Fwd: Re: Fwd: Fwd: NetRexx work for 2014

[Kermit Kiser]

Oops! Forgot to cc the list. :-)

-------- Original Message --------

Subject:	Re: Fwd: [Ibm-netrexx] Fwd: NetRexx work for 2014
Date:	Mon, 14 Apr 2014 09:04:13 -0700
From:	Kermit Kiser [hidden email]
To:	René Jansen [hidden email]

René --

I understood that there might be problems with the scripting implementation. That is one reason why it is on my list of things to look at. Another reason is that implementing JSR199 support would allow the JSR223 module to actually compile scripts to Java classes, cache them in memory and invoke them directly without interpreting them. That could improve performance although I am not sure it will fix the other issues. I am also looking at other JSR223 implementations to see how they handle script name conventions, context etc. Meanwhile the current implementation works well in some environments like Ant so I am happy with whatever you decide.

If I understand your other thought correctly, you are thinking of releasing the current SVN trunk as 3.03 now and perhaps another later release this year as 3.04 or 3.1 - is that correct? The only problem I have with that is that the current trunk does not build cleanly on Windows as I just commented on issue #102. The build fail is in the test module DiagUTF8 so it is not critical but it implies that some programs that compiled before may not compile with that code, at least on Windows systems. I would suggest backing out the change to RxVarpool until we can look at it closer if you want to do a quick 3.03 release meanwhile. What do you think about that approach?

-- Kermit

On 4/14/2014 4:36 AM, René Jansen wrote:

Kermit,

guessing that this list is just as good as the developer list to continue the discussion:

I have taken a look at the amount of things that are fixed, not only code but also documentation, examples, that I am now against doing a quick 'replacement' fix.

The one thing that held up the release of the scripting support was the fact that its context does not hold up across different invocations from the command line in a REPL loop. But I can live with that as a temporary restriction.

In your, opninion, what is holding us back to release the (vastly improved over 3.02GA) 3.03 as is?

best regards,

René Jansen.

Begin forwarded message:

From: René Jansen <[hidden email]>

Subject: Re: [Ibm-netrexx] Fwd: NetRexx work for 2014

Date: 14 april 2014 13:08:06 CEST

To: IBM Netrexx <[hidden email]>

Reply-To: IBM Netrexx <[hidden email]>

I am currently packaging the "Java 8 compatible" version, which is the 3.02GA version + the one necessary patch for the new bytecode introduced in 7 but used in 8 that Kermit prepared.

What do we all think the version number should be?

One way is to call it 3.021

Another way is to call it 3.03 and make the next update 3.1 (but that clashes with the mainframe convention to call the "3.0" version 3.1 or 4.0

I expected the 3.03 version to have the scripting enhancements, but they are not really releaseable yet. I think with the proposed enhancements, that version (with the do binary blocks etc) deserves a 4.0 name.

I'll consider all input but I am trying to release, including the new eclipse compiler, today CET, so Tom can have a go at updating the GUI installer.

best regards,

René.


On 14 apr. 2014, at 12:05, Kermit Kiser <[hidden email]> wrote:

All list members are hereby invited to help develop the next version of NetRexx!

_______________________________________________
Ibm-netrexx mailing list
[hidden email]
Online Archive : http://ibm-netrexx.215625.n3.nabble.com/

_______________________________________________
Ibm-netrexx mailing list
[hidden email]
Online Archive : http://ibm-netrexx.215625.n3.nabble.com/

[ThSITC]

Whil'st I shall not there on the 2014  *RexxLA meeting* in the US (due to budget constraints, as always the past Years),
it might and shall be a *good time* to release NetRexx 3.03 on this meeting :-)

Kermit & Rene,
My applause to all the work You did and are doing!

Kind regards,
Thomas Schneider, from Vienna.
====================================================================================

Am 14.04.2014 18:10, schrieb Kermit Kiser:

Oops! Forgot to cc the list. :-)

-------- Original Message --------

Subject:	Re: Fwd: [Ibm-netrexx] Fwd: NetRexx work for 2014
Date:	Mon, 14 Apr 2014 09:04:13 -0700
From:	Kermit Kiser [hidden email]
To:	René Jansen [hidden email]

René --

I understood that there might be problems with the scripting implementation. That is one reason why it is on my list of things to look at. Another reason is that implementing JSR199 support would allow the JSR223 module to actually compile scripts to Java classes, cache them in memory and invoke them directly without interpreting them. That could improve performance although I am not sure it will fix the other issues. I am also looking at other JSR223 implementations to see how they handle script name conventions, context etc. Meanwhile the current implementation works well in some environments like Ant so I am happy with whatever you decide.

If I understand your other thought correctly, you are thinking of releasing the current SVN trunk as 3.03 now and perhaps another later release this year as 3.04 or 3.1 - is that correct? The only problem I have with that is that the current trunk does not build cleanly on Windows as I just commented on issue #102. The build fail is in the test module DiagUTF8 so it is not critical but it implies that some programs that compiled before may not compile with that code, at least on Windows systems. I would suggest backing out the change to RxVarpool until we can look at it closer if you want to do a quick 3.03 release meanwhile. What do you think about that approach?

-- Kermit

On 4/14/2014 4:36 AM, René Jansen wrote:

Kermit,

guessing that this list is just as good as the developer list to continue the discussion:

I have taken a look at the amount of things that are fixed, not only code but also documentation, examples, that I am now against doing a quick 'replacement' fix.

The one thing that held up the release of the scripting support was the fact that its context does not hold up across different invocations from the command line in a REPL loop. But I can live with that as a temporary restriction.

In your, opninion, what is holding us back to release the (vastly improved over 3.02GA) 3.03 as is?

best regards,

René Jansen.

Begin forwarded message:

From: René Jansen <[hidden email]>

Subject: Re: [Ibm-netrexx] Fwd: NetRexx work for 2014

Date: 14 april 2014 13:08:06 CEST

To: IBM Netrexx <[hidden email]>

Reply-To: IBM Netrexx <[hidden email]>

I am currently packaging the "Java 8 compatible" version, which is the 3.02GA version + the one necessary patch for the new bytecode introduced in 7 but used in 8 that Kermit prepared.

What do we all think the version number should be?

One way is to call it 3.021

Another way is to call it 3.03 and make the next update 3.1 (but that clashes with the mainframe convention to call the "3.0" version 3.1 or 4.0

I expected the 3.03 version to have the scripting enhancements, but they are not really releaseable yet. I think with the proposed enhancements, that version (with the do binary blocks etc) deserves a 4.0 name.

I'll consider all input but I am trying to release, including the new eclipse compiler, today CET, so Tom can have a go at updating the GUI installer.

best regards,

René.


On 14 apr. 2014, at 12:05, Kermit Kiser <[hidden email]> wrote:

All list members are hereby invited to help develop the next version of NetRexx!

_______________________________________________
Ibm-netrexx mailing list
[hidden email]
Online Archive : http://ibm-netrexx.215625.n3.nabble.com/

_______________________________________________
Ibm-netrexx mailing list
[hidden email]
Online Archive : http://ibm-netrexx.215625.n3.nabble.com/

_______________________________________________
Ibm-netrexx mailing list
[hidden email]
Online Archive : http://ibm-netrexx.215625.n3.nabble.com/