Blacklist gulp-karma

[pkozlowski-opensource]

https://github.com/lazd/gulp-karma isn't a real plugin as it doesn't process any files.
It has some fundamental design issues (lazd/gulp-karma#33)
and doesn't bring enough value as Karma can be invoked directly
(see https://github.com/karma-runner/gulp-karma)

[phated]

@lazd is pretty good with gulp guidelines. We should ping him first

[pkozlowski-opensource]

Definitively! Although I had a number of exchanges with @lazd on the topic. I believe that with the current Karma design gulp-karma as a plugin is not providing any value and creates problem described in lazd/gulp-karma#21. This might change in the future when Karma exposes more hook points but this is not the case today.

I hesitated a long time before opening this PR but today I had to debug another project where https://github.com/lazd/gulp-karma was used and "watch wasn't working" and this was one time too much. Sorry for putting it straight like this but we should not encourage people to use plugins that have well-known problems and a simpler alternative exists.

[yocontra]

@pkozlowski-opensource Thanks for taking the time to help out btw

[lazd]

There isn't much we can do to gulp-karma to make it a valid gulp plugin while still offering value.

Since gulp-karma was written, a number of issues against Karma itself were closed that made what gulp-karma was trying to do less useful. I did put together a different version of gulp-karma that was gulp-friendly (but not a gulp-plugin) with a nice promise-based API, so that might be what needs to replace the gulp-karma plugin. See lazd/gulp-karma#21 (comment) for details on that (even that needs to be updated as Karma has fixed a few things it had to work around).

[sindresorhus]

So what is the conclusion on this? Blacklist or not?

[pkozlowski-opensource]

I guess @lazd argument is that there are shorcomings in Karma that - when solved - would make gulp-karma useful. I tend to agree with the fact that there are several shortcomings in the Karma but the issues raised by @lazd are not trivial to solve and I don't see them being addressed in the coming weeks / months.

So, looking at where we stand today I still claim that gulp-karma does more harm than good so I would black-list it given today's situation. Then I would love to work with @lazd / whoever else to improve gulp + karma story. But this won't happen overnight so we shouldn't be confusing users today.

[yocontra]

@pkozlowski-opensource Yep that makes sense to me. Remove from the blacklist when it is ready.

[lazd]

Agreed. @pkozlowski-opensource, let's see what we can do about making Karma play nice with gulp -- if there is an adaptor of sorts, it can live on as gulp-karma, but what we have there right now isn't doing much good.