Poll: what do you code Scala in?

Just now,

Eclipse 25% (38 votes)
IntelliJ IDEA 17% (26 votes)
NetBeans 16% (24 votes)
Emacs 13% (19 votes)
Vim 13% (20 votes)
TextMate 9% (14 votes)
JEdit 4% (6 votes)
Other: 3% (5 votes)
Total Votes: 152
2009/1/13 Alex Payne <al3x@al3x.net>

Over at Twitter we've been experimenting with different editors and
IDEs in search of the optimal Scala development experience. I'm
curious as to what the community at large is using. A poll seems like
a good way to do that.

If you'd be so kind, please take a moment to fill this out:
http://www.polldaddy.com/p/1272527/

The results will remain publicly available. Thanks!

--
Alex Payne
http://twitter.com/al3x

Actually, we frequently need ideas for new polls at scala-lang.org. If
anyone has further ideas for possible polls, we'd be happy to host them!
Toni

Rob Dickens wrote:
> Just now,
>
> Eclipse 25% (38 votes)
> IntelliJ IDEA 17% (26 votes)
> NetBeans 16% (24 votes)
> Emacs 13% (19 votes)
> Vim 13% (20 votes)
> TextMate 9% (14 votes)
> JEdit 4% (6 votes)
> Other: 3% (5 votes)
> Total Votes: 152
>
> 2009/1/13 Alex Payne
>
>> Over at Twitter we've been experimenting with different editors and
>> IDEs in search of the optimal Scala development experience. I'm
>> curious as to what the community at large is using. A poll seems like
>> a good way to do that.
>>
>> If you'd be so kind, please take a moment to fill this out:
>> http://www.polldaddy.com/p/1272527/
>>
>> The results will remain publicly available. Thanks!
>>
>> --
>> Alex Payne
>> http://twitter.com/al3x
>>
>

Here are a few ideas. Feel free to customize them.
* Poll 1 - What Java environment you use Scala in  -- Java SE (1.4)  -- Java SE (1.5)  -- Java EE (1.4)  -- Java EE (1.5)
* Poll 2 - What is the nature of your Scala projects  -- Pure Scala  -- Mixed Scala/Java, Scala mostly  -- Mixed Scala/Java, Java mostly
* Poll 3 - What is your favourite build system   -- Maven  -- Simple Build Tool  -- Ant  -- Buildr  -- A mix of the above  -- Custom
* Poll 4 - In your Scala code, do you use Java collections?   -- Yes, as they are  -- Yes, through the wrapper Scala API  -- Not often  -- No
* Poll 5 - How do you prefer to iterate a Scala collection?  -- using foreach   -- via Iterators
* Poll 6 - How do you feel about Monads in Scala  -- What are they?  -- I need more examples to understand  -- I have started to understand them, though not using them that much   -- I frequently use Scala's for as a Monadic construct  -- Can someone explain all that map/flatMap/... stuff??
Christos.
On Wed, Jan 14, 2009 at 10:53 AM, Antonio Cunei <scala_ml@cunei.com> wrote:

Actually, we frequently need ideas for new polls at scala-lang.org. If anyone has further ideas for possible polls, we'd be happy to host them!
Toni

Rob Dickens wrote:

Just now,

Eclipse 25% (38 votes)
IntelliJ IDEA 17% (26 votes)
NetBeans 16% (24 votes)
Emacs 13% (19 votes)
Vim 13% (20 votes)
TextMate 9% (14 votes)
JEdit 4% (6 votes)
Other: 3% (5 votes)
 Total Votes: 152

2009/1/13 Alex Payne <al3x@al3x.net>

Over at Twitter we've been experimenting with different editors and
IDEs in search of the optimal Scala development experience. I'm
curious as to what the community at large is using. A poll seems like
a good way to do that.

If you'd be so kind, please take a moment to fill this out:
http://www.polldaddy.com/p/1272527/

The results will remain publicly available. Thanks!

--
Alex Payne
http://twitter.com/al3x

--
 __~O
-\ <,       Christos KK Loverdos
(*)/ (*)      http://ckkloverdos.com

Not too many people even use 1.5 anymore, the current release is 1.6.

On Wed, Jan 14, 2009 at 5:21 PM, Christos KK Loverdos <loverdos@gmail.com> wrote:

Here are a few ideas. Feel free to customize them.
* Poll 1 - What Java environment you use Scala in   -- Java SE (1.4)  -- Java SE (1.5)  -- Java EE (1.4)  -- Java EE (1.5)

Interpret the suggested poll items "loosely": 1.4 means 1.4-and-previous, 1.5 means 1.5-and-next. I meant to differentiate two different JVM & language models. 1.5 and 1.6 are the same in this respect, while 1.4 is an old generation (and officially dead :-) ).
For the 1.5 vs 1.6 thing I can't tell.
Christos

On Wed, Jan 14, 2009 at 3:24 PM, Sean McDirmid <sean.mcdirmid@gmail.com> wrote:

Not too many people even use 1.5 anymore, the current release is 1.6.

On Wed, Jan 14, 2009 at 5:21 PM, Christos KK Loverdos <loverdos@gmail.com> wrote:

Here are a few ideas. Feel free to customize them.
* Poll 1 - What Java environment you use Scala in   -- Java SE (1.4)  -- Java SE (1.5)  -- Java EE (1.4)  -- Java EE (1.5)

--
 __~O
-\ <,       Christos KK Loverdos
(*)/ (*)      http://ckkloverdos.com

only those stick to 1.5 that have the "wrong mac computer" (like
me... thank you apple!)

Am 14.01.2009 um 14:24 schrieb Sean McDirmid:

> Not too many people even use 1.5 anymore, the current release is 1.6.
>
> On Wed, Jan 14, 2009 at 5:21 PM, Christos KK Loverdos
> wrote:
> Here are a few ideas. Feel free to customize them.
>
> * Poll 1 - What Java environment you use Scala in
> -- Java SE (1.4)
> -- Java SE (1.5)
> -- Java EE (1.4)
> -- Java EE (1.5)
>
>

Sure, but you are just inviting a lot of "neither" responses given that much of the world has moved on to 1.6 and might even be early testing 1.7. The question is better phrased as:
1.4 or earlier (EE/SE) 1.5 or later (EE/SE)

On Wed, Jan 14, 2009 at 9:38 PM, Christos KK Loverdos <loverdos@gmail.com> wrote:

Interpret the suggested poll items "loosely": 1.4 means 1.4-and-previous, 1.5 means 1.5-and-next. I meant to differentiate two different JVM & language models. 1.5 and 1.6 are the same in this respect, while 1.4 is an old generation (and officially dead :-) ).
For the 1.5 vs 1.6 thing I can't tell.
Christos

On Wed, Jan 14, 2009 at 3:24 PM, Sean McDirmid <sean.mcdirmid@gmail.com> wrote:

Not too many people even use 1.5 anymore, the current release is 1.6.

On Wed, Jan 14, 2009 at 5:21 PM, Christos KK Loverdos <loverdos@gmail.com> wrote:

Here are a few ideas. Feel free to customize them.
* Poll 1 - What Java environment you use Scala in   -- Java SE (1.4)  -- Java SE (1.5)  -- Java EE (1.4)  -- Java EE (1.5)

--
 __~O
-\ <,       Christos KK Loverdos
(*)/ (*)      http://ckkloverdos.com

Sean, ok, you are right and we are probably saying the same thing :)
They were poll-templates, as per my suggestion to Toni to customize them  :-)
BRChristos.

On Wed, Jan 14, 2009 at 4:12 PM, Sean McDirmid <sean.mcdirmid@gmail.com> wrote:

Sure, but you are just inviting a lot of "neither" responses given that much of the world has moved on to 1.6 and might even be early testing 1.7. The question is better phrased as:
1.4 or earlier (EE/SE) 1.5 or later (EE/SE)

On Wed, Jan 14, 2009 at 9:38 PM, Christos KK Loverdos <loverdos@gmail.com> wrote:

Interpret the suggested poll items "loosely": 1.4 means 1.4-and-previous, 1.5 means 1.5-and-next. I meant to differentiate two different JVM & language models. 1.5 and 1.6 are the same in this respect, while 1.4 is an old generation (and officially dead :-) ).
For the 1.5 vs 1.6 thing I can't tell.
Christos

On Wed, Jan 14, 2009 at 3:24 PM, Sean McDirmid <sean.mcdirmid@gmail.com> wrote:

Not too many people even use 1.5 anymore, the current release is 1.6.

On Wed, Jan 14, 2009 at 5:21 PM, Christos KK Loverdos <loverdos@gmail.com> wrote:

Here are a few ideas. Feel free to customize them.
* Poll 1 - What Java environment you use Scala in   -- Java SE (1.4)  -- Java SE (1.5)  -- Java EE (1.4)  -- Java EE (1.5)

--
 __~O
-\ <,       Christos KK Loverdos
(*)/ (*)      http://ckkloverdos.com

--
 __~O
-\ <,       Christos KK Loverdos
(*)/ (*)      http://ckkloverdos.com

I'm in the same boat. Just to clarify for the readers-at-home, Apple
only supplies JVM 1.5 on 32-bit Core and Core Duo Macs.
First-generation Intel Macs are stuck with 1.5. Core2 Macs get 1.6.

If you need 1.6 on a 32-bit Mac, look into the SoyLatte project. It's
a port of OpenJDK BSD.

Matt

On Wed, Jan 14, 2009 at 6:12 AM, Sciss wrote:
> only those stick to 1.5 that have the "wrong mac computer" (like me... thank
> you apple!)
>
> Am 14.01.2009 um 14:24 schrieb Sean McDirmid:
>
>> Not too many people even use 1.5 anymore, the current release is 1.6.

Alex,

I hope you twitter guys will post or blog about your experiences and
recommendations at some point. TIA.

dean

On Tue, Jan 13, 2009 at 2:39 PM, Alex Payne wrote:
> Over at Twitter we've been experimenting with different editors and
> IDEs in search of the optimal Scala development experience. I'm
> curious as to what the community at large is using. A poll seems like
> a good way to do that.
>
> If you'd be so kind, please take a moment to fill this out:
> http://www.polldaddy.com/p/1272527/
>
> The results will remain publicly available. Thanks!
>
> --
> Alex Payne
> http://twitter.com/al3x
>

i have soylatte installed, but honestly x11 sucks for desktop
applications IMO... and i'm a java on the desktop person ;-) apple
should open source their Cocoa-AWT bridge and donate it to the
soylatte project.

by the way scala runs fine on soylatte...

ciao, -sciss-

Am 14.01.2009 um 20:14 schrieb Matt Harrington:

> I'm in the same boat. Just to clarify for the readers-at-home, Apple
> only supplies JVM 1.5 on 32-bit Core and Core Duo Macs.
> First-generation Intel Macs are stuck with 1.5. Core2 Macs get 1.6.
>
> If you need 1.6 on a 32-bit Mac, look into the SoyLatte project. It's
> a port of OpenJDK BSD.
>
> Matt
>
>
> On Wed, Jan 14, 2009 at 6:12 AM, Sciss wrote:
>> only those stick to 1.5 that have the "wrong mac computer" (like
>> me... thank
>> you apple!)
>>
>> Am 14.01.2009 um 14:24 schrieb Sean McDirmid:
>>
>>> Not too many people even use 1.5 anymore, the current release is
>>> 1.6.

On Wed, Jan 14, 2009 at 8:38 AM, Christos KK Loverdos <loverdos@gmail.com> wrote:

Interpret the suggested poll items "loosely": 1.4 means 1.4-and-previous, 1.5 means 1.5-and-next. I meant to differentiate two different JVM & language models. 1.5 and 1.6 are the same in this respect, while 1.4 is an old generation (and officially dead :-) ).

Except for phoneME and other mobile stuff for some reason don't want to support generics etc.,

Definitely! I think we'll know a bit more about where we're it in
terms of tools in another week or so. Several of us have been
migrating from various plain-text editors like TextMate and Emacs to
IDEA.

On Wed, Jan 14, 2009 at 11:22, Dean Wampler wrote:
> Alex,
>
> I hope you twitter guys will post or blog about your experiences and
> recommendations at some point. TIA.
>
> dean
>
> On Tue, Jan 13, 2009 at 2:39 PM, Alex Payne wrote:
>> Over at Twitter we've been experimenting with different editors and
>> IDEs in search of the optimal Scala development experience. I'm
>> curious as to what the community at large is using. A poll seems like
>> a good way to do that.
>>
>> If you'd be so kind, please take a moment to fill this out:
>> http://www.polldaddy.com/p/1272527/
>>
>> The results will remain publicly available. Thanks!
>>
>> --
>> Alex Payne
>> http://twitter.com/al3x
>>
>
>
>
> --
> Dean Wampler
> http://www.objectmentor.com
> http://www.polyglotprogramming.com
> http://www.aspectprogramming.com
> http://aquarium.rubyforge.org
> http://www.contract4j.org
>

I have just been trying to achieve the following and did a quick
evaluation of IDEs:
- maven project with sub projects
- mix of java and scala code in one of the projects
- java calling scala and scala calling java

Eclipse:
- little glitchy to configure IAM (maven plugin for eclipse) to work
with scala and mix scala code and java code but is feasible.
- Calling the scala code from java wasn't too straightforward, there are
still some bugs in the eclipse plugin (I declared a Array[T] in scala
which was seen as Array in the jdt)
+ Scala seems to compile incrementally which doesn't seem to happen with
the other IDEs

The IAM developers seem to be commited to get scala to work with the
maven plugin and there is an effort of a scala plugin for IAM.

IDEA:
+ no issue for having scala and java on a project
- has been somehow unstable on linux
- requires the scala compiler to be installed separately - and
"building" my project caused a classnotfound after calling scalac

Netbeans:
+ uses the maven builder for building projects (no weird scalac
configuration)
- the java classes did not see the scala classes (this may require a
build which had syntax errors)

At this point, the scala support did not seem any stronger in IDEA than
Netbeans. I'm going to stick with the eclipse plugin for now and wait
for fixes.

Erik.
> -----Original Message-----
> From: Sciss [mailto:contact@sciss.de]
> Sent: January-14-09 2:22 PM
> To: Matt Harrington
> Cc: scala-tools@listes.epfl.ch
> Subject: Re: [scala-tools] Poll: what do you code Scala in?
>
> i have soylatte installed, but honestly x11 sucks for desktop
> applications IMO... and i'm a java on the desktop person ;-) apple
> should open source their Cocoa-AWT bridge and donate it to the
> soylatte project.
>
> by the way scala runs fine on soylatte...
>
> ciao, -sciss-
>
>
> Am 14.01.2009 um 20:14 schrieb Matt Harrington:
>
> > I'm in the same boat. Just to clarify for the readers-at-home,
Apple
> > only supplies JVM 1.5 on 32-bit Core and Core Duo Macs.
> > First-generation Intel Macs are stuck with 1.5. Core2 Macs get 1.6.
> >
> > If you need 1.6 on a 32-bit Mac, look into the SoyLatte project.
It's
> > a port of OpenJDK BSD.
> >
> > Matt
> >
> >
> > On Wed, Jan 14, 2009 at 6:12 AM, Sciss wrote:
> >> only those stick to 1.5 that have the "wrong mac computer" (like
> >> me... thank
> >> you apple!)
> >>
> >> Am 14.01.2009 um 14:24 schrieb Sean McDirmid:
> >>
> >>> Not too many people even use 1.5 anymore, the current release is
> >>> 1.6.

Out of curiosity, what did you find insufficient about plain-text editors like TextMate and Emacs?

I use TextMate plus Java/ScalaDocs (Firefox's AwesomeBar helps a lot for this) for most of my work, but I sometimes miss being able to ask the IDE what the inferred type for an expression is. However, I was constantly annoyed by the slowness of IDEs, especially for.. However, I was constantly annoyed by the slowness of IDEs, especially for such basic things as switching tabs. (To be honest, I've only used Eclipse and NetBeans, so maybe IDEA doesn't suffer from this particular fault.)

--j

On Wed, Jan 14, 2009 at 11:56 AM, Alex Payne <al3x@al3x.net> wrote:

Definitely! I think we'll know a bit more about where we're it in
terms of tools in another week or so. Several of us have been
migrating from various plain-text editors like TextMate and Emacs to
IDEA.

On Wed, Jan 14, 2009 at 11:22, Dean Wampler <deanwampler@gmail.com> wrote:
> Alex,
>
> I hope you twitter guys will post or blog about your experiences and
> recommendations at some point. TIA.
>
> dean
>
> On Tue, Jan 13, 2009 at 2:39 PM, Alex Payne <al3x@al3x.net> wrote:
>> Over at Twitter we've been experimenting with different editors and
>> IDEs in search of the optimal Scala development experience. I'm
>> curious as to what the community at large is using. A poll seems like
>> a good way to do that.
>>
>> If you'd be so kind, please take a moment to fill this out:
>> http://www.polldaddy.com/p/1272527/
>>
>> The results will remain publicly available. Thanks!
>>
>> --
>> Alex Payne
>> http://twitter.com/al3x
>>
>
>
>
> --
> Dean Wampler
> http://www.objectmentor.com
> http://www.polyglotprogramming.com
> http://www.aspectprogramming.com
> http://aquarium.rubyforge.org
> http://www.contract4j.org
>



--
Alex Payne
http://twitter.com/al3x

The main thing that's been frustrating is the lack of accurate syntax
highlighting, both in TextMate and Emacs, though it's far worse in
TextMate. I've already made one TextMate bundle (for the Puppet
syntax) and I'm hesitant to repeat the process if something else works
reliably and adds some intelligence to my development environment.

IDEA is definitely slow, even on an obscenely over-provisioned MacBook
Pro, and it annoys me too.

On Wed, Jan 14, 2009 at 12:20, Jorge Ortiz wrote:
> Out of curiosity, what did you find insufficient about plain-text editors
> like TextMate and Emacs?
>
> I use TextMate plus Java/ScalaDocs (Firefox's AwesomeBar helps a lot for
> this) for most of my work, but I sometimes miss being able to ask the IDE
> what the inferred type for an expression is. However, I was constantly
> annoyed by the slowness of IDEs, especially for such basic things as
> switching tabs. (To be honest, I've only used Eclipse and NetBeans, so maybe
> IDEA doesn't suffer from this particular fault.)
>
> --j
>
> On Wed, Jan 14, 2009 at 11:56 AM, Alex Payne wrote:
>>
>> Definitely! I think we'll know a bit more about where we're it in
>> terms of tools in another week or so. Several of us have been
>> migrating from various plain-text editors like TextMate and Emacs to
>> IDEA.
>>
>> On Wed, Jan 14, 2009 at 11:22, Dean Wampler wrote:
>> > Alex,
>> >
>> > I hope you twitter guys will post or blog about your experiences and
>> > recommendations at some point. TIA.
>> >
>> > dean
>> >
>> > On Tue, Jan 13, 2009 at 2:39 PM, Alex Payne wrote:
>> >> Over at Twitter we've been experimenting with different editors and
>> >> IDEs in search of the optimal Scala development experience. I'm
>> >> curious as to what the community at large is using. A poll seems like
>> >> a good way to do that.
>> >>
>> >> If you'd be so kind, please take a moment to fill this out:
>> >> http://www.polldaddy.com/p/1272527/
>> >>
>> >> The results will remain publicly available. Thanks!
>> >>
>> >> --
>> >> Alex Payne
>> >> http://twitter.com/al3x
>> >>
>> >
>> >
>> >
>> > --
>> > Dean Wampler
>> > http://www.objectmentor.com
>> > http://www.polyglotprogramming.com
>> > http://www.aspectprogramming.com
>> > http://aquarium.rubyforge.org
>> > http://www.contract4j.org
>> >
>>
>>
>>
>> --
>> Alex Payne
>> http://twitter.com/al3x
>
>

I've had very good experiences with jEdit's syntax highlighting. Of
course, I had to modify the mode, but the point is that it works!

The only thing I miss from an IDE is semantic highlighting.
Considering Scala's varied method dispatch syntax, this can be
extrordinarily useful at times. Popup code completion is nice too,
but I can live without it, especially when it doesn't complete for
extension methods anyway.

Daniel

On Jan 14, 2009, at 4:11 PM, "Alex Payne" wrote:

> The main thing that's been frustrating is the lack of accurate syntax
> highlighting, both in TextMate and Emacs, though it's far worse in
> TextMate. I've already made one TextMate bundle (for the Puppet
> syntax) and I'm hesitant to repeat the process if something else works
> reliably and adds some intelligence to my development environment.
>
> IDEA is definitely slow, even on an obscenely over-provisioned MacBook
> Pro, and it annoys me too.
>
> On Wed, Jan 14, 2009 at 12:20, Jorge Ortiz
> wrote:
>> Out of curiosity, what did you find insufficient about plain-text
>> editors
>> like TextMate and Emacs?
>>
>> I use TextMate plus Java/ScalaDocs (Firefox's AwesomeBar helps a
>> lot for
>> this) for most of my work, but I sometimes miss being able to ask
>> the IDE
>> what the inferred type for an expression is. However, I was
>> constantly
>> annoyed by the slowness of IDEs, especially for such basic things as
>> switching tabs. (To be honest, I've only used Eclipse and NetBeans,
>> so maybe
>> IDEA doesn't suffer from this particular fault.)
>>
>> --j
>>
>> On Wed, Jan 14, 2009 at 11:56 AM, Alex Payne wrote:
>>>
>>> Definitely! I think we'll know a bit more about where we're it in
>>> terms of tools in another week or so. Several of us have been
>>> migrating from various plain-text editors like TextMate and Emacs to
>>> IDEA.
>>>
>>> On Wed, Jan 14, 2009 at 11:22, Dean Wampler
>>> wrote:
>>>> Alex,
>>>>
>>>> I hope you twitter guys will post or blog about your experiences
>>>> and
>>>> recommendations at some point. TIA.
>>>>
>>>> dean
>>>>
>>>> On Tue, Jan 13, 2009 at 2:39 PM, Alex Payne wrote:
>>>>> Over at Twitter we've been experimenting with different editors
>>>>> and
>>>>> IDEs in search of the optimal Scala development experience. I'm
>>>>> curious as to what the community at large is using. A poll seems
>>>>> like
>>>>> a good way to do that.
>>>>>
>>>>> If you'd be so kind, please take a moment to fill this out:
>>>>> http://www.polldaddy.com/p/1272527/
>>>>>
>>>>> The results will remain publicly available. Thanks!
>>>>>
>>>>> --
>>>>> Alex Payne
>>>>> http://twitter.com/al3x
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Dean Wampler
>>>> http://www.objectmentor.com
>>>> http://www.polyglotprogramming.com
>>>> http://www.aspectprogramming.com
>>>> http://aquarium.rubyforge.org
>>>> http://www.contract4j.org
>>>>
>>>
>>>
>>>
>>> --
>>> Alex Payne
>>> http://twitter.com/al3x
>>
>>
>
>
>

On Wed, Jan 14, 2009 at 10:23 PM, Daniel Spiewak wrote:
> Popup code completion is nice too, but I can live without it,
> especially when it doesn't complete for extension methods anyway.

You mean for methods provided via implicits? That works in the Eclipse
plugin ...

Cheers,

Miles

It does now?  Sweet!  I haven't been able to use it in some time (it doesn't seem to work on Eclipse 3.5) so that's probably how I missed it.  Nice job!
Daniel

On Wed, Jan 14, 2009 at 4:34 PM, Miles Sabin <miles@milessabin.com> wrote:

On Wed, Jan 14, 2009 at 10:23 PM, Daniel Spiewak <djspiewak@gmail.com> wrote:
> Popup code completion is nice too, but I can live without it,
> especially when it doesn't complete for extension methods anyway.

You mean for methods provided via implicits? That works in the Eclipse
plugin ...

Cheers,


Miles

--
Miles Sabin
tel:    +44 (0)1273 720 779
mobile: +44 (0)7813 944 528
skype:  milessabin

[Sorry if this comes through more than once, gmail is going freaky on me]

On Wed, Jan 14, 2009 at 02:11:28PM -0800, Alex Payne wrote:
> The main thing that's been frustrating is the lack of accurate syntax
> highlighting, both in TextMate and Emacs, though it's far worse in
> TextMate. I've already made one TextMate bundle (for the Puppet
> syntax) and I'm hesitant to repeat the process if something else works
> reliably and adds some intelligence to my development environment.

I have tried out all the big IDEs repeatedly, and always come running back to textmate with my tail between my
legs. I also am usually on my overprovisioned macbook pro, and it's hard for me to see how anyone gets any
work done in eclipse, intellij, or netbeans, although netbeans is the least slow. All I can say is if you do
make any improvements to textmate's syntax highlighting or other scala-specific bits, at least one person will
use those improvements with great gusto.

BTW, with respect to adding some intelligence to your dev environment, I went with enhancing the interpreter
since at this point I have no confidence anything fancier than textmate will ever be appealing. I'd hesitate
to call it "intelligence" but I added some completion bits which save me a lot of time. I had intended to
submit it for trunk but I had to put it aside and haven't come back to it. In the below log imagine yourself
hitting "tab" on each scala> line; the fields and method signatures are all discovered via reflection. Also,
you can fire up the scaladoc/javadoc for a class in a web browser by hitting ctrl-T.

scala> import scala.col

scala> import scala.collection.

scala.collection.BitSet scala.collection.Map scala.collection.MapProxy
scala.collection.Ranged scala.collection.RollbackIterator scala.collection.Set
scala.collection.SetProxy scala.collection.Sorted scala.collection.SortedMap
scala.collection.SortedSet scala.collection.immutable. scala.collection.jcl.
scala.collection.mutable.

scala> import scala.collection.Set.

* ** apply contains equals hashCode isEmpty size
stringPrefix subsetOf toArray

scala> scala.collection.Set.contains

abstract def contains(A): Boolean

scala> java.lang.String

def this(Array[byte])
def this(StringBuffer) def this(StringBuilder)
def this(Int, Int, Array[char]) def this(Array[byte], Int, Int)
def this(Array[int], Int, Int) def this(Array[byte], Int, Int, Int)
def this(Array[byte], Int) def this(Array[byte], Int, Int, String)
def this(Array[byte], String) def this(Array[char], Int, Int)
def this(String) def this()
def this(Array[char])

A project to improve the interpreter would be much appreciated.  For the command-line version, tab completion would be really nice, and I think about as far as you could go.  I would also like to see a richer Swing/SWT version with full syntax highlighting, popup completion, etc.  I'm not sure yet how often I would use it (the CL version is just so darn convenient), but it might be a big win for all the pansies suffering from shell phobia.  :-)
Daniel

On Wed, Jan 14, 2009 at 4:57 PM, Paul Phillips <paulp@improving.org> wrote:

[Sorry if this comes through more than once, gmail is going freaky on me]

On Wed, Jan 14, 2009 at 02:11:28PM -0800, Alex Payne wrote:
> The main thing that's been frustrating is the lack of accurate syntax
> highlighting, both in TextMate and Emacs, though it's far worse in
> TextMate. I've already made one TextMate bundle (for the Puppet
> syntax) and I'm hesitant to repeat the process if something else works
> reliably and adds some intelligence to my development environment.

I have tried out all the big IDEs repeatedly, and always come running back to textmate with my tail between my
legs.  I also am usually on my overprovisioned macbook pro, and it's hard for me to see how anyone gets any
work done in eclipse, intellij, or netbeans, although netbeans is the least slow.  All I can say is if you do
make any improvements to textmate's syntax highlighting or other scala-specific bits, at least one person will
use those improvements with great gusto.

BTW, with respect to adding some intelligence to your dev environment, I went with enhancing the interpreter
since at this point I have no confidence anything fancier than textmate will ever be appealing.  I'd hesitate
to call it "intelligence" but I added some completion bits which save me a lot of time.  I had intended to
submit it for trunk but I had to put it aside and haven't come back to it.  In the below log imagine yourself
hitting "tab" on each scala> line; the fields and method signatures are all discovered via reflection.  Also,
you can fire up the scaladoc/javadoc for a class in a web browser by hitting ctrl-T.


scala> import scala.col

scala> import scala.collection.

scala.collection.BitSet             scala.collection.Map                scala.collection.MapProxy
scala.collection.Ranged             scala.collection.RollbackIterator   scala.collection.Set
scala.collection.SetProxy           scala.collection.Sorted             scala.collection.SortedMap
scala.collection.SortedSet          scala.collection.immutable.         scala.collection.jcl.
scala.collection.mutable.

scala> import scala.collection.Set.

*              **             apply          contains       equals         hashCode       isEmpty        size
stringPrefix   subsetOf       toArray

scala> scala.collection.Set.contains

abstract def contains(A): Boolean

scala> java.lang.String

def this(Array[byte])
def this(StringBuffer)                    def this(StringBuilder)
def this(Int, Int, Array[char])           def this(Array[byte], Int, Int)
def this(Array[int], Int, Int)            def this(Array[byte], Int, Int, Int)
def this(Array[byte], Int)                def this(Array[byte], Int, Int, String)
def this(Array[byte], String)             def this(Array[char], Int, Int)
def this(String)                          def this()
def this(Array[char])

--
Paul Phillips      | Beware of bugs in the above code; I have only
Apatheist          | proved it correct, not tried it.
Empiricist         |     -- Knuth
up hill, pi pals!  |----------* http://www.improving.org/paulp/ *----------

JLine does support tab completion.  I don't exactly know how the callback works, but I assume it's something that could be hooked into in order to provide the desired functionality.  I think the problem would be needing to incrementally parse partially-entered expressions.  Considering how stolid Scala's parser can be (especially with paren-matching), I wonder how possible this is.
Does anyone know the Scala shell code / parser well enough to give a definitive answer (possible/impossible) on this?

Daniel
On Wed, Jan 14, 2009 at 5:31 PM, Jorge Ortiz <jorge.ortiz@gmail.com> wrote:

I would love to have this enhancement available.

On Wed, Jan 14, 2009 at 2:57 PM, Paul Phillips <paulp@improving.org> wrote:

[Sorry if this comes through more than once, gmail is going freaky on me]

On Wed, Jan 14, 2009 at 02:11:28PM -0800, Alex Payne wrote:
> The main thing that's been frustrating is the lack of accurate syntax
> highlighting, both in TextMate and Emacs, though it's far worse in
> TextMate. I've already made one TextMate bundle (for the Puppet
> syntax) and I'm hesitant to repeat the process if something else works
> reliably and adds some intelligence to my development environment.

I have tried out all the big IDEs repeatedly, and always come running back to textmate with my tail between my
legs.  I also am usually on my overprovisioned macbook pro, and it's hard for me to see how anyone gets any
work done in eclipse, intellij, or netbeans, although netbeans is the least slow.  All I can say is if you do
make any improvements to textmate's syntax highlighting or other scala-specific bits, at least one person will
use those improvements with great gusto.

BTW, with respect to adding some intelligence to your dev environment, I went with enhancing the interpreter
since at this point I have no confidence anything fancier than textmate will ever be appealing.  I'd hesitate
to call it "intelligence" but I added some completion bits which save me a lot of time.  I had intended to
submit it for trunk but I had to put it aside and haven't come back to it.  In the below log imagine yourself
hitting "tab" on each scala> line; the fields and method signatures are all discovered via reflection.  Also,
you can fire up the scaladoc/javadoc for a class in a web browser by hitting ctrl-T.


scala> import scala.col

scala> import scala.collection.

scala.collection.BitSet             scala.collection.Map                scala.collection.MapProxy
scala.collection.Ranged             scala.collection.RollbackIterator   scala.collection.Set
scala.collection.SetProxy           scala.collection.Sorted             scala.collection.SortedMap
scala.collection.SortedSet          scala.collection.immutable.         scala.collection.jcl.
scala.collection.mutable.

scala> import scala.collection.Set.

*              **             apply          contains       equals         hashCode       isEmpty        size
stringPrefix   subsetOf       toArray

scala> scala.collection.Set.contains

abstract def contains(A): Boolean

scala> java.lang.String

def this(Array[byte])
def this(StringBuffer)                    def this(StringBuilder)
def this(Int, Int, Array[char])           def this(Array[byte], Int, Int)
def this(Array[int], Int, Int)            def this(Array[byte], Int, Int, Int)
def this(Array[byte], Int)                def this(Array[byte], Int, Int, String)
def this(Array[byte], String)             def this(Array[char], Int, Int)
def this(String)                          def this()
def this(Array[char])

--
Paul Phillips      | Beware of bugs in the above code; I have only
Apatheist          | proved it correct, not tried it.
Empiricist         |     -- Knuth
up hill, pi pals!  |----------* http://www.improving.org/paulp/ *----------

I would love to have this enhancement available.

On Wed, Jan 14, 2009 at 2:57 PM, Paul Phillips <paulp@improving.org> wrote:

[Sorry if this comes through more than once, gmail is going freaky on me]

On Wed, Jan 14, 2009 at 02:11:28PM -0800, Alex Payne wrote:
> The main thing that's been frustrating is the lack of accurate syntax
> highlighting, both in TextMate and Emacs, though it's far worse in
> TextMate. I've already made one TextMate bundle (for the Puppet
> syntax) and I'm hesitant to repeat the process if something else works
> reliably and adds some intelligence to my development environment.

I have tried out all the big IDEs repeatedly, and always come running back to textmate with my tail between my
legs.  I also am usually on my overprovisioned macbook pro, and it's hard for me to see how anyone gets any
work done in eclipse, intellij, or netbeans, although netbeans is the least slow.  All I can say is if you do
make any improvements to textmate's syntax highlighting or other scala-specific bits, at least one person will
use those improvements with great gusto.

BTW, with respect to adding some intelligence to your dev environment, I went with enhancing the interpreter
since at this point I have no confidence anything fancier than textmate will ever be appealing.  I'd hesitate
to call it "intelligence" but I added some completion bits which save me a lot of time.  I had intended to
submit it for trunk but I had to put it aside and haven't come back to it.  In the below log imagine yourself
hitting "tab" on each scala> line; the fields and method signatures are all discovered via reflection.  Also,
you can fire up the scaladoc/javadoc for a class in a web browser by hitting ctrl-T.


scala> import scala.col

scala> import scala.collection.

scala.collection.BitSet             scala.collection.Map                scala.collection.MapProxy
scala.collection.Ranged             scala.collection.RollbackIterator   scala.collection.Set
scala.collection.SetProxy           scala.collection.Sorted             scala.collection.SortedMap
scala.collection.SortedSet          scala.collection.immutable.         scala.collection.jcl.
scala.collection.mutable.

scala> import scala.collection.Set.

*              **             apply          contains       equals         hashCode       isEmpty        size
stringPrefix   subsetOf       toArray

scala> scala.collection.Set.contains

abstract def contains(A): Boolean

scala> java.lang.String

def this(Array[byte])
def this(StringBuffer)                    def this(StringBuilder)
def this(Int, Int, Array[char])           def this(Array[byte], Int, Int)
def this(Array[int], Int, Int)            def this(Array[byte], Int, Int, Int)
def this(Array[byte], Int)                def this(Array[byte], Int, Int, String)
def this(Array[byte], String)             def this(Array[char], Int, Int)
def this(String)                          def this()
def this(Array[char])

--
Paul Phillips      | Beware of bugs in the above code; I have only
Apatheist          | proved it correct, not tried it.
Empiricist         |     -- Knuth
up hill, pi pals!  |----------* http://www.improving.org/paulp/ *----------

The Eclipse IDE uses a separate pass to match parens between lexing and parsing, given that paren matching is a much more simple problem than parsing. 
Sean

On Thu, Jan 15, 2009 at 7:34 AM, Daniel Spiewak <djspiewak@gmail.com> wrote:

JLine does support tab completion.  I don't exactly know how the callback works, but I assume it's something that could be hooked into in order to provide the desired functionality.  I think the problem would be needing to incrementally parse partially-entered expressions.  Considering how stolid Scala's parser can be (especially with paren-matching), I wonder how possible this is.
Does anyone know the Scala shell code / parser well enough to give a definitive answer (possible/impossible) on this?

Daniel
On Wed, Jan 14, 2009 at 5:31 PM, Jorge Ortiz <jorge.ortiz@gmail.com> wrote:

I would love to have this enhancement available.

On Wed, Jan 14, 2009 at 2:57 PM, Paul Phillips <paulp@improving.org> wrote:

[Sorry if this comes through more than once, gmail is going freaky on me]

On Wed, Jan 14, 2009 at 02:11:28PM -0800, Alex Payne wrote:
> The main thing that's been frustrating is the lack of accurate syntax
> highlighting, both in TextMate and Emacs, though it's far worse in
> TextMate. I've already made one TextMate bundle (for the Puppet
> syntax) and I'm hesitant to repeat the process if something else works
> reliably and adds some intelligence to my development environment.

I have tried out all the big IDEs repeatedly, and always come running back to textmate with my tail between my
legs.  I also am usually on my overprovisioned macbook pro, and it's hard for me to see how anyone gets any
work done in eclipse, intellij, or netbeans, although netbeans is the least slow.  All I can say is if you do
make any improvements to textmate's syntax highlighting or other scala-specific bits, at least one person will
use those improvements with great gusto.

BTW, with respect to adding some intelligence to your dev environment, I went with enhancing the interpreter
since at this point I have no confidence anything fancier than textmate will ever be appealing.  I'd hesitate
to call it "intelligence" but I added some completion bits which save me a lot of time.  I had intended to
submit it for trunk but I had to put it aside and haven't come back to it.  In the below log imagine yourself
hitting "tab" on each scala> line; the fields and method signatures are all discovered via reflection.  Also,
you can fire up the scaladoc/javadoc for a class in a web browser by hitting ctrl-T.


scala> import scala.col

scala> import scala.collection.

scala.collection.BitSet             scala.collection.Map                scala.collection.MapProxy
scala.collection.Ranged             scala.collection.RollbackIterator   scala.collection.Set
scala.collection.SetProxy           scala.collection.Sorted             scala.collection.SortedMap
scala.collection.SortedSet          scala.collection.immutable.         scala.collection.jcl.
scala.collection.mutable.

scala> import scala.collection.Set.

*              **             apply          contains       equals         hashCode       isEmpty        size
stringPrefix   subsetOf       toArray

scala> scala.collection.Set.contains

abstract def contains(A): Boolean

scala> java.lang.String

def this(Array[byte])
def this(StringBuffer)                    def this(StringBuilder)
def this(Int, Int, Array[char])           def this(Array[byte], Int, Int)
def this(Array[int], Int, Int)            def this(Array[byte], Int, Int, Int)
def this(Array[byte], Int)                def this(Array[byte], Int, Int, String)
def this(Array[byte], String)             def this(Array[char], Int, Int)
def this(String)                          def this()
def this(Array[char])

--
Paul Phillips      | Beware of bugs in the above code; I have only
Apatheist          | proved it correct, not tried it.
Empiricist         |     -- Knuth
up hill, pi pals!  |----------* http://www.improving.org/paulp/ *----------

Sounds pretty interesting! Have you considered pushing the changes
out in a Git fork or something to that effect? That way, we can enjoy
the fruits of your labor even if the mighty Powers on High are
reluctant to integrate the patch.

Daniel

On Jan 14, 2009, at 5:47 PM, Paul Phillips wrote:

> On Wed, Jan 14, 2009 at 05:34:15PM -0600, Daniel Spiewak wrote:
>> JLine does support tab completion. I don't exactly know how the
>> callback works, but I assume it's something
>> that could be hooked into in order to provide the desired
>> functionality. I think the problem would be
>> needing to incrementally parse partially-entered expressions.
>> Considering how stolid Scala's parser can be
>> (especially with paren-matching), I wonder how possible this is.
>
> When I was working on this my expression ambition wasn't any deeper
> than getting this:
>
> val x: String = "foo"
> x.
>
> ...to show String's instance methods. Sadly even that turned out to
> require more abuse to the interpreter
> than I was up for at the time. The way the existing code works is
> to find the name of every class on the
> classpath on startup, distill out the package paths into a jline
> dictionary, and look up methods and fields on
> demand with Class.forName if you ask for them -- meaning it works
> from the name of the class, not from
> anything it figures out about your current repl session.
>
> The interpreter at present isn't written with the expectation that
> other scala code will be popping in to ask
> about types. I assume it wouldn't be too much work to get it to
> reveal the type of "x" in the above example,
> but I ran into some no-longer-remembered roadblock a few months ago,
> possibly just a reluctance to make deep
> changes because I thought it would impair my chances of getting a
> patch accepted.
>

On Wed, Jan 14, 2009 at 05:34:15PM -0600, Daniel Spiewak wrote:
> JLine does support tab completion. I don't exactly know how the callback works, but I assume it's something
> that could be hooked into in order to provide the desired functionality. I think the problem would be
> needing to incrementally parse partially-entered expressions. Considering how stolid Scala's parser can be
> (especially with paren-matching), I wonder how possible this is.

When I was working on this my expression ambition wasn't any deeper than getting this:

val x: String = "foo"
x.

...to show String's instance methods. Sadly even that turned out to require more abuse to the interpreter
....to show String's instance methods. Sadly even that turned out to require more abuse to the interpreter
than I was up for at the time. The way the existing code works is to find the name of every class on the
classpath on startup, distill out the package paths into a jline dictionary, and look up methods and fields on
demand with Class.forName if you ask for them -- meaning it works from the name of the class, not from
anything it figures out about your current repl session.

The interpreter at present isn't written with the expectation that other scala code will be popping in to ask
about types. I assume it wouldn't be too much work to get it to reveal the type of "x" in the above example,
but I ran into some no-longer-remembered roadblock a few months ago, possibly just a reluctance to make deep
changes because I thought it would impair my chances of getting a patch accepted.

On Wed, Jan 14, 2009 at 05:53:25PM -0600, Daniel Spiewak wrote:
> Sounds pretty interesting! Have you considered pushing the changes out in
> a Git fork or something to that effect? That way, we can enjoy the fruits
> of your labor even if the mighty Powers on High are reluctant to integrate
> the patch.

I have about eight git forks of the scala source going with different experiments in progress; I'm happy to
share them all with the two or three of us willing to maintain our own patch sets. I need to polish it up a
little and rebase it against trunk though. My github scala mirror is here:

http://github.com/paulp/scala

That's a straight mirror of the official svn right now, but I'll add a couple branches with my more plausibly
useful changes as soon as I can get to it.

Sounds great! I'm already working with a Git fork to patch the Vector
stuff; what's another branch among friends?

Daniel

On Jan 14, 2009, at 5:47 PM, Paul Phillips wrote:

> On Wed, Jan 14, 2009 at 05:34:15PM -0600, Daniel Spiewak wrote:
>> JLine does support tab completion. I don't exactly know how the
>> callback works, but I assume it's something
>> that could be hooked into in order to provide the desired
>> functionality. I think the problem would be
>> needing to incrementally parse partially-entered expressions.
>> Considering how stolid Scala's parser can be
>> (especially with paren-matching), I wonder how possible this is.
>
> When I was working on this my expression ambition wasn't any deeper
> than getting this:
>
> val x: String = "foo"
> x.
>
> ...to show String's instance methods. Sadly even that turned out to
> require more abuse to the interpreter
> than I was up for at the time. The way the existing code works is
> to find the name of every class on the
> classpath on startup, distill out the package paths into a jline
> dictionary, and look up methods and fields on
> demand with Class.forName if you ask for them -- meaning it works
> from the name of the class, not from
> anything it figures out about your current repl session.
>
> The interpreter at present isn't written with the expectation that
> other scala code will be popping in to ask
> about types. I assume it wouldn't be too much work to get it to
> reveal the type of "x" in the above example,
> but I ran into some no-longer-remembered roadblock a few months ago,
> possibly just a reluctance to make deep
> changes because I thought it would impair my chances of getting a
> patch accepted.
>

I had exactly the same needs. I was impressed how well the maven
plugin worked.

With intellij I found the need to tell it to compile scala first
(Which is not the default) so I could have fully mixed. I also needed
to add scala "facets" to each sub project (module) even if they were
only java source (known bug).

So is the netbeans one up with intellij? I found intellij quite good,
code nav, running unit tests etc all worked quite well.

Sent from my phone.

On 15/01/2009, at 6:56, "Putrycz, Erik"
wrote:

> I have just been trying to achieve the following and did a quick
> evaluation of IDEs:
> - maven project with sub projects
> - mix of java and scala code in one of the projects
> - java calling scala and scala calling java
>
> Eclipse:
> - little glitchy to configure IAM (maven plugin for eclipse) to work
> with scala and mix scala code and java code but is feasible.
> - Calling the scala code from java wasn't too straightforward, there
> are
> still some bugs in the eclipse plugin (I declared a Array[T] in scala
> which was seen as Array in the jdt)
> + Scala seems to compile incrementally which doesn't seem to happen
> with
> the other IDEs
>
> The IAM developers seem to be commited to get scala to work with the
> maven plugin and there is an effort of a scala plugin for IAM.
>
> IDEA:
> + no issue for having scala and java on a project
> - has been somehow unstable on linux
> - requires the scala compiler to be installed separately - and
> "building" my project caused a classnotfound after calling scalac
>
> Netbeans:
> + uses the maven builder for building projects (no weird scalac
> configuration)
> - the java classes did not see the scala classes (this may require a
> build which had syntax errors)
>
> At this point, the scala support did not seem any stronger in IDEA
> than
> Netbeans. I'm going to stick with the eclipse plugin for now and wait
> for fixes.
>
> Erik.
>> -----Original Message-----
>> From: Sciss [mailto:contact@sciss.de]
>> Sent: January-14-09 2:22 PM
>> To: Matt Harrington
>> Cc: scala-tools@listes.epfl.ch
>> Subject: Re: [scala-tools] Poll: what do you code Scala in?
>>
>> i have soylatte installed, but honestly x11 sucks for desktop
>> applications IMO... and i'm a java on the desktop person ;-) apple
>> should open source their Cocoa-AWT bridge and donate it to the
>> soylatte project.
>>
>> by the way scala runs fine on soylatte...
>>
>> ciao, -sciss-
>>
>>
>> Am 14.01.2009 um 20:14 schrieb Matt Harrington:
>>
>>> I'm in the same boat. Just to clarify for the readers-at-home,
> Apple
>>> only supplies JVM 1.5 on 32-bit Core and Core Duo Macs.
>>> First-generation Intel Macs are stuck with 1.5. Core2 Macs get 1.6.
>>>
>>> If you need 1.6 on a 32-bit Mac, look into the SoyLatte project.
> It's
>>> a port of OpenJDK BSD.
>>>
>>> Matt
>>>
>>>
>>> On Wed, Jan 14, 2009 at 6:12 AM, Sciss wrote:
>>>> only those stick to 1.5 that have the "wrong mac computer" (like
>>>> me... thank
>>>> you apple!)
>>>>
>>>> Am 14.01.2009 um 14:24 schrieb Sean McDirmid:
>>>>
>>>>> Not too many people even use 1.5 anymore, the current release is
>>>>> 1.6.
>

I've tried Eclipse, IDEA and NetBeans on OS X and Windows in the last few
weeks and come down strongly in favour of IDEA.

Eclipse was unusable because ctrl-space would sometimes cause the whole IDE
to hang completely for a minute at a time. This was happening every few
minutes so I quickly lost patience. But even before then I was often seeing
the editor breaking completely, flooding the error window with messages and
refusing to accept any input. Closing and reopening the file generally
fixed it but the whole thing was an exercise in frustration.

NetBeans occasionally showed a similar problem to Eclipse with locking up
but generally felt usable but a bit basic.

IDEA felt a lot more solid and polished than the other two and generally the
features that are included work well. The other two (and particularly
Eclipse) felt half-finished in places.

Chris

On Thu, Jan 15, 2009 at 8:46 AM, cjkent wrote:
> The other two (and particularly Eclipse) felt half-finished in places.

Probably because it was. It's now about two thirds finished and I
recommend you give it another try ;-)

Cheers,

Miles

>Probably because it was. It's now about two thirds finished and I
>recommend you give it another try ;-)
So i just did. I was using some 2.7.2 Version.
1. I try to open ANY of my existing sources -> Editor crash/fail
2. I press F5 to refresh, all my sources are gone!Puff!
3. the project is empty so i click on create new Scala Applciation. Because
of "default" package the complie crashes. I create a new package named
"basics" i try to drag the Scala File from default to basics. The process
crashes gives me an error creates a Java?! and a scala file in the basics
package...
4. I go back to netbeans :D

1. You need a project to edit your sources. Otherwise, how can the Eclipse IDE magically figure out the classpath you intended for the file? 2. Have no clue. 3. This is BS, the Scala application wizard won't let you create a Scala application without specifying a package.  4. If NetBeans works for you don't change. Eclipse is for people who have to hack on real projects, say 2000 line Scala files. If you are creating small hello world Scala files in default packages without any formal project setup, I don't think Eclipse is for you. 
On Fri, Jan 16, 2009 at 6:42 AM, andreas s. <andreas_scheinert@web.de> wrote:

>Probably because it was. It's now about two thirds finished and I
>recommend you give it another try ;-)
So i just did. I was using some 2.7.2 Version.
1. I try to open ANY of my existing sources -> Editor crash/fail
2. I press F5 to refresh, all my sources are gone!Puff!
3. the project is empty so i click on create new Scala Applciation. Because
of "default" package the complie crashes. I create a new package named
"basics" i try to drag the Scala File from default to basics. The process
crashes gives me an error creates a Java?! and a scala file in the basics
package...
4. I go back to netbeans :D
--
View this message in context: http://www.nabble.com/-scala-tools--Poll%3A-what-do-you-code-Scala-in--tp21444077p21488496.html
Sent from the Scala - Tools mailing list archive at Nabble.com.

On Thu, Jan 15, 2009 at 11:56 PM, Sean McDirmid wrote:

I think what Sean _meant_ to say was: thanks for your feedback, and if
you can provide us with reproducible test cases for the problems
you've been having we'll do our best to sort them out ;-)

Cheers,

Miles

I added some comments inline.
On Jan 15, 2009, at 6:56 PM, Sean McDirmid <sean.mcdirmid@gmail.com> wrote:

1. You need a project to edit your sources. Otherwise, how can the Eclipse IDE magically figure out the classpath you intended for the file?

For most other eclipse editors, it at least provides syntax highlighting and the contents of the file

 2. Have no clue. 

Sounds like your project does not have a scala nature.  If it does, then I have no clue.

3. This is BS, the Scala application wizard won't let you create a Scala application without specifying a package. 

You can create files without using the scala wizard.  In this case, the scala plugin *does* freak out.  If I had enough motivation, I'd attempt to fix this, but to be honest Miles fixed enough bugs that I naturally avoid doing thing like using non-scala wizards to create scala files, and it doesn't bother me except for once in awhile.

4. If NetBeans works for you don't change. Eclipse is for people who have to hack on real projects, say 2000 line Scala files. If you are creating small hello world Scala files in default packages without any formal project setup, I don't think Eclipse is for you. 
On Fri, Jan 16, 2009 at 6:42 AM, andreas s. < (andreas_scheinert [at] web [dot] de> wrote:

>Probably because it was. It's now about two thirds finished and I
>recommend you give it another try ;-)
So i just did. I was using some 2.7.2 Version.
1. I try to open ANY of my existing sources -> Editor crash/fail
2. I press F5 to refresh, all my sources are gone!Puff!
3. the project is empty so i click on create new Scala Applciation. Because
of "default" package the complie crashes. I create a new package named
"basics" i try to drag the Scala File from default to basics. The process
crashes gives me an error creates a Java?! and a scala file in the basics
package...
4. I go back to netbeans :D
--
View this message in context: http://www.nabble.com/-scala-tools--Poll%3A-what-do-you-code-Scala-in--tp21444077p21488496.html
Sent from the Scala - Tools mailing list archive at Nabble.com.

Actually, I meant we haven't optimized for people just playing around with Scala yet. This is probably something that needs to be fixed, but no one has fixed it yet. Instead, we've focused on the large project use cases, trying to get that to work right. 

On Fri, Jan 16, 2009 at 9:56 AM, Miles Sabin <miles@milessabin.com> wrote:

On Thu, Jan 15, 2009 at 11:56 PM, Sean McDirmid <sean.mcdirmid@gmail.com> wrote:
<snip/>

I think what Sean _meant_ to say was: thanks for your feedback, and if
you can provide us with reproducible test cases for the problems
you've been having we'll do our best to sort them out ;-)

Cheers,


Miles

--
Miles Sabin
tel:    +44 (0)1273 720 779
mobile: +44 (0)7813 944 528
skype:  milessabin

1. You need a project to edit your sources. Otherwise, how can the Eclipse IDE magically figure out the classpath you intended for the file?

For most other eclipse editors, it at least provides syntax highlighting and the contents of the file

Possibly, but then we have to disable a lot of functionality (e.g., code completion). We could have a default project hanging around with a basic default classpath for people who just want to open up/edit files, and then hide semantic errors in these files. 

 2. Have no clue. 

Sounds like your project does not have a scala nature.  If it does, then I have no clue. 

3. This is BS, the Scala application wizard won't let you create a Scala application without specifying a package. 

You can create files without using the scala wizard.  In this case, the scala plugin *does* freak out.  If I had enough motivation, I'd attempt to fix this, but to be honest Miles fixed enough bugs that I naturally avoid doing thing like using non-scala wizards to create scala files, and it doesn't bother me except for once in awhile.

I think the original message specifically said "created a Scala Application", so I was assuming he did this through the menu system.  

On Fri, Jan 16, 2009 at 12:15 AM, Sean McDirmid <sean.mcdirmid@gmail.com> wrote:

1. You need a project to edit your sources. Otherwise, how can the Eclipse IDE magically figure out the classpath you intended for the file?

For most other eclipse editors, it at least provides syntax highlighting and the contents of the file

Possibly, but then we have to disable a lot of functionality (e.g., code completion). We could have a default project hanging around with a basic default classpath for people who just want to open up/edit files, and then hide semantic errors in these files. 

I don't see why it's hard to do what almost every other language extension in Eclipse does.... Yes it would require a bit of rearchitecting the current plugin, but I believe the user experience would be far greater (in my opinion).  I think things like code-completion and syntax highlighting should never get in the way of text editing, which the SDT editor frequently does.  I'm hoping that with time the bugs are worked out, but I'm not so sure the approach (re-using the language parser for the IDE) has proven itself viable.

 2. Have no clue. 

Sounds like your project does not have a scala nature.  If it does, then I have no clue. 

3. This is BS, the Scala application wizard won't let you create a Scala application without specifying a package. 

You can create files without using the scala wizard.  In this case, the scala plugin *does* freak out.  If I had enough motivation, I'd attempt to fix this, but to be honest Miles fixed enough bugs that I naturally avoid doing thing like using non-scala wizards to create scala files, and it doesn't bother me except for once in awhile.

I think the original message specifically said "created a Scala Application", so I was assuming he did this through the menu system. 

You know what they say about asumming... it makes an ass of u and me.

 

 

Josh Suereth wrote:
>
> On Fri, Jan 16, 2009 at 12:15 AM, Sean McDirmid
> wrote:
>
>>
>> 1. You need a project to edit your sources. Otherwise, how can the
>> Eclipse
>>> IDE magically figure out the classpath you intended for the file?
>>>
>>> For most other eclipse editors, it at least provides syntax highlighting
>>> and the contents of the file
>>>
>>
>> Possibly, but then we have to disable a lot of functionality (e.g., code
>> completion). We could have a default project hanging around with a basic
>> default classpath for people who just want to open up/edit files, and
>> then
>> hide semantic errors in these files.
>>
>
> I don't see why it's hard to do what almost every other language extension
> in Eclipse does.... Yes it would require a bit of rearchitecting the
> current
> plugin, but I believe the user experience would be far greater (in my
> opinion). I think things like code-completion and syntax highlighting
> should never get in the way of text editing, which the SDT editor
> frequently
> does. I'm hoping that with time the bugs are worked out, but I'm not so
> sure the approach (re-using the language parser for the IDE) has proven
> itself viable.
>
>> 2. Have no clue.
>>>
>>> Sounds like your project does not have a scala nature. If it does, then
>>> I
>>> have no clue.
>>>
>>
>>> 3. This is BS, the Scala application wizard won't let you create a Scala
>>> application without specifying a package.
>>>
>>>
>>> You can create files without using the scala wizard. In this case, the
>>> scala plugin *does* freak out. If I had enough motivation, I'd attempt
>>> to
>>> fix this, but to be honest Miles fixed enough bugs that I naturally
>>> avoid
>>> doing thing like using non-scala wizards to create scala files, and it
>>> doesn't bother me except for once in awhile.
>>>
>>
>> I think the original message specifically said "created a Scala
>> Application", so I was assuming he did this through the menu system.
>>
>
> You know what they say about asumming... it makes an ass of u and me.
>
>
>
>>
>>
>
>
hm i see i need to be a bit more specific:
I had already an existing scala project with existing packages. And to
create the Scala Application i used right click + Scala Application. When i
find some time i try some other things too and report.

On Wed, Jan 14, 2009 at 03:31:16PM -0800, Jorge Ortiz wrote:
> I would love to have this enhancement available.

OK, as indicated I have added a branch called "repl-completion" to the repo here:

http://github.com/paulp/scala

It comes with a jar which contains most of the support code, which if you want to build it yourself is here:

http://github.com/paulp/scala-improving

(It's called "scala-improving" not because I am presumptuous but because, well, look at my email address.)

While I'm at it I have another not really announced repo here, a rewrite of fjbg in scala I called ssbg:

http://github.com/paulp/ssbg

I spent a lot of time tweaking it trying to make it as fast as fjbg and never got there, although I was close
by the time I'd removed every scala-ish feature and made it look like semicolon-free java. Maybe some of the
optimizations being investigated for 2.8 will help, but I despaired of continuing since my own motivation
involves using scala's appealing features and it was only going to get slower down that road. Also, there
were some big timing differences I didn't finish figuring out depending on which ant task I used (i.e. ssbg
and fjbg would have similar build times under one of scalacfork/fsc/scala, but then ssbg would be much slower
under one of the others.)

I make no assurances about how well any of this works, especially because both ssbg and the repl code have
been sitting untouched for weeks, but I built the repl and it did about what I remembered. If anyone on the
scala team would like to give a little guidance about what they'd want to see for that to go into trunk, I can
look into accomodating.

On Fri, Jan 16, 2009 at 8:47 PM, Josh Suereth <joshua.suereth@gmail.com> wrote:

I don't see why it's hard to do what almost every other language extension in Eclipse does.... Yes it would require a bit of rearchitecting the current plugin, but I believe the user experience would be far greater (in my opinion).  I think things like code-completion and syntax highlighting should never get in the way of text editing, which the SDT editor frequently does.  I'm hoping that with time the bugs are worked out, but I'm not so sure the approach (re-using the language parser for the IDE) has proven itself viable.

No its not difficult at all to do this. We just have to have someone do it: create a default project profile somewhere, and when a project-less file is open use that profile (sit in the debugger to figure out where the crashes occur and redirect). 
Re-using the language parser in the IDE is the easy part (well, if you ignore XML). The place where we break down is that the type checker evolves without concern for the IDE. I also don't know if the approach is viable, I think its technically viable, but it might not be socially viable (e.g., getting the compiler and IDE teams to work together). Scala is very much meant as a command line language. 

You know what they say about asumming... it makes an ass of u and me.

I'm pretty sure this is what he meant.