Hi,

I just discovered Io, and I found it great, it's just the language I've
been waiting for. I just have little questions :

Apparently, the end of line is seen as a separator like the semicolon.
Why ? I understand that putting semicolons at end of each line can be
very painful but doesn't this rule create problems when you want to
split long messages across multiples lines ?
Is it even possible ?

And another tiny question, I came from the lua world and I like very
much the portability of the lua code. I know lua can be run almost
everywhere ... even on small embedded devices. Because of the C source
and the wery small footprint.
Even if I don't need it (yet) I want to know if Io is as small and
portable.

Thanks
--
Mildred       <xmpp:mildred@...> <http://mildred632.free.fr/>
Clef GPG :    <hkp://pgp.mit.edu> ou <http://mildred632.free.fr/gpg_key>
Fingerprint : 197C A7E6 645B 4299 6D37 684B 6F9D A8D6 [9A7D 2E2B]

I'm looking for a call that can run a command and get the stdout
results that it spits out, as opposed to just the return value. Is
there any such thing in Io, or should I just pipe the output to a
file and read from there?

Thanks,

Morgan Aldridge
--
morgant@...
http://www.makkintosshu.com/

On 17 Nov 2006, at 11:48 am, Mildred wrote:
> Apparently, the end of line is seen as a separator like the semicolon.
> Why ? I understand that putting semicolons at end of each line can be
> very painful but doesn't this rule create problems when you want to
> split long messages across multiples lines ?
> Is it even possible ?

That's why word processors have line wrapping. ;-)

Here's one trick:

unsplit:

123 sin cos println

split:

123 sin(
) cos println

Maybe we should make \ used for line wrapping (this could actually be
implemented in Io).

> And another tiny question, I came from the lua world and I like very
> much the portability of the lua code. I know lua can be run almost
> everywhere ... even on small embedded devices. Because of the C source
> and the wery small footprint.
> Even if I don't need it (yet) I want to know if Io is as small and
> portable.

Yes, though you may have to add a coro entry for your processor if
it's not on an OS that supports the ucontext or fiber APIs.

- Steve

On 06-11-17, at 14:48, Mildred wrote:

> Hi,
>
> And another tiny question, I came from the lua world and I like very
> much the portability of the lua code. I know lua can be run almost
> everywhere ... even on small embedded devices. Because of the C source
> and the wery small footprint.
> Even if I don't need it (yet) I want to know if Io is as small and
> portable.

Io is nowhere near as light on memory as lua is. Infact, Io is about
as heavy on memory as an equivalent program written in Python and
Ruby *combined*. While this may sound bad, for most embedded devices,
it's not an issue except if you're trying to stuff Io into NOR --
then good luck.

That said, I am working on a dialect of Io which is much lighter, and
in fact, targeted towards small memory devices; but it's nowhere near
usable at the moment (still working out the concurrency model). In
any event, it'll be about as light as lua, if not a little lighter
(early estimates, take with a grain of salt).

>
> Thanks
> --
> Mildred       <xmpp:mildred@...> <http://mildred632.free.fr/>
> Clef GPG :    <hkp://pgp.mit.edu> ou <http://mildred632.free.fr/
> gpg_key>
> Fingerprint : 197C A7E6 645B 4299 6D37 684B 6F9D A8D6 [9A7D 2E2B]
>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
> !DSPAM:455e2e04954941604914933!
>


--
Jeremy Tregunna
jtregunna@...

On 06-11-09, at 16:05, Morgan Aldridge wrote:

> I'm looking for a call that can run a command and get the stdout
> results that it spits out, as opposed to just the return value. Is
> there any such thing in Io, or should I just pipe the output to a
> file and read from there?

You could do one of two things; use File popen, or SystemCall -- both
of which don't currently work on Windows. (popen should work on NT+
maybe)

>
> Thanks,
>
> Morgan Aldridge
> --
> morgant@...
> http://www.makkintosshu.com/
>
>
>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
> !DSPAM:455e2e50955041605182660!
>


--
Jeremy Tregunna
jtregunna@...

On Nov 17, 2006, at 4:15 PM, Brian L. wrote:
> So if I were to theoretically force a full garbage collection
> between each of the lines, the weakrefs would die along with my 'f'
> property because there would be nothing to keep the string alive
> except for the symbol table (implying that the behavior in this
> example can't be depended upon in general).

String literal's are referenced by the Message objects that cache
them. Example:

	 Io> message("foo") cachedResult
	 ==> foo

Likewise for number literals:

	 Io> message(123) cachedResult
	 ==> 123

While the result of a .. is only referenced if you keep a reference
to it.

Here's how you can test this:

a := method("foobar123")
getSlot("a") message cachedResult uniqueId
"foobar123" uniqueId // it's the same because the symbol was in the
symbol table
removeSlot("a") // removes the reference to that symbol
Collector collect
"foobar123" uniqueId // a different id because it's a new symbol

- Steve

On 9 Nov 2006, at 01:05 pm, Morgan Aldridge wrote:
> I'm looking for a call that can run a command and get the stdout
> results that it spits out, as opposed to just the return value. Is
> there any such thing in Io, or should I just pipe the output to a
> file and read from there?


SystemCall clone setCommand("ls") run stdout readLines foreach(println)

I notice that the removal of ':' break ObjcBridge:

  2006 11 22 steve removed : from identifier character list - it's now
a proper operator

':' is commonly used in objc methods.
Any solution ?
Thanx.

Yen-Ju

Hi Yen-Ju,

We're scheming about supporting obj-c style method calls. [object
deleteFrom:start to:end]. Hopefully the : handling won't be broken for
long.

Thanks,
Jonathan.

On 11/25/06, Yen-Ju Chen <yjchenx@...> wrote:
> I notice that the removal of ':' break ObjcBridge:
>
>  2006 11 22 steve removed : from identifier character list - it's now
> a proper operator
>
> ':' is commonly used in objc methods.
> Any solution ?
> Thanx.
>
> Yen-Ju
>
>
>
> Yahoo! Groups Links
>
>
>
>

Color used to be a prototype object, and in the online reference it still is
(under Math), but
when I try to run some old scripts that use Color, using the current darcs Io on
a PPC Mac,
then I get errors. What's happening?

/Jon

It is good to hear that.
Thanks.

Yen-Ju

On 11/24/06, Quag <quaggy@...> wrote:
>
> Hi Yen-Ju,
>
>  We're scheming about supporting obj-c style method calls. [object
>  deleteFrom:start to:end]. Hopefully the : handling won't be broken for
>  long.
>
>  Thanks,
>  Jonathan.
>
>
>  On 11/25/06, Yen-Ju Chen <yjchenx@...> wrote:
>  > I notice that the removal of ':' break ObjcBridge:
>  >
>  >  2006 11 22 steve removed : from identifier character list - it's now
>  > a proper operator
>  >
>  > ':' is commonly used in objc methods.
>  > Any solution ?
>  > Thanx.
>  >
>  > Yen-Ju
>  >
>  >
>  >
>  > Yahoo! Groups Links
>  >
>  >
>  >

Hi Chris, thanks for your reply. So you did a port to symbian? Could
you point me in the right direction in terms of starting this off? If
I understand correctly, the core of IO doesn't really care about the
availability of posix support or the FS, correct? If I wanted to
simply get the VM compiled for the DS, then hack up a simple
interactive console that pipes lines of code into the VM, do you have
any recommendations on where I should begin the task?

thanks

slim

--- In iolanguage@yahoogroups.com, "Chris Double" <chris.double@...>
wrote:
>
> On 9/23/06, dJsLiM <djslim@...> wrote:
> > Hi guys, I was wondering if anyone has tried porting IO to the
Nintendo
> >  DS?
>
> You would probably have to compile a cut down version similar to the
> Symbian port. The DS doesn't have a file system and no posix
support,
> although devkitpro does attempt to emulate some functionality.
>
> Lua has been ported so it certainly seems possible for Io too:
> http://www.dslua.com/
>
> I anyone's interested and have some information on Nintendo DS
> homebrew development here:
>
> http://www.double.co.nz/nintendo_ds
>
> Chris
> --
> http://www.bluishcoder.co.nz
>

Hi,

I have pg in /Library/PostgreSQL8 (symlink from /usr/local/pgsql) so
the current "Io/addons/Postgres/build.io" will not find it. Added the
following:

	 if(platform == "darwin", appendLibSearchPath("/usr/local/pgsql/lib"))
	 if(platform == "darwin", appendHeaderSearchPath("/usr/local/pgsql/
include"))

Not optimal but handles a non-uncommon pg installation location w/
MacOS X.

Best Wishes,
Kon

Hi,

- PortAudio

   Since I have V19 w/ all the API incompatibilities no surprise here.
Be nice though if a method of determining the pa version at compile
time was possible (I only see a runtime method). Seems(?) the actual
code changes to AudioDevice.[ch] are minor otherwise.

Best Wishes,
Kon

Hi,

Needed to add:

	 if(platform == "darwin", dependsOnFramework("AGL"))

to the GLFW build.io.

Also, anyone have problems using the default build of glfw on MacOS
X? I had to make some changes to stop the dreaded "common symbols not
allowed with MH_DYLIB output format" link error.

Best Wishes,
Kon

Hi,

To have libstdc++.dylib found for SoundTouch I added:

	 if(platform == "darwin", appendLibSearchPath("/usr/lib/gcc/powerpc-
apple-darwin8/4.0.1"))

to Io/addons/SoundTouch/build.io

Obviously sub-optimal.

Best Wishes,
Kon

Hi,

When all addons build:

Io/Makefile
86c86
<       cat errors
---
  >       @if [ -f errors ]; then cat errors; fi

Best Wishes,
Kon

On Dec 1, 2006, at 7:05 PM, Kon Lovett wrote:

> Hi,
>
> Needed to add:
>
> if(platform == "darwin", dependsOnFramework("AGL"))
>
> to the GLFW build.io.
>
> Also, anyone have problems using the default build of glfw on MacOS
> X? I had to make some changes to stop the dreaded "common symbols not
> allowed with MH_DYLIB output format" link error.

Baptiste Heymann and I fixed the GLFW building issues today.  The
patches are in the public repo.


> Best Wishes,
> Kon

Trevor Fancher

On Dec 1, 2006, at 8:16 PM, Kon Lovett wrote:

> Hi,
>
> When all addons build:
>
> Io/Makefile
> 86c86
> <       cat errors
> ---
>>       @if [ -f errors ]; then cat errors; fi

I was actually complaining about that today, but was too lazy to make a
patch.  I added the change to my public repo.  Should be in the main
repo shortly.


> Best Wishes,
> Kon

Thanks,
Trevor Fancher

I can build the DBI addon, but when I attempt to run the samples
included in addons/DBI/samples/, I get:

   Exception: *** IoDBI error during dbi_initialize

I've got a from-source install of libdbi-0.8.1.
Any ideas?  Is the addon not complete?

-Erik

On Dec 1, 2006, at 3:42 PM, Kon Lovett wrote:

> Hi,
>
> I have pg in /Library/PostgreSQL8 (symlink from /usr/local/pgsql) so
> the current "Io/addons/Postgres/build.io" will not find it. Added the
> following:
>
>  if(platform == "darwin", appendLibSearchPath("/usr/local/pgsql/lib"))
>  if(platform == "darwin", appendHeaderSearchPath("/usr/local/pgsql/
> include"))
>
> Not optimal but handles a non-uncommon pg installation location w/
> MacOS X.

Thanks, I've added your change. Btw, darcs is also handy for
submitting fixes.

Cheers,
- Steve

I got an error:

   Exception: Message does not respond to 'setAttached'
   ---------
   Message setAttached                 A4_Exception.io 11


A := Object clone do(
  f := method(x,
   x writeln
  )
)

B := Object clone do(
  a := nil
  forward := method(
   call delegateTo(self a)
  )
  init := method(
   self a = A clone
   f("error")
  )
)

b := B clone
---

Is it a bug, or am i abusing Io here?

On Dec 3, 2006, at 4:05 AM, Bart Botma wrote:

>   Exception: Message does not respond to 'setAttached'

Hi Bart,

Messages no longer have attached messages. Now they only have "next"
Messages and end-of-lines (";") are explicit Messages.

- Steve

Hi Bart,

After the next/attached change, there was a bug in delegateTo. The
delegateTo bug has been fixed in a later patch. If you don't want to
pull the fix using darcs, change delegateTo in
libs/iovm/A4_Exception.io to:

         delegateTo := method(target, altSender,
                 target doMessage(self message clone setNext, altSender
ifNilEval(self sender))
         )

         delegateToMethod := method(target, methodName,
                 target doMessage(self message clone setNext
setName(methodName), self sender)
         )


Thanks,
Jonathan.

On 12/4/06, Bart Botma <bart_botma@...> wrote:
> I got an error:
>
>   Exception: Message does not respond to 'setAttached'
>   ---------
>   Message setAttached                 A4_Exception.io 11
>
>
> A := Object clone do(
>  f := method(x,
>   x writeln
>  )
> )
>
> B := Object clone do(
>  a := nil
>  forward := method(
>   call delegateTo(self a)
>  )
>  init := method(
>   self a = A clone
>   f("error")
>  )
> )
>
> b := B clone
> ---
>
> Is it a bug, or am i abusing Io here?
>
>
>
>
> Yahoo! Groups Links
>
>
>
>

Thanks, this solved the problem. Strange however that my darcs get or
pull does not (yet) contain the fix you mentioned.

Remark:
Maybe in the future I should also distribute a frozen "full" Io
version with any Io programs I release to public repositories to avoid
these kinds of problems for end users. Actually, a program I released
to sourceforge recently contained a minimal Io runtime but after
giving it a second thought I decided to avoid license issues and
removed the Io runtime from the distribution.

Bart Botma

--- In iolanguage@yahoogroups.com, Quag <quaggy@...> wrote:
>
> Hi Bart,
>
> After the next/attached change, there was a bug in delegateTo. The
> delegateTo bug has been fixed in a later patch. If you don't want to
> pull the fix using darcs, change delegateTo in
> libs/iovm/A4_Exception.io to:
>
>         delegateTo := method(target, altSender,
>                 target doMessage(self message clone setNext, altSender
> ifNilEval(self sender))
>         )
>
>         delegateToMethod := method(target, methodName,
>                 target doMessage(self message clone setNext
> setName(methodName), self sender)
>         )
>
>
> Thanks,
> Jonathan.
>
> On 12/4/06, Bart Botma <bart_botma@...> wrote:
> > I got an error:
> >
> >   Exception: Message does not respond to 'setAttached'
> >   ---------
> >   Message setAttached                 A4_Exception.io 11
> >
> >
> > A := Object clone do(
> >  f := method(x,
> >   x writeln
> >  )
> > )
> >
> > B := Object clone do(
> >  a := nil
> >  forward := method(
> >   call delegateTo(self a)
> >  )
> >  init := method(
> >   self a = A clone
> >   f("error")
> >  )
> > )
> >
> > b := B clone
> > ---
> >
> > Is it a bug, or am i abusing Io here?
> >
> >
> >
> >
> > Yahoo! Groups Links
> >
> >
> >
> >
>

On 3 Dec 2006, at 11:53 am, Bart Botma wrote:

> Thanks, this solved the problem. Strange however that my darcs get or
> pull does not (yet) contain the fix you mentioned.
>
> Remark:
> Maybe in the future I should also distribute a frozen "full" Io
> version with any Io programs I release to public repositories to avoid
> these kinds of problems for end users. Actually, a program I released
> to sourceforge recently contained a minimal Io runtime but after
> giving it a second thought I decided to avoid license issues and
> removed the Io runtime from the distribution.

Io is BSD licensed, so there shouldn't be any license issues.

- Steve

Hi Bart,

On 12/4/06, Bart Botma <bart_botma@...> wrote:
> Thanks, this solved the problem. Strange however that my darcs get or
> pull does not (yet) contain the fix you mentioned.

Maybe the patch hasn't made it into Steve's repo yet. It will the next
time he pulls.

> Remark:
> Maybe in the future I should also distribute a frozen "full" Io
> version with any Io programs I release to public repositories to avoid
> these kinds of problems for end users. Actually, a program I released
> to sourceforge recently contained a minimal Io runtime but after
> giving it a second thought I decided to avoid license issues and
> removed the Io runtime from the distribution.

What program did you put up on sourceforge with Io in it?

It is a good idea to embed Io within your application. As Io is a long
way off being mature we regularly make breaking changes. Having your
own copy is important for keeping your app running.

Thanks,
Jonathan.

> Bart Botma
>
> --- In iolanguage@yahoogroups.com, Quag <quaggy@...> wrote:
> >
> > Hi Bart,
> >
> > After the next/attached change, there was a bug in delegateTo. The
> > delegateTo bug has been fixed in a later patch. If you don't want to
> > pull the fix using darcs, change delegateTo in
> > libs/iovm/A4_Exception.io to:
> >
> >         delegateTo := method(target, altSender,
> >                 target doMessage(self message clone setNext, altSender
> > ifNilEval(self sender))
> >         )
> >
> >         delegateToMethod := method(target, methodName,
> >                 target doMessage(self message clone setNext
> > setName(methodName), self sender)
> >         )
> >
> >
> > Thanks,
> > Jonathan.
> >
> > On 12/4/06, Bart Botma <bart_botma@...> wrote:
> > > I got an error:
> > >
> > >   Exception: Message does not respond to 'setAttached'
> > >   ---------
> > >   Message setAttached                 A4_Exception.io 11
> > >
> > >
> > > A := Object clone do(
> > >  f := method(x,
> > >   x writeln
> > >  )
> > > )
> > >
> > > B := Object clone do(
> > >  a := nil
> > >  forward := method(
> > >   call delegateTo(self a)
> > >  )
> > >  init := method(
> > >   self a = A clone
> > >   f("error")
> > >  )
> > > )
> > >
> > > b := B clone
> > > ---
> > >
> > > Is it a bug, or am i abusing Io here?
> > >
> > >
> > >
> > >
> > > Yahoo! Groups Links
> > >
> > >
> > >
> > >
> >
>
>
>
>
>
> Yahoo! Groups Links
>
>
>
>

Steve,

Thanks much for the update. I was still running a previous release
that didn't have the SystemCall add-on (and the docs on the site at
the time didn't list it, so I was trying to work with System alone).

Morgan Aldridge
--
morgant@...
http://www.makkintosshu.com/


On Nov 17, 2006, at 5:15 PM, Steve Dekorte wrote:

>
> On 9 Nov 2006, at 01:05 pm, Morgan Aldridge wrote:
> > I'm looking for a call that can run a command and get the stdout
> > results that it spits out, as opposed to just the return value. Is
> > there any such thing in Io, or should I just pipe the output to a
> > file and read from there?
>
> SystemCall clone setCommand("ls") run stdout readLines foreach
> (println)
>
>