scalariform zombieware?

[Sam Halliday]

Hi all,

Have any of you noticed that scalariform has become zombieware?

  https://github.com/mdr/scalariform

I have tried to contact the author but not received a response. I hope nothing bad has happened to him.

This chap has been pumping out a few updates and fixing bugs

  https://github.com/daniel-trinh/scalariform

(especially correctly implementing "newline for hanging brace", which almost everybody agrees is the prettier way to do things)

but he has to push to his own organisation on maven central. Therefore many people won't have seen it and certainly won't get automatic updates (and that introduces its own versioning problems).

Given how important source code formatting is to the Scala IDE (and we use it in ENSIME), is this a project that should perhaps be pulled under the scala-ide organisation on github, with either a new maven org or try to get access to org.scalariform via sonatype?

Best regards,
Sam

[Matthew Farwell]

I'd certainly be interested in something 'official' for 2.11, because scalariform gets used in Scalastyle.

Matthew Farwell.

--
You received this message because you are subscribed to the Google Groups "Scala IDE Dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to scala-ide-de...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/scala-ide-dev/94e3fb89-f82c-4ec3-a207-d452d7b5c98b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[iulian dragos]

Hi Sam,

We are in touch with Daniel, and we'll probably switch to his branch for the next release of the IDE. He's been very responsive (see our merged PRs from skyluc).

cheers,

iulian

To view this discussion on the web visit https://groups.google.com/d/msgid/scala-ide-dev/CAOBkoFVPMLLhYRyXzbCCxvCEDN82HT1HHx0WFUNmpTi1aQTBYA%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--
« Je déteste la montagne, ça cache le paysage »
Alphonse Allais

[Rafał Krzewski]

I'm wondering if there is any progress with switching to Daniel's branch, and / or reclaiming org.scalariform group id?

I've posted a PR to that repo a few days ago (https://github.com/daniel-trinh/scalariform/pull/35) but there was no activitiy on it so far...

Cheers,
Rafał

[Matthew Farwell]

Yep, I've switched for Scalastyle 0.6.0, which will be being released this weekend.

Have fun,

Matthew Farwell.

To view this discussion on the web visit https://groups.google.com/d/msgid/scala-ide-dev/708dd9bc-2e11-4901-a8cd-e7f17efc9c4a%40googlegroups.com.

[Sam Halliday]

Does anybody know any more about this?

In ENSIME we're in a difficult position because Daniel's branch is not completely backwards compatible with the canon version, and is published under a different groupid.

It would really help us out if either daniel's branch becomes canon, or all the classes in his branch are moved to a different FQN domain (with a corresponding scalariform sbt plugin)

The problem we're having:

* we have an sbt plugin that will produce an ensime config file from an sbt definition

* we want our sbt plugin to depend on our new S-Expression parser, which (one aspect of it) will convert scalariform settings into sexpressions, pulling in daniel's branch

* if we do this, as soon as users - who are using any release of scalariform - pull in our plugin, they could get two scalariforms on their classpath and all hell will break loose.

Any workaround to this involves users having to put magical exclusion filters into their plugins.sbt file.

[Matthew Farwell]

Can you not do something funky with classpaths? In your sbt plugin, you can dynamically add the scalariform classes to the classpath, and even use a different classloader. This would probably protect you and the users from any wierdness.

Have fun,

Matthew Farwell.

To view this discussion on the web visit https://groups.google.com/d/msgid/scala-ide-dev/898cae2e-ff95-46fc-a2e8-978b38bd3308%40googlegroups.com.

[Sam Halliday]

On Sunday, 11 January 2015 19:08:22 UTC, Matthew Farwell wrote:

Can you not do something funky with classpaths? In your sbt plugin, you can dynamically add the scalariform classes to the classpath, and even use a different classloader. This would probably protect you and the users from any wierdness.

SBT is difficult enough to work with.

It would simpler to fork daniel's repo and build 2.9, 2.10 and 2.11 artefacts under the 'org.ensime' groupId, while also moving all the package names to the `org.ensime.scalariform` FQN.

I'd rather not have to go there. I think it can be solved if we can just get in contact with Matt Russell to either push a new release or give scala-ide / danieltrinh / both permission to push to org.scalariform. (He was last active a month ago on Github. Does anybody know him / have his contact details?)

[iulian dragos]

On Mon, Jan 12, 2015 at 9:50 PM, Sam Halliday <sam.ha...@gmail.com> wrote:

On Sunday, 11 January 2015 19:08:22 UTC, Matthew Farwell wrote:

Can you not do something funky with classpaths? In your sbt plugin, you can dynamically add the scalariform classes to the classpath, and even use a different classloader. This would probably protect you and the users from any wierdness.

SBT is difficult enough to work with.

It would simpler to fork daniel's repo and build 2.9, 2.10 and 2.11 artefacts under the 'org.ensime' groupId, while also moving all the package names to the `org.ensime.scalariform` FQN.

I'd rather not have to go there. I think it can be solved if we can just get in contact with Matt Russell to either push a new release or give scala-ide / danieltrinh / both permission to push to org.scalariform. (He was last active a month ago on Github. Does anybody know him / have his contact details?)

I know him, but he didn't reply to any of my emails or direct tweets when asking about scalariform. So I'd say that the project is dead and convince Daniel to publish under a different org ID.

 

To view this discussion on the web visit https://groups.google.com/d/msgid/scala-ide-dev/8eb1fce1-969d-489e-b897-8b87951bbc1f%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Matthew Farwell]

2015-01-12 21:50 GMT+01:00 Sam Halliday <sam.ha...@gmail.com>:

SBT is difficult enough to work with.

 

No it isn't. All you need is a good book :-)

http://manning.com/suereth2/

Matthew Farwell.

[Sam Halliday]

I've been deep in the details of SBT and have implemented several non trivial plugins. I can assure you that no book could possibly make it enjoyable to work with, unless it is a book on making drugs.

--
You received this message because you are subscribed to a topic in the Google Groups "Scala IDE Dev" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/scala-ide-dev/65JHMO6Kkao/unsubscribe.
To unsubscribe from this group and all its topics, send an email to scala-ide-de...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/scala-ide-dev/CAOBkoFXwiS14U39tdnrU3RTMEFFeiHe47NhGzp-MQ_kFrCrtZg%40mail.gmail.com.