Good question. These issues were closed on my request. I've done similar clean-ups in the past. The issue tracker has gotten to large to be effective in help guide our work. We need to clean it up. In addition, lots of these issues weren't linked to the original reporter, making it less likely that the original reporter would step up with more information if needed, etc. On Tue, Feb 24, 2015 at 11:12 PM, Thomas Tuegel wrote:

...

hi,

i noticed today's run at the closing "inactive" issues on the tracker. i would like to to ask an innocent question: what exactly is the benefit of

...

(i disclose that it seems to me that valid, if inactive, issues are being closed, which i do not like. but before complaining, i want to know

On Tue, Feb 24, 2015 at 3:52 PM, lennart spitzner wrote: this action? the

...

counter-arguments).

The most important reason is that we do not now, nor will we ever, have the human resources to fix all those issues. When we have done this before, we usually get a few people who chime in about an issue that still affects them. This allows us to prioritize on issues that cause actual developers actual problems. It also lets us find these issues; there are still valid issues back there, on Page 28 of our GitHub issue tracker, but I know I never venture back there.

If an issue was closed that still causes you problems, you should by all means request that it be reopened.

-- Thomas Tuegel _______________________________________________ cabal-devel mailing list cabal-devel@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/cabal-devel

On 25-02-2015 19:21, lennart spitzner wrote:

I am not convinced. how does closing ~40 out of ~700 open tickets make the contributors more effective? that demand exceeds resources is true, but it is no argument for closing issues. many of the issues represent sensible ideas for features that do not need new feedback.

Well, it's a *start* at reducing the ridiculous number of outdated issues. Nobody is served by having huge numbers of outdated issues in an issue tracker. It's demotivating and the likelihood of an issue being fixed (or implemented, or...) decreases exponentially the longer it's been in a tracker... which is usually fair enough since it must mean that it's not *that* important after all.

I'd say the general lack of stability and the recently mentioned lack of tests are the main problems of Cabal; to a degree this looks like shooting at symptoms.

That may certainly be the case. You should feel to contribute fixes for any of the existing issues -- that would help the Cabal maintainer(s) enormously, I suspect. Regards,

On Wed, Feb 25, 2015 at 12:36 PM, Bardur Arantsson wrote:

On 25-02-2015 19:21, lennart spitzner wrote:

...

I am not convinced. how does closing ~40 out of ~700 open tickets make the contributors more effective? that demand exceeds resources is true, but it is no argument for closing issues. many of the issues represent sensible ideas for features that do not need new feedback.

Well, it's a *start* at reducing the ridiculous number of outdated issues. Nobody is served by having huge numbers of outdated issues in an issue tracker. It's demotivating and the likelihood of an issue being fixed (or implemented, or...) decreases exponentially the longer it's been in a tracker... which is usually fair enough since it must mean that it's not *that* important after all.

...

I'd say the general lack of stability and the recently mentioned lack of tests are the main problems of Cabal; to a degree this looks like shooting at symptoms.

That may certainly be the case. You should feel to contribute fixes for any of the existing issues -- that would help the Cabal maintainer(s) enormously, I suspect.

On a related note, I'm tagging issues with 'documentation' or 'easy' as I find them. Either should be do-able for a first-time contributor. In particular, the 'easy' issues are ones I think I could talk a Haskell programmer through in a 1-2 paragraphs; something I think a first-time contributor could knock out in an afternoon. Periodically, folks ask about small projects for advanced students. I know Cabal's not hip and exciting, or whatever, but please think of us. -- Thomas Tuegel

On Wed, Feb 25, 2015 at 1:53 PM, Thomas Tuegel wrote:

On Wed, Feb 25, 2015 at 12:21 PM, lennart spitzner wrote:

...

I am not convinced. how does closing ~40 out of ~700 open tickets make the contributors more effective? that demand exceeds resources is true, but it is no argument for closing issues. many of the issues represent sensible ideas for features that do not need new feedback.

At the risk of beating a dead horse, I just wanted to point out an exchange on Twitter [1] which _proves_ that having a large number of open tickets discourages our users from opening new issues when they encounter bugs, even severe performance regressions. I recommend, if you think there is any reason to believe an issue is inactive, close it! We can always re-open issues if the re-occur. [1]. https://twitter.com/shebang/status/590578911148380160 -- Thomas Tuegel

GHC has a process for systematically downgrading inactive tickets. See https://ghc.haskell.org/trac/ghc/wiki/WorkingConventions/BugTracker under "Re-milestoning after a release". Something like that might be useful for Cabal? Mind you, I'm not sure that we have executed this algorithm following the 7.10 release (Austin NB). Simon | -----Original Message----- | From: cabal-devel [mailto:cabal-devel-bounces@haskell.org] On Behalf Of | Kosyrev Serge | Sent: 21 April 2015 19:38 | To: Thomas Tuegel | Cc: cabal-devel | Subject: Re: inactive issues | | Thomas Tuegel writes: | > On Wed, Feb 25, 2015 at 1:53 PM, Thomas Tuegel | wrote: | >> On Wed, Feb 25, 2015 at 12:21 PM, lennart spitzner | >> wrote: | >>> I am not convinced. how does closing ~40 out of ~700 open tickets | make | >>> the contributors more effective? that demand exceeds resources is | >>> true, but it is no argument for closing issues. many of the issues | >>> represent sensible ideas for features that do not need new feedback. | > | > At the risk of beating a dead horse, I just wanted to point out an | > exchange on Twitter [1] which _proves_ that having a large number of | > open tickets discourages our users from opening new issues when they | > encounter bugs, even severe performance regressions. | > | > I recommend, if you think there is any reason to believe an issue is | > inactive, close it! We can always re-open issues if the re-occur. | | As a bystander and purely philosophically, the action of "closing" feels | gratuitiously non-injective, since it conflates "inactivity" with | "completion". | | Could it be that we could have a more discerning tracker, which would | show the "not inactive" subset of "open" issues by default? | | Hardly so, with github, and yet.. | | -- | regards, | Косырев Серёга | _______________________________________________ | cabal-devel mailing list | cabal-devel@haskell.org | http://mail.haskell.org/cgi-bin/mailman/listinfo/cabal-devel