CTG Games Ltd
Wellington, New Zealand

Report message to a moderator

Re: Old bugs [message #1731258 is a reply to message #1731255]

Tue, 03 May 2016 20:54

Stephan Herrmann

Messages: 1853
Registered: July 2009

Senior Member

If today you look at https://projects.eclipse.org/projects/eclipse.jdt/who you'll see that 1/3 of commits to JDT during last 3 months are from people in the categories [Unaffeliated] and [Contributor]. We, too, are Eclipse users, who only became Eclipse developers because we chose to help.

I can't say much about the majority of old unattended bugs, most of them pre-date my joining the team.

I can, however, assure that every new bug is being looked at, huge efforts are invested in trying to reproduce, asking back for more info, and finally fixing - if we have enough information. So, the least you can do is: find the existing bug for any problem you are observing, check its state, and provide information necessary for reproducing the issue, or if you observe a new problem, file a new bug. A handful of excellent bug reports may have similar value as a bug fix.

There is no free lunch but the kitchen is open to everybody. If you can't or don't want to cook you can still slice the vegetable Smile

Last but not least: if you choose not to get involved with your time, you can still donate money, which since recently will also be used for improving Eclipse.

best,
Stephan

Report message to a moderator

Re: Old bugs [message #1731260 is a reply to message #1731258]

Tue, 03 May 2016 21:13

Stephan Herrmann

Messages: 1853
Registered: July 2009

Senior Member

I can also say a word regarding cleanup vs. new features.

I'd say the biggest share of work in JDT is: bug fixing, so we are already doing a lot of what you are begging for.

This may be directly followed by developing support for new Java versions. You wouldn't want JDT to stop supporting new Java versions, would you?

In JDT/Core perhaps one of the biggest "new features" during the last years is annotation-based null analysis. Guess what: this is being developed in my spare time. The interesting twist here is: its largely due to the opportunity to contribute new functionality that I enjoy working on JDT. This is to say: there is not as much competition between feature work and bug fixing as you may think. To some extent it's actually synergy.

Another "biggest new feature" is a complete new Java formatter, after the old formatter was getting unmaintainable. So here the "new feature" is actually a pre-requisite for fixing the exact thing you are whining about. Have you tried the new formatter? If you still see a problem, now is the time to help Mateusz to identify and fix it, he's doing a heroic job at it, believe me.

Stephan

Report message to a moderator

Re: Old bugs [message #1731382 is a reply to message #1731260]

Wed, 04 May 2016 19:13

Colin Sharples

Messages: 96
Registered: July 2009
Location: Wellington, New Zealand

Member

I have been a friend of Eclipse for many years, I have submitted many bug reports, I am actively supporting the development of GEF4, which is where my main expertise and interest lies. This is not whining, nor am I expecting others to do my work for me - the spare time I have to offer to Eclipse is better used where it is useful, which is not JDT.

I am just pointing out that there's a lot of open bugs, and no visibility of what might be happening with them. Yes, new features would be great, but not if the existing features suffer as a result. How about, instead of just brushing off your users as "whiners", you actually engage with them?

Colin Sharples
CTG Games Ltd
Wellington, New Zealand

Report message to a moderator

Re: Old bugs [message #1731393 is a reply to message #1731382]

Wed, 04 May 2016 22:11

Stephan Herrmann

Messages: 1853
Registered: July 2009

Senior Member

Quote:

Yes, new features would be great, but not if the existing features suffer as a result.

I tried to explain that there's more synergy than competition between new feature work and fixing old bugs.

Quote:

How about, instead of just brushing off your users as "whiners", you actually engage with them?

What kind of engagement do you have in mind?

Report message to a moderator

Re: Old bugs [message #1731405 is a reply to message #1731393]

Thu, 05 May 2016 00:44

Colin Sharples

Messages: 96
Registered: July 2009
Location: Wellington, New Zealand

Member

A great start would be a plan for how to start reducing that backlog. It may very well be that new features have made some, possibly a lot, of those old bugs redundant, so I imagine a first step would probably be along the lines of getting people to test and see if these old bugs are still a problem. Winnowing out the non-existent or superseded bugs would reduce the task greatly. From there, it would make sense to find out the ones that are causing most pain, so they can be targeted as a priority. To be clear, what I am suggesting here is something beyond what is already available in Bugzilla (votes/cc etc), but an additional set of resources specifically set up to support a campaign to reduce the backlog. The more people are encouraged to get involved and help fix the bugs, of course the easier it will be, but I think the first step is to acknowledge that there is an issue with the backlog, and that it will take some effort to sort it out.

Colin Sharples
CTG Games Ltd
Wellington, New Zealand

Report message to a moderator

Re: Old bugs [message #1731512 is a reply to message #1731405]

Fri, 06 May 2016 11:04

Stephan Herrmann

Messages: 1853
Registered: July 2009

Senior Member

Eclipse carries out some compaigns like "great fix" contests. Maybe a "clear the backlog" campaign could be blended into that strategy.

A few things to consider:

Are we discussing how prioritization of bug fixing should be improved or are we looking for ways to get more people involved in fixing bugs? You personally seem to be out of reach regarding the latter, so let's focus on the former: given fixed resources for fixing bugs, how can we make sure we're fixing the most relevant bugs. Additionally, some clean-up in bugzilla (like closing obsolete bugs) can improve the picture we have about the status.

Regarding available resources (developer time): you asked "Is there a resourcing issue with JDT? ". I didn't answer to that because I believe the answer for any software project will be "yes, of course". For JDT this implies that any time that JDT committers invest in organizing a campaign, any time we spend in mining old bugs means fewer bug fixes in a given interval of time. BTW, when talking about number of bugs, just keep in mind that JDT is a "big" project with a total of over 54k bugs.

Regarding relevance: We currently focus very much on bugs with "recent activity", like: recently reported, recently commented with new examples etc. I assume that activity like this is already a pretty good indicator about relevance of bugs to users. Any campaign we'd start, should be careful not to unearth seemingly interesting old bugs, which effectively don't bother anybody in real life today.

I am answering from the perspective of a JDT committer. Maybe at this point I should pass the question on to other users reading this forum:

Are you worried about JDT's backlog?
If so, are you willing to help sift through the backlog and provide interesting new information regarding old bugs?

Here, "Interesting new information" could be a demonstration that the bug has been fixed (since which version?), or a reproducing example for a bug which so far couldn't be reproduced etc.

If we get lots of positive responses we might get something new moving. Let's see.
Stephan

Report message to a moderator