ibm-netrexx

Reading ZIP/JAR Files in Java &/ NetRexx

Classic

List

Threaded

18 messages Options

Thomas.Schneider.Wien

Reading ZIP/JAR Files in Java &/ NetRexx

Hi there again,
first, I must admit that I am *NO* Java *NOR* NetRexx Expert.
second, I would like to enhance my FILE-Access routines (current doc
in \Rexx2RT\RexxFile
(in the Run-Time package of www.Rexx2Nrx.com since years) to do the
following:

-- using a special (computer independent) Syntax to read 'the members of:

-- a JAR File
-- a ZIP File
-- an IBM zOS PDS file

ON demand (e.g. to use linein, lineout, and all the other things already
implemented
in RexxFile (without the need to CHANGE the 'calling program')

Also, the '/' may now be used as a GENRIC File-separator (translated to
'\' in the
Windows case...)

Questions:

1.) What do you suggest as the proper syntax for the LIBRARY Name
preceeding
the filename (may it be a JAR, a ZIP, or an IBM z-OS PDS)

2.) Does anyone of you have a (small) example how I can read/write the LINES
of a specified MEMBER of a JAR/ZIP file in Java/NetRexx..

Tom.

PS: I already got from Walter Pachl some (very old & very good) examples
how to read an IBM PDS File thru EXECIO.

I am currently testing my own implementation of EXECIO for Rexx2Nrx.
Thomas Schneider.

Tom. (ths@db-123.com)

David Requena

Reading ZIP/JAR Files in Java &/ NetRexx

Thomas,

ZIP and JAR are the same thing.

There is standard stuff in the java library for manipulating this format.
Look at the package java.util.zip classes.

David

2010/2/12 Thomas Schneider <[hidden email]>

> Hi there again,
> first, I must admit that I am *NO* Java *NOR* NetRexx Expert.
> second, I would like to enhance my FILE-Access routines (current doc in
> \Rexx2RT\RexxFile
> (in the Run-Time package of www.Rexx2Nrx.com since years) to do the
> following:
>
> -- using a special (computer independent) Syntax to read 'the members of:
>
> -- a JAR File
> -- a ZIP File
> -- an IBM zOS PDS file
>
> ON demand (e.g. to use linein, lineout, and all the other things already
> implemented
> in RexxFile (without the need to CHANGE the 'calling program')
>
> Also, the '/' may now be used as a GENRIC File-separator (translated to '\'
> in the
> Windows case...)
>
> Questions:
>
> 1.) What do you suggest as the proper syntax for the LIBRARY Name
> preceeding
> the filename (may it be a JAR, a ZIP, or an IBM z-OS PDS)
>
> 2.) Does anyone of you have a (small) example how I can read/write the
> LINES
> of a specified MEMBER of a JAR/ZIP file in Java/NetRexx..
>
> Tom.
>
> PS: I already got from Walter Pachl some (very old & very good) examples
> how to read an IBM PDS File thru EXECIO.
>
> I am currently testing my own implementation of EXECIO for Rexx2Nrx.
> Thomas Schneider.
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>

--
Saludos / Regards,
David Requena
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ns.hursley.ibm.com/pipermail/ibm-netrexx/attachments/20100213/39757a34/attachment.html

Thomas.Schneider.Wien

Reading ZIP/JAR Files in Java &/ NetRexx

Hi Divid,
.. many thanks for the Details.
... MFC already said me the same thins!

Isn't that (OUR) Universe ...

..... a UNIQUE thing ??
Thomas.
... will read your message (below) tomorrow in more detail.
= Did archive it now, I'm gooing a'sleeeeep ...
=================================================================
David Requena schrieb:

> Thomas,
>
> ZIP and JAR are the same thing.
>
> There is standard stuff in the java library for manipulating this format.
> Look at the package java.util.zip classes.
>
> David
>
> 2010/2/12 Thomas Schneider <[hidden email] <mailto:[hidden email]>>
>
> Hi there again,
> first, I must admit that I am *NO* Java *NOR* NetRexx Expert.
> second, I would like to enhance my FILE-Access routines (current
> doc in \Rexx2RT\RexxFile
> (in the Run-Time package of www.Rexx2Nrx.com
> <http://www.Rexx2Nrx.com> since years) to do the following:
>
> -- using a special (computer independent) Syntax to read 'the
> members of:
>
> -- a JAR File
> -- a ZIP File
> -- an IBM zOS PDS file
>
> ON demand (e.g. to use linein, lineout, and all the other things
> already implemented
> in RexxFile (without the need to CHANGE the 'calling program')
>
> Also, the '/' may now be used as a GENRIC File-separator
> (translated to '\' in the
> Windows case...)
>
> Questions:
>
> 1.) What do you suggest as the proper syntax for the LIBRARY Name
> preceeding
> the filename (may it be a JAR, a ZIP, or an IBM z-OS PDS)
>
> 2.) Does anyone of you have a (small) example how I can read/write
> the LINES
> of a specified MEMBER of a JAR/ZIP file in Java/NetRexx..
>
> Tom.
>
> PS: I already got from Walter Pachl some (very old & very good)
> examples
> how to read an IBM PDS File thru EXECIO.
>
> I am currently testing my own implementation of EXECIO for Rexx2Nrx.
> Thomas Schneider.
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email] <mailto:[hidden email]>
>
>
>
>
> --
> Saludos / Regards,
> David Requena
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>

Tom. (ths@db-123.com)

Thomas.Schneider.Wien

Reading ZIP/JAR Files in Java &/ NetRexx

In reply to this post by David Requena

An HTML attachment was scrubbed...
URL: http://ns.hursley.ibm.com/pipermail/ibm-netrexx/attachments/20100213/5c656603/attachment.html

Tom. (ths@db-123.com)

Kermit Kiser

Reading ZIP/JAR Files in Java &/ NetRexx

In reply to this post by Thomas.Schneider.Wien

Thomas ;

There was an exchange about reading Zip/Jar files on the list last
month. I don't know if you were on the list then, so I am attaching it here.

It sounds like you might be interested in the actual code to read the
data also, so here it is:

------------------------------------------------------ general purpose routines -------------------------------------------------------

method copyjarentry(jf=JarFile,je=JarEntry,scriptdir=String) -- method to copy a jarfile entry to a directory

if df then trace results

nrscript=scriptdir||fs||je.getName

nrs=File(nrscript)

if \nrs.exists then do

rc=copyjarentrytofile(jf,je,nrs)

if rc=1 then say je.getName "copied to:" scriptdir

end

method copyjarentrytofile(jf=JarFile,je=JarEntry,nrs=File)

if df then trace results

do

scriptstream=jf.getInputStream(je)

if Rexx(je.getName).pos(".class")=0 then do

scriptreader=BufferedReader(InputStreamReader(scriptstream))

-- scriptwriter=BufferedWriter(OutputStreamWriter(outstream))

scriptwriter=BufferedWriter(FileWriter(nrs))

copyfile(scriptreader,scriptwriter)

scriptwriter.close

end

else do

outstream=FileOutputStream(nrs)

copybin(scriptstream,outstream)

outstream.close

end

catch badguy=Exception

say je.getName "copy error =" badguy

return 0

end

return 1

/* methods to copy a file: */

method copyfile(ifq=java.io.BufferedReader,ofq=java.io.BufferedWriter) signals IOException

if df then trace results

line=ifq.readline

if line = null then return

ofq.write(string line,0,line.length)

trace off

loop forever

line=ifq.readline

if line = null then leave

ofq.newline()

ofq.write(string line,0,line.length)

end

method copybin(ifq=InputStream,ofq=OutputStream) binary signals IOException

bite=ifq.read

loop while bite \= -1

ofq.write(bite)

bite=ifq.read

end

---------------------------------------------------------------------------------------------------------------------------------------------------

Kinda crude, but maybe it helps.

note: fs=File.separator

-- Kermit

Thomas Schneider wrote:

> Hi Divid,
> .. many thanks for the Details.
> ... MFC already said me the same thins!
>
> Isn't that (OUR) Universe ...
>
> ..... a UNIQUE thing ??
> Thomas.
> ... will read your message (below) tomorrow in more detail.
> = Did archive it now, I'm gooing a'sleeeeep ...
> =================================================================
> David Requena schrieb:
>> Thomas,
>>
>> ZIP and JAR are the same thing.
>>
>> There is standard stuff in the java library for manipulating this
>> format.
>> Look at the package java.util.zip classes.
>>
>> David
>>
>> 2010/2/12 Thomas Schneider <[hidden email] <mailto:[hidden email]>>
>>
>> Hi there again,
>> first, I must admit that I am *NO* Java *NOR* NetRexx Expert.
>> second, I would like to enhance my FILE-Access routines (current
>> doc in \Rexx2RT\RexxFile
>> (in the Run-Time package of www.Rexx2Nrx.com
>> <http://www.Rexx2Nrx.com> since years) to do the following:
>>
>> -- using a special (computer independent) Syntax to read 'the
>> members of:
>>
>> -- a JAR File
>> -- a ZIP File
>> -- an IBM zOS PDS file
>>
>> ON demand (e.g. to use linein, lineout, and all the other things
>> already implemented
>> in RexxFile (without the need to CHANGE the 'calling program')
>>
>> Also, the '/' may now be used as a GENRIC File-separator
>> (translated to '\' in the
>> Windows case...)
>>
>> Questions:
>>
>> 1.) What do you suggest as the proper syntax for the LIBRARY Name
>> preceeding
>> the filename (may it be a JAR, a ZIP, or an IBM z-OS PDS)
>>
>> 2.) Does anyone of you have a (small) example how I can read/write
>> the LINES
>> of a specified MEMBER of a JAR/ZIP file in Java/NetRexx..
>>
>> Tom.
>>
>> PS: I already got from Walter Pachl some (very old & very good)
>> examples
>> how to read an IBM PDS File thru EXECIO.
>>
>> I am currently testing my own implementation of EXECIO for Rexx2Nrx.
>> Thomas Schneider.
>>
>> _______________________________________________
>> Ibm-netrexx mailing list
>> [hidden email] <mailto:[hidden email]>
>>
>>
>>
>>
>> --
>> Saludos / Regards,
>> David Requena
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> Ibm-netrexx mailing list
>> [hidden email]
>>
>>
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>
>

-------------- next part --------------
An embedded message was scrubbed...
From: Kermit Kiser <[hidden email]>
Subject: Re: [Ibm-netrexx] Help Needed -
Date: Wed, 13 Jan 2010 17:19:33 -0800
Size: 8746
Url: http://ns.hursley.ibm.com/pipermail/ibm-netrexx/attachments/20100212/b142b609/Ibm-netrexxHelpNeeded--0001.eml

David Requena

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx

Kermit,

hmm.. got curious. I noticed the following on your supplied code snipped:

if df then trace results

Here I'm supposing df is a property appropriately set up upstream as needed.

This is a construct I've used often in other languages but in NetRexx I
tend to use trace statements liberally as needed. Then for production
build I use 'options notrace' so no overhead is incurred.

Any reason why would you prefer the conditional check over NetRexx'
options nobinary idiom?

Maybe I'm being dumb but seems to me the conditional would be carried to
production code whereas options nobinary would eliminate any trace
(yeah, redundant) of trace code.

---
Saludos / Kind regards.
David Requena

El 13/02/2010 8:16, Kermit Kiser escribi?:

> Thomas ;
>
> There was an exchange about reading Zip/Jar files on the list last
> month. I don't know if you were on the list then, so I am attaching it
> here.
>
> It sounds like you might be interested in the actual code to read the
> data also, so here it is:
>
> ------------------------------------------------------ general purpose
> routines -------------------------------------------------------
>
> method copyjarentry(jf=JarFile,je=JarEntry,scriptdir=String) --
> method to copy a jarfile entry to a directory
>
> if df then trace results
> nrscript=scriptdir||fs||je.getName
>
> nrs=File(nrscript)
>
> if \nrs.exists then do
>
> rc=copyjarentrytofile(jf,je,nrs)
>
> if rc=1 then say je.getName "copied to:" scriptdir
>
> end
>
>
> method copyjarentrytofile(jf=JarFile,je=JarEntry,nrs=File)
>
> if df then trace results
>
> do
>
> scriptstream=jf.getInputStream(je)
>
> if Rexx(je.getName).pos(".class")=0 then do
>
>
> scriptreader=BufferedReader(InputStreamReader(scriptstream))
>
> --
> scriptwriter=BufferedWriter(OutputStreamWriter(outstream))
>
> scriptwriter=BufferedWriter(FileWriter(nrs))
>
> copyfile(scriptreader,scriptwriter)
>
> scriptwriter.close
>
> end
>
> else do
>
> outstream=FileOutputStream(nrs)
>
> copybin(scriptstream,outstream)
>
> outstream.close
>
> end
>
> catch badguy=Exception
>
> say je.getName "copy error =" badguy
>
> return 0
>
> end
>
> return 1
>
>
> /* methods to copy a file: */
>
> method copyfile(ifq=java.io.BufferedReader,ofq=java.io.BufferedWriter)
> signals IOException
>
> if df then trace results
>
> line=ifq.readline
>
> if line = null then return
>
> ofq.write(string line,0,line.length)
>
> trace off
>
> loop forever
>
> line=ifq.readline
>
> if line = null then leave
>
> ofq.newline()
>
> ofq.write(string line,0,line.length)
>
> end
> method copybin(ifq=InputStream,ofq=OutputStream) binary signals
> IOException
>
> bite=ifq.read
>
> loop while bite \= -1
>
> ofq.write(bite)
>
> bite=ifq.read
>
> end
> ---------------------------------------------------------------------------------------------------------------------------------------------------
>
>
>
> Kinda crude, but maybe it helps.
>
> note: fs=File.separator
>
> -- Kermit
>
>
> Thomas Schneider wrote:
>> Hi Divid,
>> .. many thanks for the Details.
>> ... MFC already said me the same thins!
>>
>> Isn't that (OUR) Universe ...
>>
>> ..... a UNIQUE thing ??
>> Thomas.
>> ... will read your message (below) tomorrow in more detail.
>> = Did archive it now, I'm gooing a'sleeeeep ...
>> =================================================================
>> David Requena schrieb:
>>> Thomas,
>>>
>>> ZIP and JAR are the same thing.
>>>
>>> There is standard stuff in the java library for manipulating this
>>> format.
>>> Look at the package java.util.zip classes.
>>>
>>> David
>>>
>>> 2010/2/12 Thomas Schneider <[hidden email] <mailto:[hidden email]>>
>>>
>>> Hi there again,
>>> first, I must admit that I am *NO* Java *NOR* NetRexx Expert.
>>> second, I would like to enhance my FILE-Access routines (current
>>> doc in \Rexx2RT\RexxFile
>>> (in the Run-Time package of www.Rexx2Nrx.com
>>> <http://www.Rexx2Nrx.com> since years) to do the following:
>>>
>>> -- using a special (computer independent) Syntax to read 'the
>>> members of:
>>>
>>> -- a JAR File
>>> -- a ZIP File
>>> -- an IBM zOS PDS file
>>>
>>> ON demand (e.g. to use linein, lineout, and all the other things
>>> already implemented
>>> in RexxFile (without the need to CHANGE the 'calling program')
>>>
>>> Also, the '/' may now be used as a GENRIC File-separator
>>> (translated to '\' in the
>>> Windows case...)
>>>
>>> Questions:
>>>
>>> 1.) What do you suggest as the proper syntax for the LIBRARY Name
>>> preceeding
>>> the filename (may it be a JAR, a ZIP, or an IBM z-OS PDS)
>>>
>>> 2.) Does anyone of you have a (small) example how I can read/write
>>> the LINES
>>> of a specified MEMBER of a JAR/ZIP file in Java/NetRexx..
>>>
>>> Tom.
>>>
>>> PS: I already got from Walter Pachl some (very old & very good)
>>> examples
>>> how to read an IBM PDS File thru EXECIO.
>>>
>>> I am currently testing my own implementation of EXECIO for
>>> Rexx2Nrx.
>>> Thomas Schneider.
>>>
>>> _______________________________________________
>>> Ibm-netrexx mailing list
>>> [hidden email] <mailto:[hidden email]>
>>>
>>>
>>>
>>>
>>> --
>>> Saludos / Regards,
>>> David Requena
>>>
>>> ------------------------------------------------------------------------
>>>
>>>
>>> _______________________________________________
>>> Ibm-netrexx mailing list
>>> [hidden email]
>>>
>>
>> _______________________________________________
>> Ibm-netrexx mailing list
>> [hidden email]
>>
>>
>>
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ns.hursley.ibm.com/pipermail/ibm-netrexx/attachments/20100215/b077b83f/attachment-0001.html

Thomas.Schneider.Wien

SPAM Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4)

Hi David,
I'm using a similar approach than Kermit, to have an OPTION to
*Dynamically* force DEBUG.
I'm rarely using TRACE, but I do have a lot of conditional debugs in my
program, which I can even
use in the PRODUCTION version just in case a not anticipated situation
occurs in the program
of a client.

This is just for your info.

Similarly, I do use a lot of assertions in my code, and did invent
some special notation
for those in my newer programs, e.g.:

x=3
y=x ::= 3

where :: introduces an assertion, which is checked at run tyme (in this
case, y MUST be 3)

when it is NOT 3, an ASSERTION error message is given (and when not
otherwise wanted,
the program aborts immediately.

Tom.
========================================================
Kermit, please give us the reson of cyour decision as well.

David Requena schrieb:

> Kermit,
>
> hmm.. got curious. I noticed the following on your supplied code snipped:
>
> if df then trace results
>
> Here I'm supposing df is a property appropriately set up upstream as
> needed.
>
> This is a construct I've used often in other languages but in NetRexx
> I tend to use trace statements liberally as needed. Then for
> production build I use 'options notrace' so no overhead is incurred.
>
> Any reason why would you prefer the conditional check over NetRexx'
> options nobinary idiom?
>
> Maybe I'm being dumb but seems to me the conditional would be carried
> to production code whereas options nobinary would eliminate any trace
> (yeah, redundant) of trace code.
>
> ---
> Saludos / Kind regards.
> David Requena
>
> El 13/02/2010 8:16, Kermit Kiser escribi?:
>> Thomas ;
>>
>> There was an exchange about reading Zip/Jar files on the list last
>> month. I don't know if you were on the list then, so I am attaching
>> it here.
>>
>> It sounds like you might be interested in the actual code to read the
>> data also, so here it is:
>>
>> ------------------------------------------------------ general
>> purpose routines -------------------------------------------------------
>>
>> method copyjarentry(jf=JarFile,je=JarEntry,scriptdir=String) --
>> method to copy a jarfile entry to a directory
>>
>> if df then trace results
>> nrscript=scriptdir||fs||je.getName
>>
>> nrs=File(nrscript)
>>
>> if \nrs.exists then do
>>
>> rc=copyjarentrytofile(jf,je,nrs)
>>
>> if rc=1 then say je.getName "copied to:" scriptdir
>>
>> end
>>
>>
>> method copyjarentrytofile(jf=JarFile,je=JarEntry,nrs=File)
>>
>> if df then trace results
>>
>> do
>>
>> scriptstream=jf.getInputStream(je)
>>
>> if Rexx(je.getName).pos(".class")=0 then do
>>
>>
>> scriptreader=BufferedReader(InputStreamReader(scriptstream))
>>
>> --
>> scriptwriter=BufferedWriter(OutputStreamWriter(outstream))
>>
>> scriptwriter=BufferedWriter(FileWriter(nrs))
>>
>> copyfile(scriptreader,scriptwriter)
>>
>> scriptwriter.close
>>
>> end
>>
>> else do
>>
>> outstream=FileOutputStream(nrs)
>>
>> copybin(scriptstream,outstream)
>>
>> outstream.close
>>
>> end
>>
>> catch badguy=Exception
>>
>> say je.getName "copy error =" badguy
>>
>> return 0
>>
>> end
>>
>> return 1
>>
>>
>> /* methods to copy a file: */
>>
>> method
>> copyfile(ifq=java.io.BufferedReader,ofq=java.io.BufferedWriter)
>> signals IOException
>>
>> if df then trace results
>>
>> line=ifq.readline
>>
>> if line = null then return
>>
>> ofq.write(string line,0,line.length)
>>
>> trace off
>>
>> loop forever
>>
>> line=ifq.readline
>>
>> if line = null then leave
>>
>> ofq.newline()
>>
>> ofq.write(string line,0,line.length)
>>
>> end
>> method copybin(ifq=InputStream,ofq=OutputStream) binary signals
>> IOException
>>
>> bite=ifq.read
>>
>> loop while bite \= -1
>>
>> ofq.write(bite)
>>
>> bite=ifq.read
>>
>> end
>> ---------------------------------------------------------------------------------------------------------------------------------------------------
>>
>>
>>
>> Kinda crude, but maybe it helps.
>>
>> note: fs=File.separator
>>
>> -- Kermit
>>
>>
>> Thomas Schneider wrote:
>>> Hi Divid,
>>> .. many thanks for the Details.
>>> ... MFC already said me the same thins!
>>>
>>> Isn't that (OUR) Universe ...
>>>
>>> ..... a UNIQUE thing ??
>>> Thomas.
>>> ... will read your message (below) tomorrow in more detail.
>>> = Did archive it now, I'm gooing a'sleeeeep ...
>>> =================================================================
>>> David Requena schrieb:
>>>> Thomas,
>>>>
>>>> ZIP and JAR are the same thing.
>>>>
>>>> There is standard stuff in the java library for manipulating this
>>>> format.
>>>> Look at the package java.util.zip classes.
>>>>
>>>> David
>>>>
>>>> 2010/2/12 Thomas Schneider <[hidden email] <mailto:[hidden email]>>
>>>>
>>>> Hi there again,
>>>> first, I must admit that I am *NO* Java *NOR* NetRexx Expert.
>>>> second, I would like to enhance my FILE-Access routines (current
>>>> doc in \Rexx2RT\RexxFile
>>>> (in the Run-Time package of www.Rexx2Nrx.com
>>>> <http://www.Rexx2Nrx.com> since years) to do the following:
>>>>
>>>> -- using a special (computer independent) Syntax to read 'the
>>>> members of:
>>>>
>>>> -- a JAR File
>>>> -- a ZIP File
>>>> -- an IBM zOS PDS file
>>>>
>>>> ON demand (e.g. to use linein, lineout, and all the other things
>>>> already implemented
>>>> in RexxFile (without the need to CHANGE the 'calling program')
>>>>
>>>> Also, the '/' may now be used as a GENRIC File-separator
>>>> (translated to '\' in the
>>>> Windows case...)
>>>>
>>>> Questions:
>>>>
>>>> 1.) What do you suggest as the proper syntax for the LIBRARY Name
>>>> preceeding
>>>> the filename (may it be a JAR, a ZIP, or an IBM z-OS PDS)
>>>>
>>>> 2.) Does anyone of you have a (small) example how I can read/write
>>>> the LINES
>>>> of a specified MEMBER of a JAR/ZIP file in Java/NetRexx..
>>>>
>>>> Tom.
>>>>
>>>> PS: I already got from Walter Pachl some (very old & very good)
>>>> examples
>>>> how to read an IBM PDS File thru EXECIO.
>>>>
>>>> I am currently testing my own implementation of EXECIO for
>>>> Rexx2Nrx.
>>>> Thomas Schneider.
>>>>
>>>> _______________________________________________
>>>> Ibm-netrexx mailing list
>>>> [hidden email] <mailto:[hidden email]>
>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Saludos / Regards,
>>>> David Requena
>>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>>
>>>> _______________________________________________
>>>> Ibm-netrexx mailing list
>>>> [hidden email]
>>>>
>>>>
>>>
>>> _______________________________________________
>>> Ibm-netrexx mailing list
>>> [hidden email]
>>>
>>>
>>>
>>
>> _______________________________________________
>> Ibm-netrexx mailing list
>> [hidden email]
>>
>>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>

Tom. (ths@db-123.com)

David Requena

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4)

Thomas,

That is indeed a very valid point.
Thought of the possibility of activating debug mode on production code
by using some command line switch just after sending my previous email.

Thanks,
David

---
Saludos / Kind regards.
David Requena

El 15/02/2010 13:44, Thomas Schneider escribi?:

> Hi David,
> I'm using a similar approach than Kermit, to have an OPTION to
> *Dynamically* force DEBUG.
> I'm rarely using TRACE, but I do have a lot of conditional debugs in
> my program, which I can even
> use in the PRODUCTION version just in case a not anticipated situation
> occurs in the program
> of a client.
>
> This is just for your info.
>
> Similarly, I do use a lot of assertions in my code, and did invent
> some special notation
> for those in my newer programs, e.g.:
>
> x=3
> y=x ::= 3
>
> where :: introduces an assertion, which is checked at run tyme (in
> this case, y MUST be 3)
>
> when it is NOT 3, an ASSERTION error message is given (and when not
> otherwise wanted,
> the program aborts immediately.
>
> Tom.
> ========================================================
> Kermit, please give us the reson of cyour decision as well.
>
> David Requena schrieb:
>> Kermit,
>>
>> hmm.. got curious. I noticed the following on your supplied code
>> snipped:
>>
>> if df then trace results
>>
>> Here I'm supposing df is a property appropriately set up upstream as
>> needed.
>>
>> This is a construct I've used often in other languages but in NetRexx
>> I tend to use trace statements liberally as needed. Then for
>> production build I use 'options notrace' so no overhead is incurred.
>>
>> Any reason why would you prefer the conditional check over NetRexx'
>> options nobinary idiom?
>>
>> Maybe I'm being dumb but seems to me the conditional would be carried
>> to production code whereas options nobinary would eliminate any trace
>> (yeah, redundant) of trace code.
>>
>> ---
>> Saludos / Kind regards.
>> David Requena
>>
>> El 13/02/2010 8:16, Kermit Kiser escribi?:
>>> Thomas ;
>>>
>>> There was an exchange about reading Zip/Jar files on the list last
>>> month. I don't know if you were on the list then, so I am attaching
>>> it here.
>>>
>>> It sounds like you might be interested in the actual code to read
>>> the data also, so here it is:
>>>
>>> ------------------------------------------------------ general
>>> purpose routines
>>> -------------------------------------------------------
>>>
>>> method copyjarentry(jf=JarFile,je=JarEntry,scriptdir=String)
>>> -- method to copy a jarfile entry to a directory
>>>
>>> if df then trace results nrscript=scriptdir||fs||je.getName
>>>
>>> nrs=File(nrscript)
>>>
>>> if \nrs.exists then do
>>>
>>> rc=copyjarentrytofile(jf,je,nrs)
>>>
>>> if rc=1 then say je.getName "copied to:" scriptdir
>>>
>>> end
>>>
>>> method copyjarentrytofile(jf=JarFile,je=JarEntry,nrs=File)
>>>
>>> if df then trace results
>>>
>>> do
>>>
>>> scriptstream=jf.getInputStream(je)
>>>
>>> if Rexx(je.getName).pos(".class")=0 then do
>>>
>>>
>>> scriptreader=BufferedReader(InputStreamReader(scriptstream))
>>>
>>> --
>>> scriptwriter=BufferedWriter(OutputStreamWriter(outstream))
>>>
>>> scriptwriter=BufferedWriter(FileWriter(nrs))
>>>
>>> copyfile(scriptreader,scriptwriter)
>>>
>>> scriptwriter.close
>>>
>>> end
>>>
>>> else do
>>>
>>> outstream=FileOutputStream(nrs)
>>>
>>> copybin(scriptstream,outstream)
>>>
>>> outstream.close
>>>
>>> end
>>>
>>> catch badguy=Exception
>>>
>>> say je.getName "copy error =" badguy
>>>
>>> return 0
>>>
>>> end
>>>
>>> return 1
>>>
>>> /* methods to copy a file: */
>>>
>>> method
>>> copyfile(ifq=java.io.BufferedReader,ofq=java.io.BufferedWriter)
>>> signals IOException
>>>
>>> if df then trace results
>>>
>>> line=ifq.readline
>>>
>>> if line = null then return
>>>
>>> ofq.write(string line,0,line.length)
>>>
>>> trace off
>>>
>>> loop forever
>>>
>>> line=ifq.readline
>>>
>>> if line = null then leave
>>>
>>> ofq.newline()
>>>
>>> ofq.write(string line,0,line.length)
>>>
>>> end method copybin(ifq=InputStream,ofq=OutputStream) binary
>>> signals IOException
>>>
>>> bite=ifq.read
>>>
>>> loop while bite \= -1
>>>
>>> ofq.write(bite)
>>>
>>> bite=ifq.read
>>>
>>> end
>>> ---------------------------------------------------------------------------------------------------------------------------------------------------
>>>
>>>
>>>
>>> Kinda crude, but maybe it helps.
>>>
>>> note: fs=File.separator
>>>
>>> -- Kermit
>>>
>>>
>>> Thomas Schneider wrote:
>>>> Hi Divid,
>>>> .. many thanks for the Details.
>>>> ... MFC already said me the same thins!
>>>>
>>>> Isn't that (OUR) Universe ...
>>>>
>>>> ..... a UNIQUE thing ??
>>>> Thomas.
>>>> ... will read your message (below) tomorrow in more detail.
>>>> = Did archive it now, I'm gooing a'sleeeeep ...
>>>> =================================================================
>>>> David Requena schrieb:
>>>>> Thomas,
>>>>>
>>>>> ZIP and JAR are the same thing.
>>>>>
>>>>> There is standard stuff in the java library for manipulating this
>>>>> format.
>>>>> Look at the package java.util.zip classes.
>>>>>
>>>>> David
>>>>>
>>>>> 2010/2/12 Thomas Schneider <[hidden email] <mailto:[hidden email]>>
>>>>>
>>>>> Hi there again,
>>>>> first, I must admit that I am *NO* Java *NOR* NetRexx Expert.
>>>>> second, I would like to enhance my FILE-Access routines (current
>>>>> doc in \Rexx2RT\RexxFile
>>>>> (in the Run-Time package of www.Rexx2Nrx.com
>>>>> <http://www.Rexx2Nrx.com> since years) to do the following:
>>>>>
>>>>> -- using a special (computer independent) Syntax to read 'the
>>>>> members of:
>>>>>
>>>>> -- a JAR File
>>>>> -- a ZIP File
>>>>> -- an IBM zOS PDS file
>>>>>
>>>>> ON demand (e.g. to use linein, lineout, and all the other things
>>>>> already implemented
>>>>> in RexxFile (without the need to CHANGE the 'calling program')
>>>>>
>>>>> Also, the '/' may now be used as a GENRIC File-separator
>>>>> (translated to '\' in the
>>>>> Windows case...)
>>>>>
>>>>> Questions:
>>>>>
>>>>> 1.) What do you suggest as the proper syntax for the LIBRARY Name
>>>>> preceeding
>>>>> the filename (may it be a JAR, a ZIP, or an IBM z-OS PDS)
>>>>>
>>>>> 2.) Does anyone of you have a (small) example how I can
>>>>> read/write
>>>>> the LINES
>>>>> of a specified MEMBER of a JAR/ZIP file in Java/NetRexx..
>>>>>
>>>>> Tom.
>>>>>
>>>>> PS: I already got from Walter Pachl some (very old & very good)
>>>>> examples
>>>>> how to read an IBM PDS File thru EXECIO.
>>>>>
>>>>> I am currently testing my own implementation of EXECIO for
>>>>> Rexx2Nrx.
>>>>> Thomas Schneider.
>>>>>
>>>>> _______________________________________________
>>>>> Ibm-netrexx mailing list
>>>>> [hidden email] <mailto:[hidden email]>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Saludos / Regards,
>>>>> David Requena
>>>>>
>>>>> ------------------------------------------------------------------------
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Ibm-netrexx mailing list
>>>>> [hidden email]
>>>>>
>>>>
>>>> _______________________________________________
>>>> Ibm-netrexx mailing list
>>>> [hidden email]
>>>>
>>>>
>>>>
>>>
>>> _______________________________________________
>>> Ibm-netrexx mailing list
>>> [hidden email]
>>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> Ibm-netrexx mailing list
>> [hidden email]
>>
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>

Thomas.Schneider.Wien

Rey OPTIONS.

Hello David,
I am (first time) trying to attach a PIECE of my source-code, namely
the Options.nrx file, for two purposes:

1.) To inform you of the Options available in Rey.
2.) To *test* (for my own knowledge) whether the IBM NetRexx Groups does
allow to
send ATTACHED files (which REXXLA does NOT!).

Tom.

David Requena schrieb:

> Thomas,
>
> That is indeed a very valid point.
> Thought of the possibility of activating debug mode on production code
> by using some command line switch just after sending my previous email.
>
> Thanks,
> David
>
> ---
> Saludos / Kind regards.
> David Requena
>
>
> El 15/02/2010 13:44, Thomas Schneider escribi?:
>> Hi David,
>> I'm using a similar approach than Kermit, to have an OPTION to
>> *Dynamically* force DEBUG.
>> I'm rarely using TRACE, but I do have a lot of conditional debugs in
>> my program, which I can even
>> use in the PRODUCTION version just in case a not anticipated
>> situation occurs in the program
>> of a client.
>>
>> This is just for your info.
>>
>> Similarly, I do use a lot of assertions in my code, and did invent
>> some special notation
>> for those in my newer programs, e.g.:
>>
>> x=3
>> y=x ::= 3
>>
>> where :: introduces an assertion, which is checked at run tyme (in
>> this case, y MUST be 3)
>>
>> when it is NOT 3, an ASSERTION error message is given (and when not
>> otherwise wanted,
>> the program aborts immediately.
>>
>> Tom.
>> ========================================================
>> Kermit, please give us the reson of cyour decision as well.
>>
>> David Requena schrieb:
>>> Kermit,
>>>
>>> hmm.. got curious. I noticed the following on your supplied code
>>> snipped:
>>>
>>> if df then trace results
>>>
>>> Here I'm supposing df is a property appropriately set up upstream as
>>> needed.
>>>
>>> This is a construct I've used often in other languages but in
>>> NetRexx I tend to use trace statements liberally as needed. Then for
>>> production build I use 'options notrace' so no overhead is incurred.
>>>
>>> Any reason why would you prefer the conditional check over NetRexx'
>>> options nobinary idiom?
>>>
>>> Maybe I'm being dumb but seems to me the conditional would be
>>> carried to production code whereas options nobinary would eliminate
>>> any trace (yeah, redundant) of trace code.
>>>
>>> ---
>>> Saludos / Kind regards.
>>> David Requena
>>>
>>> El 13/02/2010 8:16, Kermit Kiser escribi?:
>>>> Thomas ;
>>>>
>>>> There was an exchange about reading Zip/Jar files on the list last
>>>> month. I don't know if you were on the list then, so I am attaching
>>>> it here.
>>>>
>>>> It sounds like you might be interested in the actual code to read
>>>> the data also, so here it is:
>>>>
>>>> ------------------------------------------------------ general
>>>> purpose routines
>>>> -------------------------------------------------------
>>>>
>>>> method copyjarentry(jf=JarFile,je=JarEntry,scriptdir=String)
>>>> -- method to copy a jarfile entry to a directory
>>>>
>>>> if df then trace results nrscript=scriptdir||fs||je.getName
>>>>
>>>> nrs=File(nrscript)
>>>>
>>>> if \nrs.exists then do
>>>>
>>>> rc=copyjarentrytofile(jf,je,nrs)
>>>>
>>>> if rc=1 then say je.getName "copied to:" scriptdir
>>>>
>>>> end
>>>>
>>>> method copyjarentrytofile(jf=JarFile,je=JarEntry,nrs=File)
>>>>
>>>> if df then trace results
>>>>
>>>> do
>>>>
>>>> scriptstream=jf.getInputStream(je)
>>>>
>>>> if Rexx(je.getName).pos(".class")=0 then do
>>>>
>>>>
>>>> scriptreader=BufferedReader(InputStreamReader(scriptstream))
>>>>
>>>> --
>>>> scriptwriter=BufferedWriter(OutputStreamWriter(outstream))
>>>>
>>>> scriptwriter=BufferedWriter(FileWriter(nrs))
>>>>
>>>> copyfile(scriptreader,scriptwriter)
>>>>
>>>> scriptwriter.close
>>>>
>>>> end
>>>>
>>>> else do
>>>>
>>>> outstream=FileOutputStream(nrs)
>>>>
>>>> copybin(scriptstream,outstream)
>>>>
>>>> outstream.close
>>>>
>>>> end
>>>>
>>>> catch badguy=Exception
>>>>
>>>> say je.getName "copy error =" badguy
>>>>
>>>> return 0
>>>>
>>>> end
>>>>
>>>> return 1
>>>>
>>>> /* methods to copy a file: */
>>>>
>>>> method
>>>> copyfile(ifq=java.io.BufferedReader,ofq=java.io.BufferedWriter)
>>>> signals IOException
>>>>
>>>> if df then trace results
>>>>
>>>> line=ifq.readline
>>>>
>>>> if line = null then return
>>>>
>>>> ofq.write(string line,0,line.length)
>>>>
>>>> trace off
>>>>
>>>> loop forever
>>>>
>>>> line=ifq.readline
>>>>
>>>> if line = null then leave
>>>>
>>>> ofq.newline()
>>>>
>>>> ofq.write(string line,0,line.length)
>>>>
>>>> end method copybin(ifq=InputStream,ofq=OutputStream)
>>>> binary signals IOException
>>>>
>>>> bite=ifq.read
>>>>
>>>> loop while bite \= -1
>>>>
>>>> ofq.write(bite)
>>>>
>>>> bite=ifq.read
>>>>
>>>> end
>>>> ---------------------------------------------------------------------------------------------------------------------------------------------------
>>>>
>>>>
>>>>
>>>> Kinda crude, but maybe it helps.
>>>>
>>>> note: fs=File.separator
>>>>
>>>> -- Kermit
>>>>
>>>>
>>>> Thomas Schneider wrote:
>>>>> Hi Divid,
>>>>> .. many thanks for the Details.
>>>>> ... MFC already said me the same thins!
>>>>>
>>>>> Isn't that (OUR) Universe ...
>>>>>
>>>>> ..... a UNIQUE thing ??
>>>>> Thomas.
>>>>> ... will read your message (below) tomorrow in more detail.
>>>>> = Did archive it now, I'm gooing a'sleeeeep ...
>>>>> =================================================================
>>>>> David Requena schrieb:
>>>>>> Thomas,
>>>>>>
>>>>>> ZIP and JAR are the same thing.
>>>>>>
>>>>>> There is standard stuff in the java library for manipulating this
>>>>>> format.
>>>>>> Look at the package java.util.zip classes.
>>>>>>
>>>>>> David
>>>>>>
>>>>>> 2010/2/12 Thomas Schneider <[hidden email] <mailto:[hidden email]>>
>>>>>>
>>>>>> Hi there again,
>>>>>> first, I must admit that I am *NO* Java *NOR* NetRexx Expert.
>>>>>> second, I would like to enhance my FILE-Access routines
>>>>>> (current
>>>>>> doc in \Rexx2RT\RexxFile
>>>>>> (in the Run-Time package of www.Rexx2Nrx.com
>>>>>> <http://www.Rexx2Nrx.com> since years) to do the following:
>>>>>>
>>>>>> -- using a special (computer independent) Syntax to read 'the
>>>>>> members of:
>>>>>>
>>>>>> -- a JAR File
>>>>>> -- a ZIP File
>>>>>> -- an IBM zOS PDS file
>>>>>>
>>>>>> ON demand (e.g. to use linein, lineout, and all the other things
>>>>>> already implemented
>>>>>> in RexxFile (without the need to CHANGE the 'calling program')
>>>>>>
>>>>>> Also, the '/' may now be used as a GENRIC File-separator
>>>>>> (translated to '\' in the
>>>>>> Windows case...)
>>>>>>
>>>>>> Questions:
>>>>>>
>>>>>> 1.) What do you suggest as the proper syntax for the LIBRARY
>>>>>> Name
>>>>>> preceeding
>>>>>> the filename (may it be a JAR, a ZIP, or an IBM z-OS PDS)
>>>>>>
>>>>>> 2.) Does anyone of you have a (small) example how I can
>>>>>> read/write
>>>>>> the LINES
>>>>>> of a specified MEMBER of a JAR/ZIP file in Java/NetRexx..
>>>>>>
>>>>>> Tom.
>>>>>>
>>>>>> PS: I already got from Walter Pachl some (very old & very good)
>>>>>> examples
>>>>>> how to read an IBM PDS File thru EXECIO.
>>>>>>
>>>>>> I am currently testing my own implementation of EXECIO for
>>>>>> Rexx2Nrx.
>>>>>> Thomas Schneider.
>>>>>>
>>>>>> _______________________________________________
>>>>>> Ibm-netrexx mailing list
>>>>>> [hidden email] <mailto:[hidden email]>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Saludos / Regards,
>>>>>> David Requena
>>>>>>
>>>>>> ------------------------------------------------------------------------
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Ibm-netrexx mailing list
>>>>>> [hidden email]
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Ibm-netrexx mailing list
>>>>> [hidden email]
>>>>>
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> Ibm-netrexx mailing list
>>>> [hidden email]
>>>>
>>> ------------------------------------------------------------------------
>>>
>>>
>>> _______________________________________________
>>> Ibm-netrexx mailing list
>>> [hidden email]
>>>
>>
>> _______________________________________________
>> Ibm-netrexx mailing list
>> [hidden email]
>>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>

-------------- next part --------------
/**********************************************************************/
/* Options: generalized PP Options module */
/* (c) Th. Schneider, May 2002 */
/**********************************************************************/
/* 16.11.2003: options now moved to RUNTIME-package, as the various */
/* : options are used for Instance by External etc */
/**********************************************************************/
/* 07.10.2002: option 'PRO=pp_project' implemented:needed for Global */
/* 13.10.2002: opt_load and opt_use now 1 by default */
/* : trace_args added (for COB2RWPP) */
/* : NOPAUSE is the default now */
/* 17.05.2003: trace_items added */
/* 24.07.2003: opt_source=1 by default agaIn (TEST Rexx2Nrx 5.00) */
/* 01.09.2003: opt_GERMAN implemented, opt_CMD=1 by default */
/* 15.09.2003: opt_MAP implemented (display MAPS at end of parse) */
/* 02.10.2003: log_by_module (option LBM) implemented */
/* : default log-file is still PP.log */
/* 17.10.2003: option RESERVED implemented (reserved words as ID's */
/* : allowed In REXX by default */
/* 20.10.2003: option trace_use (TU) implemented (In External.nrx) */
/* 03.11.2003: opt_quiet implemented (corrected) */
/* 04.11.2003: opt_ANSII for ANSII standard compliance implemented */
/* 09.11.2003: opt_SHELL implemented */
/* 13.12.2003: note that the _in_Option routine is included In Params*/
/* 01.01.2004: set_pp_options moved from Params to Options, as */
/* : NetRexx uses clause does not work correctly */
/* : option 'TRV' (trace Resolve) added */
/* : option 'TBN' (trace builtin) added */
/* 18.02.2004: pp_project added */
/* 27.06.2004: build 'XREF' by default */
/* : new option FILEDIR */
/* : new option TOPCODES (trace_opcodes) */
/* : new option METRICS (display program metrics) */
/* : new options SCAN, PARSE, ANALYSE, GENERATE, EXEC */
/**********************************************************************/
options binary

package Rey.RT

import Rey.RT.ReyMsg
import Rey.RT.External

class Options public uses ReyMsg, External

/*trace methods*/
properties public static
options_initialized=boolean 0

/* ... Declare Global Switches */
/* 09.09.2008: new options LOWER and UPPER for target name conversions */
opt_lowercase=boolean 1 /* default is now Lowercase names */
opt_uppercase=boolean 0 /* and NO upper case conversion */
opt_scan = boolean 0 /* 1 = SCAN only */
opt_parse = boolean 0 /* 1 = PARSE only */
opt_analyse = boolean 0 /* 1 = ANALYSE only */
opt_generate = boolean 0 /* 1 = GENERATE ONLY */
opt_metrics = boolean 0 /* 1 = Display program Metrics (after SCAN & PARSE) */
opt_exec = boolean 0 /* 1 = EXEC only */
statement_debug=boolean 0
trace_builtin= boolean 0 /* 1= trace builtin Functions as well */

opt_filedir = boolean 0 /* 1 = change components of name to directories */
opt_file = Rexx '' /* may be 'DIR' or 'FILE' */
opt_debug = 0 /* 1 = force OPTION DEBUG */
opt_quiet=0 /* 1 quiet messages to LOG file only ... */
opt_banner = boolean 1 /* copy banner from source to target */
opt_xref=boolean 1 /* build xref by default */
class_by_module = boolean 0
log_by_module = boolean 0
dates_by_module = boolean 0
doit_by_module = boolean 0
opt_CORE = boolean 0
opt_SHELL = boolean 0
opt_ANSII = boolean 0

opt_reserved = boolean 0
opt_MAP = boolean 0
money_by_module = boolean 0
build_uref = boolean 0
opt_comment = boolean 0
opt_GERMAN = boolean 0
opt_MFC = boolean 0
opt_save = boolean 0
opt_note = boolean 1
opt_RTP = boolean 1
opt_BIF = boolean 1
opt_USE = boolean 1
opt_CMD = boolean 1
opt_load = boolean 1
opt_auto_add = boolean 1
opt_localize = boolean 0
opt_explicit = boolean 1
parsing_replacements = boolean 0
change_attributes = boolean 1
opt_package = boolean 1
opt_build = boolean 0
opt_list = boolean 0
opt_long_strings = boolean 0
opt_long_note = boolean 0
opt_short = boolean 0
opt_long = boolean 0
opt_type = boolean 1
opt_attention = boolean 1
opt_Stack = boolean 0
only_rexx = boolean 0
opt_symbolic = boolean 0
opt_source = boolean 1
opt_format = boolean 1
opt_binary = boolean 0
opt_logo = boolean 1
opt_time = boolean 0 -- /* 1 = display detailed timings */
opt_strictassign = boolean 0
opt_strictcase = boolean 0
opt_stats = boolean 0
opt_all_money = boolean 0
warn_missing = boolean 1
opt_test = boolean 0
opt_check = boolean 1
opt_RW = boolean 0
opt_nest = boolean 0
save_doc = boolean 0

trace_opcodes = boolean 0
trace_expose = boolean 0
trace_resolve = boolean 0 /* release 5.00 */
trace_use = boolean 0

trace_items = boolean 0
trace_args = boolean 0
trace_labels = boolean 0
trace_recursion = boolean 0
trace_code = boolean 0
trace_analyser = boolean 0
trace_in = boolean 0
trace_include = boolean 0
trace_tokens = boolean 0
trace_out = boolean 0
trace_value = boolean 0
trace_defs = boolean 0
trace_class = boolean 0
trace_assign = boolean 0
trace_funcs = boolean 0
trace_call = boolean 0
trace_nest = boolean 0
trace_templ = boolean 0
trace_hash = boolean 0
debug_defs = boolean 0
debug_scan = boolean 0
trace_copy = boolean 0
trace_parse = boolean 0
trace_line = boolean 0
trace_verbs = boolean 0
trace_decl = boolean 0
trace_date = boolean 0
debug_date = boolean 0
trace_source = boolean 0
trace_cond = boolean 0
trace_expr = boolean 0
trace_item = boolean 0
trace_checks = boolean 0
trace_flags = boolean 0
trace_perform = boolean 0
trace_exec = boolean 0
trace_replace = boolean 0
trace_money = boolean 0
trace_infect = boolean 0
debug_Infect = boolean 0
trace_para = boolean 0
trace_record = boolean 0
exec_trafo = boolean 1
debug_replace = boolean 0
debug_Global = boolean 0
debug_money = boolean 0
opt_cob122 = boolean 0
debug_RW_gen = boolean 0
opt_CICS = boolean 0
process_money = boolean 0
process_date = boolean 0
trace_phrase = boolean 0
trace_func_defs = boolean 0
trace_func_refs = boolean 0
trace_convert = boolean 0
trace_generator = boolean 0
debug_cics = boolean 0
debug_xref = boolean 0
flag1 = boolean 0
IO_method = int 3
optimize = boolean 0
opt_pack = boolean 0
opt_crossref = boolean 0
opt_used = boolean 0
opt_hash = boolean 0
master = boolean 0
save_code = boolean 1
opts_initialized = boolean 1

method init private static
return
method init_options() static public ;
init()
opts_initialized=1
return
method use_options() static public
if \opts_initialized then init_options
return

method set_pp_options(opt_params = Rexx) static public ;
/* param must be REXX, as words etc is used */
opt_params=opt_params.upper()
-- info('set_pp_options:' opt_params)
if Rey.RT.ReyMsg.line_no > 0 & opt_verbose > 3 then do
Rey.RT.ReyMsg.infol('executIng PP option:'||opt_params)
end--if
loop iw_option = 1 to opt_params.words()
opt0 = opt_params.word(iw_option)
if opt0.pos('=') > 0 then do
parse opt0 opt1 '=' val1
flag1 = 0
if val1 = 'ON' then do
flag1 = 1
end--if
if val1 = 'OFF' then do
flag1 = 0
end--if
end
else do
if opt0.substr(1, 2) = 'NO' then do
opt1 = opt0.substr(3)
flag1 = 0
end
else do
opt1 = opt0
flag1 = 1
end--if
val1 = flag1
/* call info 'set PP-options: opt1='opt1 'val1=' val1 */
end--if
opt1=opt1.upper -- needed for select case
select case opt1

/* new options LOWER and UPPER for Name conversions */
when 'LOWER' then do
opt_lowercase=flag1
opt_uppercase=\flag1
end
when 'UPPER' then do
opt_lowercase=\flag1
opt_uppercase=flag1
end

when 'SCAN' then opt_scan=flag1
when 'PARSE' then opt_parse=flag1
when 'ANALYSE' then opt_analyse=flag1
when 'GENERATE' then opt_generate=flag1
when 'EXEC' then opt_exec=flag1
when 'METRICS' then opt_metrics=flag1 /* release 5.00*/
when 'BANNER' then opt_banner = flag1 /* release 5.00 */
when 'TBN' then trace_builtin = flag1 /* release 5.00 */
when 'TRV' then trace_resolve = flag1 /* release 5.00 */
/* new NetRexx & Java options */
when 'PRO','PROJECT' then do
if val1 = 1 then do
val1 = Rexx(opt_params).word(iw_option + 1)
iw_option = iw_option + 1
end--if
Global.set_project(val1)

end--when
when 'LNOTE' then opt_long_note = flag1

when 'CORE' then do
opt_CORE = flag1
end--when
when 'SHELL' then do
opt_SHELL = flag1
end--when
when 'ANSII' then do
opt_ANSII = flag1
end--when
when 'TU' then do
trace_use = flag1
end--when
when 'TARGS' then do
trace_args = flag1
end--when
when 'TOPCODES' then trace_opcodes=flag1
when 'RESERVED' then do
opt_reserved = flag1
end--when
when 'LBM' then do
log_by_module = flag1
end--when
when 'MAP' then do
opt_MAP = flag1
end--when
when 'GERMAN' then do
opt_GERMAN = flag1
end--when
when 'RTP' then do
opt_RTP = flag1
end--when
when 'BIF' then do
opt_BIF = flag1
end--when
when 'USE' then do
opt_USE = flag1
end--when
when 'CMD' then do
opt_CMD = flag1
end--when
when 'LOCALIZE' then do
opt_localize = flag1
end--when
when 'LOAD' then do
opt_load = flag1
end--when
when 'EXPLICIT' then do
opt_explicit = flag1
end--when
when 'BUILD' then do
opt_build = flag1
end--when
when 'PACKAGE' then do
opt_package = flag1
end--when
when 'Lstring' then do
opt_long_strings = flag1
end--when
when 'IO1' then do
IO_method = 1
end--when
when 'IO2' then do
IO_method = 2
end--when
when 'IO3' then do
IO_method = 3
end--when
when 'OPTIMIZE' then do
optimize = flag1
end--when
when 'TYPE' then do
opt_type = flag1
end--when
when 'SHORT' then do
opt_short = flag1
end--when
when 'LONG' then do
opt_long = flag1
end--when
when 'XREF' then do
opt_xref = flag1
end--when
when 'STACK' then do
opt_Stack = flag1
end--when
when 'LOGO' then do
opt_logo = flag1
end--when
when 'REXX' then do
only_rexx = flag1
end--when
when 'VERBOSE0' then do
Rey.RT.ReyMsg.opt_verbose = 0
end--when
when 'VERBOSE1' then do
Rey.RT.ReyMsg.opt_verbose = 1
end--when
when 'VERBOSE2' then do
Rey.RT.ReyMsg.opt_verbose = 2
end--when
when 'VERBOSE3' then do
Rey.RT.ReyMsg.opt_verbose = 3
end--when
when 'VERBOSE4' then do
Rey.RT.ReyMsg.opt_verbose = 4
end--when
when 'VERBOSE5' then do
Rey.RT.ReyMsg.opt_verbose = 5
end--when
when 'STRICTASSIGN' then do
opt_strictassign = flag1
end--when
when 'STRICTCASE' then do
opt_strictcase = flag1
end--when
when 'TIME' then do
opt_time = flag1
end--when
when 'binary' then do
opt_binary = flag1
end--when
when 'PAUSE' then do
Rey.RT.ReyMsg.opt_pause = flag1
end--when
when 'STATS' then do
opt_stats = flag1
/* my own parser options */
end--when
when 'SOURCE' then do
opt_source = flag1
end--when
when 'FORMAT' then do
opt_format = flag1
end--when
when 'TCODE' then do
trace_code = flag1
end--when
when 'TT' then do
trace_tokens = flag1
end--when
when 'TGEN' then do
trace_generator = flag1
end--when
when 'TTEMPL' then do
trace_templ = flag1
end--when
when 'SYMBOLIC' then do
opt_symbolic = flag1
end--when
when 'PACK' then do
opt_pack = flag1
end--when
when 'WARN' then do
Rey.RT.ReyMsg.opt_warn = flag1
end--when
when 'ATTENTION' then do
opt_attention = flag1
end--when
when 'CROSSREF' then do
opt_crossref = flag1
end--when
when 'NEST' then do
opt_nest = flag1
end--when
when 'CICS' then do
opt_CICS = flag1
end--when
when 'TEST' then do
opt_test = flag1
end--when
when 'USED' then do
opt_used = flag1
end--when
when 'HASH' then do
opt_hash = flag1
end--when
when 'ALL-MON' then do
opt_all_money = flag1
end--when
when 'LIST' then do
opt_list = flag1
end--when
when 'MASTER' then do
master = flag1
end--when
when 'DEBUG' then do
ReyMsg.opt_debug = flag1
end--when
when 'DS' then do
statement_debug=flag1
end--when
when 'DSCAN' then do
debug_scan = 1
end--when
when 'DC' then do
debug_cics = flag1
end--when
when 'DG' then do
debug_Global = flag1
end--when
when 'DDATE' then do
debug_date = flag1
end--when
when 'DM' then do
debug_money = flag1
end--when
when 'DR' then do
debug_replace = flag1
end--when
when 'SAVE' then do
opt_save = flag1
end--when
when 'SAVECODE' then do
save_code = 1
end--when
when 'QUIET' then do
Rey.RT.ReyMsg.opt_quiet = flag1
end--when
when 'TRACE' then do
trace_in = flag1
end--when
when 'TREC' then do
trace_recursion = flag1
end--when
when 'TH' then do
trace_hash = flag1
end--when
when 'TPHR' then do
trace_phrase = flag1
end--when
when 'TDATE' then do
trace_date = flag1
end--when
when 'TS' then do
trace_source = flag1
end--when
when 'TO' then do
trace_out = flag1
end--when
when 'TC' then do
trace_cond = flag1
end--when
when 'TCV' then do
trace_convert = flag1
end--when
when 'TCL' then do
trace_class = flag1
end--when
when 'TCOPY' then do
trace_copy = flag1
end--when
when 'TN','TNEST' then do
trace_nest = flag1
end--when
when 'TY' then do
trace_checks = flag1
end--when
when 'TE' then do
trace_expr = flag1

end--when
when 'TFL' then do
trace_flags = flag1
end--when
when 'TFU' then do
trace_funcs = flag1
end--when
when 'TFD' then do
trace_func_defs = flag1
end--when
when 'TFR' then do
trace_func_refs = flag1
end--when
when 'TI' then do
trace_item = flag1
end--when
when 'TInCL' then do
trace_include = flag1
end--when
when 'TL' then do
trace_line = flag1
end--when
when 'TLAB' then do
trace_labels = flag1
end--when
when 'TP' then do
trace_perform = flag1
end--when
when 'TA' then do
trace_assign = flag1
end--when
when 'TAN' then do
trace_analyser = flag1
end--when
when 'TD' then do
trace_decl = flag1
end--when
when 'TDEF' then do
trace_defs = flag1
end--when
when 'TV' then do
trace_verbs = flag1
end--when
when 'TVAL' then do
trace_value = flag1
end--when
when 'TX' then do
External.trace_externals = flag1
end--when
when 'TEXEC' then do
trace_exec = flag1
end--when
when 'TZ' then do
trace_parse = flag1
end--when
when 'TR' then do
trace_replace = flag1
end--when
when 'TM' then do
trace_money = flag1
end--when
otherwise do
do
if Rey.RT.ReyMsg.line_no > 0 then do
end
else do
Rey.RT.ReyMsg.warning('unknown PP option: '||opt0 'ignored')
end--if
end
end--otherwise
end--select
end--loop iw_option
/* default dependecies of options */
if trace_items | trace_expr | trace_tokens | trace_verbs then do
trace_source = 1
end--if
if opt_binary then do
opt_explicit = 1
end--if
if trace_checks then do
trace_date = 1
end--if
if statement_debug then do
trace_source = 1
end--if
if trace_source then do
opt_source = 1
end--if
if \ opt_format then do
opt_note = 0
end--if

options_initialized=1 /* now the OPTIONS ARE initialized*/
return

Tom. (ths@db-123.com)

Kermit Kiser

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4)

In reply to this post by David Requena

You are close to my idea. There is an undocumented option in
NetRexxScript to turn on "df" (the debug flag) - the option is only
visible on the plugin options screen if you are running as account
"dad". ;-) If not under that account, the option can be turned on/off
via a script. Here is the debug.nrx script from the
//useracct//.jedit/NetRexx directory:

---------------------------------------------------------------------------------------------------------------------

-- This script turns the plugin debug flag on or off
-- ***** warning - setting flag on can produce lots of log and other output

trace results

OPTION_PREFIX = String "options.netrexxscript."
t=boolean 1 -- "true"
f=boolean 0 -- "false"

jEdit.setBooleanProperty(OPTION_PREFIX"debug",\jEdit.getBooleanProperty(OPTION_PREFIX"debug",f))

----------------------------------------------------------------------------------------------------------------------

I guess that I was not confident the code was bulletproof yet, so I left
in the option to trace most of the code in case unexpected problems
might show up. It came in very handy in the development phases to limit
the large volume of debug output to periods of looking at a specific bug!

-- Kermit

David Requena wrote:

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ns.hursley.ibm.com/pipermail/ibm-netrexx/attachments/20100215/ed3b6db5/attachment-0001.html

Thomas.Schneider.Wien

SPAM Re: Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4) (5.9)

Didn't know that up to now.... sorry :-(

Great, thanks for the info.
Tom.
=============================================================================

Kermit Kiser schrieb:

> You are close to my idea. There is an undocumented option in
> NetRexxScript to turn on "df" (the debug flag) - the option is only
> visible on the plugin options screen if you are running as account
> "dad". ;-) If not under that account, the option can be turned on/off
> via a script. Here is the debug.nrx script from the
> //useracct//.jedit/NetRexx directory:
>
> ---------------------------------------------------------------------------------------------------------------------
>
> -- This script turns the plugin debug flag on or off
> -- ***** warning - setting flag on can produce lots of log and other
> output
>
> trace results
>
> OPTION_PREFIX = String "options.netrexxscript."
> t=boolean 1 -- "true"
> f=boolean 0 -- "false"
>
>
> jEdit.setBooleanProperty(OPTION_PREFIX"debug",\jEdit.getBooleanProperty(OPTION_PREFIX"debug",f))
>
> ----------------------------------------------------------------------------------------------------------------------
>
> I guess that I was not confident the code was bulletproof yet, so I
> left in the option to trace most of the code in case unexpected
> problems might show up. It came in very handy in the development
> phases to limit the large volume of debug output to periods of looking
> at a specific bug!
>
> -- Kermit
>
>
> David Requena wrote:
>> Thomas,
>>
>> That is indeed a very valid point.
>> Thought of the possibility of activating debug mode on production
>> code by using some command line switch just after sending my previous
>> email.
>>
>> Thanks,
>> David
>>
>> ---
>> Saludos / Kind regards.
>> David Requena
>>
>>
>> El 15/02/2010 13:44, Thomas Schneider escribi?:
>>> Hi David,
>>> I'm using a similar approach than Kermit, to have an OPTION to
>>> *Dynamically* force DEBUG.
>>> I'm rarely using TRACE, but I do have a lot of conditional debugs in
>>> my program, which I can even
>>> use in the PRODUCTION version just in case a not anticipated
>>> situation occurs in the program
>>> of a client.
>>>
>>> This is just for your info.
>>>
>>> Similarly, I do use a lot of assertions in my code, and did invent
>>> some special notation
>>> for those in my newer programs, e.g.:
>>>
>>> x=3
>>> y=x ::= 3
>>>
>>> where :: introduces an assertion, which is checked at run tyme (in
>>> this case, y MUST be 3)
>>>
>>> when it is NOT 3, an ASSERTION error message is given (and when not
>>> otherwise wanted,
>>> the program aborts immediately.
>>>
>>> Tom.
>>> ========================================================
>>> Kermit, please give us the reson of cyour decision as well.
>>>
>>> David Requena schrieb:
>>>> Kermit,
>>>>
>>>> hmm.. got curious. I noticed the following on your supplied code
>>>> snipped:
>>>>
>>>> if df then trace results
>>>>
>>>> Here I'm supposing df is a property appropriately set up upstream
>>>> as needed.
>>>>
>>>> This is a construct I've used often in other languages but in
>>>> NetRexx I tend to use trace statements liberally as needed. Then
>>>> for production build I use 'options notrace' so no overhead is
>>>> incurred.
>>>>
>>>> Any reason why would you prefer the conditional check over NetRexx'
>>>> options nobinary idiom?
>>>>
>>>> Maybe I'm being dumb but seems to me the conditional would be
>>>> carried to production code whereas options nobinary would eliminate
>>>> any trace (yeah, redundant) of trace code.
>>>>
>>>> ---
>>>> Saludos / Kind regards.
>>>> David Requena
>>>>
>>>> El 13/02/2010 8:16, Kermit Kiser escribi?:
>>>>> Thomas ;
>>>>>
>>>>> There was an exchange about reading Zip/Jar files on the list last
>>>>> month. I don't know if you were on the list then, so I am
>>>>> attaching it here.
>>>>>
>>>>> It sounds like you might be interested in the actual code to read
>>>>> the data also, so here it is:
>>>>>
>>>>> ------------------------------------------------------ general
>>>>> purpose routines
>>>>> -------------------------------------------------------
>>>>>
>>>>> method copyjarentry(jf=JarFile,je=JarEntry,scriptdir=String)
>>>>> -- method to copy a jarfile entry to a directory
>>>>>
>>>>> if df then trace results nrscript=scriptdir||fs||je.getName
>>>>>
>>>>> nrs=File(nrscript)
>>>>>
>>>>> if \nrs.exists then do
>>>>>
>>>>> rc=copyjarentrytofile(jf,je,nrs)
>>>>>
>>>>> if rc=1 then say je.getName "copied to:" scriptdir
>>>>>
>>>>> end
>>>>>
>>>>> method copyjarentrytofile(jf=JarFile,je=JarEntry,nrs=File)
>>>>>
>>>>> if df then trace results
>>>>>
>>>>> do
>>>>>
>>>>> scriptstream=jf.getInputStream(je)
>>>>>
>>>>> if Rexx(je.getName).pos(".class")=0 then do
>>>>>
>>>>>
>>>>> scriptreader=BufferedReader(InputStreamReader(scriptstream))
>>>>>
>>>>> --
>>>>> scriptwriter=BufferedWriter(OutputStreamWriter(outstream))
>>>>>
>>>>> scriptwriter=BufferedWriter(FileWriter(nrs))
>>>>>
>>>>> copyfile(scriptreader,scriptwriter)
>>>>>
>>>>> scriptwriter.close
>>>>>
>>>>> end
>>>>>
>>>>> else do
>>>>>
>>>>> outstream=FileOutputStream(nrs)
>>>>>
>>>>> copybin(scriptstream,outstream)
>>>>>
>>>>> outstream.close
>>>>>
>>>>> end
>>>>>
>>>>> catch badguy=Exception
>>>>>
>>>>> say je.getName "copy error =" badguy
>>>>>
>>>>> return 0
>>>>>
>>>>> end
>>>>>
>>>>> return 1
>>>>>
>>>>> /* methods to copy a file: */
>>>>>
>>>>> method
>>>>> copyfile(ifq=java.io.BufferedReader,ofq=java.io.BufferedWriter)
>>>>> signals IOException
>>>>>
>>>>> if df then trace results
>>>>>
>>>>> line=ifq.readline
>>>>>
>>>>> if line = null then return
>>>>>
>>>>> ofq.write(string line,0,line.length)
>>>>>
>>>>> trace off
>>>>>
>>>>> loop forever
>>>>>
>>>>> line=ifq.readline
>>>>>
>>>>> if line = null then leave
>>>>>
>>>>> ofq.newline()
>>>>>
>>>>> ofq.write(string line,0,line.length)
>>>>>
>>>>> end method copybin(ifq=InputStream,ofq=OutputStream)
>>>>> binary signals IOException
>>>>>
>>>>> bite=ifq.read
>>>>>
>>>>> loop while bite \= -1
>>>>>
>>>>> ofq.write(bite)
>>>>>
>>>>> bite=ifq.read
>>>>>
>>>>> end
>>>>> ---------------------------------------------------------------------------------------------------------------------------------------------------
>>>>>
>>>>>
>>>>>
>>>>> Kinda crude, but maybe it helps.
>>>>>
>>>>> note: fs=File.separator
>>>>>
>>>>> -- Kermit
>>>>>
>>>>>
>>>>> Thomas Schneider wrote:
>>>>>> Hi Divid,
>>>>>> .. many thanks for the Details.
>>>>>> ... MFC already said me the same thins!
>>>>>>
>>>>>> Isn't that (OUR) Universe ...
>>>>>>
>>>>>> ..... a UNIQUE thing ??
>>>>>> Thomas.
>>>>>> ... will read your message (below) tomorrow in more detail.
>>>>>> = Did archive it now, I'm gooing a'sleeeeep ...
>>>>>> =================================================================
>>>>>> David Requena schrieb:
>>>>>>> Thomas,
>>>>>>>
>>>>>>> ZIP and JAR are the same thing.
>>>>>>>
>>>>>>> There is standard stuff in the java library for manipulating
>>>>>>> this format.
>>>>>>> Look at the package java.util.zip classes.
>>>>>>>
>>>>>>> David
>>>>>>>
>>>>>>> 2010/2/12 Thomas Schneider <[hidden email] <mailto:[hidden email]>>
>>>>>>>
>>>>>>> Hi there again,
>>>>>>> first, I must admit that I am *NO* Java *NOR* NetRexx Expert.
>>>>>>> second, I would like to enhance my FILE-Access routines
>>>>>>> (current
>>>>>>> doc in \Rexx2RT\RexxFile
>>>>>>> (in the Run-Time package of www.Rexx2Nrx.com
>>>>>>> <http://www.Rexx2Nrx.com> since years) to do the following:
>>>>>>>
>>>>>>> -- using a special (computer independent) Syntax to read 'the
>>>>>>> members of:
>>>>>>>
>>>>>>> -- a JAR File
>>>>>>> -- a ZIP File
>>>>>>> -- an IBM zOS PDS file
>>>>>>>
>>>>>>> ON demand (e.g. to use linein, lineout, and all the other
>>>>>>> things
>>>>>>> already implemented
>>>>>>> in RexxFile (without the need to CHANGE the 'calling program')
>>>>>>>
>>>>>>> Also, the '/' may now be used as a GENRIC File-separator
>>>>>>> (translated to '\' in the
>>>>>>> Windows case...)
>>>>>>>
>>>>>>> Questions:
>>>>>>>
>>>>>>> 1.) What do you suggest as the proper syntax for the LIBRARY
>>>>>>> Name
>>>>>>> preceeding
>>>>>>> the filename (may it be a JAR, a ZIP, or an IBM z-OS PDS)
>>>>>>>
>>>>>>> 2.) Does anyone of you have a (small) example how I can
>>>>>>> read/write
>>>>>>> the LINES
>>>>>>> of a specified MEMBER of a JAR/ZIP file in Java/NetRexx..
>>>>>>>
>>>>>>> Tom.
>>>>>>>
>>>>>>> PS: I already got from Walter Pachl some (very old & very good)
>>>>>>> examples
>>>>>>> how to read an IBM PDS File thru EXECIO.
>>>>>>>
>>>>>>> I am currently testing my own implementation of EXECIO for
>>>>>>> Rexx2Nrx.
>>>>>>> Thomas Schneider.
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Ibm-netrexx mailing list
>>>>>>> [hidden email]
>>>>>>> <mailto:[hidden email]>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Saludos / Regards,
>>>>>>> David Requena
>>>>>>>
>>>>>>> ------------------------------------------------------------------------
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Ibm-netrexx mailing list
>>>>>>> [hidden email]
>>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Ibm-netrexx mailing list
>>>>>> [hidden email]
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Ibm-netrexx mailing list
>>>>> [hidden email]
>>>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>>
>>>> _______________________________________________
>>>> Ibm-netrexx mailing list
>>>> [hidden email]
>>>>
>>>
>>> _______________________________________________
>>> Ibm-netrexx mailing list
>>> [hidden email]
>>>
>> _______________________________________________
>> Ibm-netrexx mailing list
>> [hidden email]
>>
>>
>>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>

Tom. (ths@db-123.com)

David Requena

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4)

In reply to this post by Kermit Kiser

Kermit,

Yes, I see. It is in fact a very similar pattern to what I've used often
when working with big Lostusscript programs in products like Lotus
Workfow, where you could attach the debugger.

I guess I jumped too fast on that one without really thinking much about
what I was actually going to ask.Or in other words: yes, I was being
dumb :-)

According to the language definition 1.00:
"In some implementations, the use of trace instructions may
substantially increase the size of classes and the execution time of
methods affected by tracing."

I wonder how much that increase might be. To be honest I always assumed
that if it was big enough to be mentioned in the docs, it must be quite
significant.

---
Saludos / Kind regards.
David Requena

El 15/02/2010 17:18, Kermit Kiser escribi?:

> You are close to my idea. There is an undocumented option in
> NetRexxScript to turn on "df" (the debug flag) - the option is only
> visible on the plugin options screen if you are running as account
> "dad". ;-) If not under that account, the option can be turned on/off
> via a script. Here is the debug.nrx script from the
> //useracct//.jedit/NetRexx directory:
>
> ---------------------------------------------------------------------------------------------------------------------
>
> -- This script turns the plugin debug flag on or off
> -- ***** warning - setting flag on can produce lots of log and other
> output
>
> trace results
>
> OPTION_PREFIX = String "options.netrexxscript."
> t=boolean 1 -- "true"
> f=boolean 0 -- "false"
>
>
> jEdit.setBooleanProperty(OPTION_PREFIX"debug",\jEdit.getBooleanProperty(OPTION_PREFIX"debug",f))
>
> ----------------------------------------------------------------------------------------------------------------------
>
> I guess that I was not confident the code was bulletproof yet, so I
> left in the option to trace most of the code in case unexpected
> problems might show up. It came in very handy in the development
> phases to limit the large volume of debug output to periods of looking
> at a specific bug!
>
> -- Kermit
>
>
> David Requena wrote:
>> Thomas,
>>
>> That is indeed a very valid point.
>> Thought of the possibility of activating debug mode on production
>> code by using some command line switch just after sending my previous
>> email.
>>
>> Thanks,
>> David
>>
>> ---
>> Saludos / Kind regards.
>> David Requena
>>
>>
>> El 15/02/2010 13:44, Thomas Schneider escribi?:
>>> Hi David,
>>> I'm using a similar approach than Kermit, to have an OPTION to
>>> *Dynamically* force DEBUG.
>>> I'm rarely using TRACE, but I do have a lot of conditional debugs in
>>> my program, which I can even
>>> use in the PRODUCTION version just in case a not anticipated
>>> situation occurs in the program
>>> of a client.
>>>
>>> This is just for your info.
>>>
>>> Similarly, I do use a lot of assertions in my code, and did invent
>>> some special notation
>>> for those in my newer programs, e.g.:
>>>
>>> x=3
>>> y=x ::= 3
>>>
>>> where :: introduces an assertion, which is checked at run tyme (in
>>> this case, y MUST be 3)
>>>
>>> when it is NOT 3, an ASSERTION error message is given (and when not
>>> otherwise wanted,
>>> the program aborts immediately.
>>>
>>> Tom.
>>> ========================================================
>>> Kermit, please give us the reson of cyour decision as well.
>>>
>>> David Requena schrieb:
>>>> Kermit,
>>>>
>>>> hmm.. got curious. I noticed the following on your supplied code
>>>> snipped:
>>>>
>>>> if df then trace results
>>>>
>>>> Here I'm supposing df is a property appropriately set up upstream
>>>> as needed.
>>>>
>>>> This is a construct I've used often in other languages but in
>>>> NetRexx I tend to use trace statements liberally as needed. Then
>>>> for production build I use 'options notrace' so no overhead is
>>>> incurred.
>>>>
>>>> Any reason why would you prefer the conditional check over NetRexx'
>>>> options nobinary idiom?
>>>>
>>>> Maybe I'm being dumb but seems to me the conditional would be
>>>> carried to production code whereas options nobinary would eliminate
>>>> any trace (yeah, redundant) of trace code.
>>>>
>>>> ---
>>>> Saludos / Kind regards.
>>>> David Requena
>>>>
>>>> El 13/02/2010 8:16, Kermit Kiser escribi?:
>>>>> Thomas ;
>>>>>
>>>>> There was an exchange about reading Zip/Jar files on the list last
>>>>> month. I don't know if you were on the list then, so I am
>>>>> attaching it here.
>>>>>
>>>>> It sounds like you might be interested in the actual code to read
>>>>> the data also, so here it is:
>>>>>
>>>>> ------------------------------------------------------ general
>>>>> purpose routines
>>>>> -------------------------------------------------------
>>>>>
>>>>> method copyjarentry(jf=JarFile,je=JarEntry,scriptdir=String)
>>>>> -- method to copy a jarfile entry to a directory
>>>>>
>>>>> if df then trace results nrscript=scriptdir||fs||je.getName
>>>>>
>>>>> nrs=File(nrscript)
>>>>>
>>>>> if \nrs.exists then do
>>>>>
>>>>> rc=copyjarentrytofile(jf,je,nrs)
>>>>>
>>>>> if rc=1 then say je.getName "copied to:" scriptdir
>>>>>
>>>>> end
>>>>>
>>>>> method copyjarentrytofile(jf=JarFile,je=JarEntry,nrs=File)
>>>>>
>>>>> if df then trace results
>>>>>
>>>>> do
>>>>>
>>>>> scriptstream=jf.getInputStream(je)
>>>>>
>>>>> if Rexx(je.getName).pos(".class")=0 then do
>>>>>
>>>>>
>>>>> scriptreader=BufferedReader(InputStreamReader(scriptstream))
>>>>>
>>>>> --
>>>>> scriptwriter=BufferedWriter(OutputStreamWriter(outstream))
>>>>>
>>>>> scriptwriter=BufferedWriter(FileWriter(nrs))
>>>>>
>>>>> copyfile(scriptreader,scriptwriter)
>>>>>
>>>>> scriptwriter.close
>>>>>
>>>>> end
>>>>>
>>>>> else do
>>>>>
>>>>> outstream=FileOutputStream(nrs)
>>>>>
>>>>> copybin(scriptstream,outstream)
>>>>>
>>>>> outstream.close
>>>>>
>>>>> end
>>>>>
>>>>> catch badguy=Exception
>>>>>
>>>>> say je.getName "copy error =" badguy
>>>>>
>>>>> return 0
>>>>>
>>>>> end
>>>>>
>>>>> return 1
>>>>>
>>>>> /* methods to copy a file: */
>>>>>
>>>>> method
>>>>> copyfile(ifq=java.io.BufferedReader,ofq=java.io.BufferedWriter)
>>>>> signals IOException
>>>>>
>>>>> if df then trace results
>>>>>
>>>>> line=ifq.readline
>>>>>
>>>>> if line = null then return
>>>>>
>>>>> ofq.write(string line,0,line.length)
>>>>>
>>>>> trace off
>>>>>
>>>>> loop forever
>>>>>
>>>>> line=ifq.readline
>>>>>
>>>>> if line = null then leave
>>>>>
>>>>> ofq.newline()
>>>>>
>>>>> ofq.write(string line,0,line.length)
>>>>>
>>>>> end method copybin(ifq=InputStream,ofq=OutputStream)
>>>>> binary signals IOException
>>>>>
>>>>> bite=ifq.read
>>>>>
>>>>> loop while bite \= -1
>>>>>
>>>>> ofq.write(bite)
>>>>>
>>>>> bite=ifq.read
>>>>>
>>>>> end
>>>>> ---------------------------------------------------------------------------------------------------------------------------------------------------
>>>>>
>>>>>
>>>>>
>>>>> Kinda crude, but maybe it helps.
>>>>>
>>>>> note: fs=File.separator
>>>>>
>>>>> -- Kermit
>>>>>
>>>>>
>>>>> Thomas Schneider wrote:
>>>>>> Hi Divid,
>>>>>> .. many thanks for the Details.
>>>>>> ... MFC already said me the same thins!
>>>>>>
>>>>>> Isn't that (OUR) Universe ...
>>>>>>
>>>>>> ..... a UNIQUE thing ??
>>>>>> Thomas.
>>>>>> ... will read your message (below) tomorrow in more detail.
>>>>>> = Did archive it now, I'm gooing a'sleeeeep ...
>>>>>> =================================================================
>>>>>> David Requena schrieb:
>>>>>>> Thomas,
>>>>>>>
>>>>>>> ZIP and JAR are the same thing.
>>>>>>>
>>>>>>> There is standard stuff in the java library for manipulating
>>>>>>> this format.
>>>>>>> Look at the package java.util.zip classes.
>>>>>>>
>>>>>>> David
>>>>>>>
>>>>>>> 2010/2/12 Thomas Schneider <[hidden email] <mailto:[hidden email]>>
>>>>>>>
>>>>>>> Hi there again,
>>>>>>> first, I must admit that I am *NO* Java *NOR* NetRexx Expert.
>>>>>>> second, I would like to enhance my FILE-Access routines
>>>>>>> (current
>>>>>>> doc in \Rexx2RT\RexxFile
>>>>>>> (in the Run-Time package of www.Rexx2Nrx.com
>>>>>>> <http://www.Rexx2Nrx.com> since years) to do the following:
>>>>>>>
>>>>>>> -- using a special (computer independent) Syntax to read 'the
>>>>>>> members of:
>>>>>>>
>>>>>>> -- a JAR File
>>>>>>> -- a ZIP File
>>>>>>> -- an IBM zOS PDS file
>>>>>>>
>>>>>>> ON demand (e.g. to use linein, lineout, and all the other
>>>>>>> things
>>>>>>> already implemented
>>>>>>> in RexxFile (without the need to CHANGE the 'calling program')
>>>>>>>
>>>>>>> Also, the '/' may now be used as a GENRIC File-separator
>>>>>>> (translated to '\' in the
>>>>>>> Windows case...)
>>>>>>>
>>>>>>> Questions:
>>>>>>>
>>>>>>> 1.) What do you suggest as the proper syntax for the LIBRARY
>>>>>>> Name
>>>>>>> preceeding
>>>>>>> the filename (may it be a JAR, a ZIP, or an IBM z-OS PDS)
>>>>>>>
>>>>>>> 2.) Does anyone of you have a (small) example how I can
>>>>>>> read/write
>>>>>>> the LINES
>>>>>>> of a specified MEMBER of a JAR/ZIP file in Java/NetRexx..
>>>>>>>
>>>>>>> Tom.
>>>>>>>
>>>>>>> PS: I already got from Walter Pachl some (very old & very good)
>>>>>>> examples
>>>>>>> how to read an IBM PDS File thru EXECIO.
>>>>>>>
>>>>>>> I am currently testing my own implementation of EXECIO for
>>>>>>> Rexx2Nrx.
>>>>>>> Thomas Schneider.
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Ibm-netrexx mailing list
>>>>>>> [hidden email] <mailto:[hidden email]>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Saludos / Regards,
>>>>>>> David Requena
>>>>>>>
>>>>>>> ------------------------------------------------------------------------
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Ibm-netrexx mailing list
>>>>>>> [hidden email]
>>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Ibm-netrexx mailing list
>>>>>> [hidden email]
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Ibm-netrexx mailing list
>>>>> [hidden email]
>>>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>>
>>>> _______________________________________________
>>>> Ibm-netrexx mailing list
>>>> [hidden email]
>>>>
>>>
>>> _______________________________________________
>>> Ibm-netrexx mailing list
>>> [hidden email]
>>>
>> _______________________________________________
>> Ibm-netrexx mailing list
>> [hidden email]
>>
>>
>>
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ns.hursley.ibm.com/pipermail/ibm-netrexx/attachments/20100215/aecdb213/attachment-0001.html

Mike Cowlishaw

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Filesin Java &/ NetRexx (5.4)

According to the language definition 1.00:
"In some implementations, the use of trace instructions may substantially
increase the size of classes and the execution time of methods affected by
tracing."

I wonder how much that increase might be. To be honest I always assumed that
if it was big enough to be mentioned in the docs, it must be quite
significant.

Yes .. the code to generate the traces has to be in place (unless turned
off by the options/compile-time flag). To decide whether that's significant
or not, in your context, choose a reasonably large class, and compile it.
Observe the size of the class file.

Next, add a trace statement at the top of the class -- again compile and
observe the resulting size. This will vary depening on the trace option,
of course, so several variations would be interesting. (Report back here?)

Size of class might not matter too much nowadays, although it will slow down
loading. But worth being aware of it.

Mike

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ns.hursley.ibm.com/pipermail/ibm-netrexx/attachments/20100215/601158a3/attachment.html

Thomas.Schneider.Wien

SPAM Re: Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4) (6.2)

In reply to this post by David Requena

Hi David, Kermit,
I did use (long time ago, say 6-7 years) look at Mike F. Cowlishaws
generated Java Code,
... and I also seem to remember *some* discussions of them:

What MIKE is, obviously, doing, is putting *all String* Literals at the top.
Named $1,$2, etc

He put's those String literals all in a VERY BIG Line (hence you will
see them NOT with an Editor)

I think, he made this decision, to keeep the string's (or Rexx Strings)
*initialized* once forwever.

As far as I can re-call my tests (years ago, sorry) I think I cann
recall he does have the
same strategy for *the source Code*.

I'm having another one, but that should no be discussion point yet ...

When I might help you, simply give me a mail...
Thomas.
=======================================================================
David Requena schrieb:

> Kermit,
>
> Yes, I see. It is in fact a very similar pattern to what I've used
> often when working with big Lostusscript programs in products like
> Lotus Workfow, where you could attach the debugger.
>
> I guess I jumped too fast on that one without really thinking much
> about what I was actually going to ask.Or in other words: yes, I was
> being dumb :-)
>
> According to the language definition 1.00:
> "In some implementations, the use of trace instructions may
> substantially increase the size of classes and the execution time of
> methods affected by tracing."
>
> I wonder how much that increase might be. To be honest I always
> assumed that if it was big enough to be mentioned in the docs, it must
> be quite significant.
> ---
> Saludos / Kind regards.
> David Requena
>
> El 15/02/2010 17:18, Kermit Kiser escribi?:
>> You are close to my idea. There is an undocumented option in
>> NetRexxScript to turn on "df" (the debug flag) - the option is only
>> visible on the plugin options screen if you are running as account
>> "dad". ;-) If not under that account, the option can be turned
>> on/off via a script. Here is the debug.nrx script from the
>> //useracct//.jedit/NetRexx directory:
>>
>> ---------------------------------------------------------------------------------------------------------------------
>>
>> -- This script turns the plugin debug flag on or off
>> -- ***** warning - setting flag on can produce lots of log and other
>> output
>>
>> trace results
>>
>> OPTION_PREFIX = String "options.netrexxscript."
>> t=boolean 1 -- "true"
>> f=boolean 0 -- "false"
>>
>>
>> jEdit.setBooleanProperty(OPTION_PREFIX"debug",\jEdit.getBooleanProperty(OPTION_PREFIX"debug",f))
>>
>> ----------------------------------------------------------------------------------------------------------------------
>>
>> I guess that I was not confident the code was bulletproof yet, so I
>> left in the option to trace most of the code in case unexpected
>> problems might show up. It came in very handy in the development
>> phases to limit the large volume of debug output to periods of
>> looking at a specific bug!
>>
>> -- Kermit
>>
>>
>> David Requena wrote:
>>> Thomas,
>>>
>>> That is indeed a very valid point.
>>> Thought of the possibility of activating debug mode on production
>>> code by using some command line switch just after sending my
>>> previous email.
>>>
>>> Thanks,
>>> David
>>>
>>> ---
>>> Saludos / Kind regards.
>>> David Requena
>>>
>>>
>>> El 15/02/2010 13:44, Thomas Schneider escribi?:
>>>> Hi David,
>>>> I'm using a similar approach than Kermit, to have an OPTION to
>>>> *Dynamically* force DEBUG.
>>>> I'm rarely using TRACE, but I do have a lot of conditional debugs
>>>> in my program, which I can even
>>>> use in the PRODUCTION version just in case a not anticipated
>>>> situation occurs in the program
>>>> of a client.
>>>>
>>>> This is just for your info.
>>>>
>>>> Similarly, I do use a lot of assertions in my code, and did invent
>>>> some special notation
>>>> for those in my newer programs, e.g.:
>>>>
>>>> x=3
>>>> y=x ::= 3
>>>>
>>>> where :: introduces an assertion, which is checked at run tyme (in
>>>> this case, y MUST be 3)
>>>>
>>>> when it is NOT 3, an ASSERTION error message is given (and when
>>>> not otherwise wanted,
>>>> the program aborts immediately.
>>>>
>>>> Tom.
>>>> ========================================================
>>>> Kermit, please give us the reson of cyour decision as well.
>>>>
>>>> David Requena schrieb:
>>>>> Kermit,
>>>>>
>>>>> hmm.. got curious. I noticed the following on your supplied code
>>>>> snipped:
>>>>>
>>>>> if df then trace results
>>>>>
>>>>> Here I'm supposing df is a property appropriately set up upstream
>>>>> as needed.
>>>>>
>>>>> This is a construct I've used often in other languages but in
>>>>> NetRexx I tend to use trace statements liberally as needed. Then
>>>>> for production build I use 'options notrace' so no overhead is
>>>>> incurred.
>>>>>
>>>>> Any reason why would you prefer the conditional check over
>>>>> NetRexx' options nobinary idiom?
>>>>>
>>>>> Maybe I'm being dumb but seems to me the conditional would be
>>>>> carried to production code whereas options nobinary would
>>>>> eliminate any trace (yeah, redundant) of trace code.
>>>>>
>>>>> ---
>>>>> Saludos / Kind regards.
>>>>> David Requena
>>>>>
>>>>> El 13/02/2010 8:16, Kermit Kiser escribi?:
>>>>>> Thomas ;
>>>>>>
>>>>>> There was an exchange about reading Zip/Jar files on the list
>>>>>> last month. I don't know if you were on the list then, so I am
>>>>>> attaching it here.
>>>>>>
>>>>>> It sounds like you might be interested in the actual code to read
>>>>>> the data also, so here it is:
>>>>>>
>>>>>> ------------------------------------------------------ general
>>>>>> purpose routines
>>>>>> -------------------------------------------------------
>>>>>>
>>>>>> method copyjarentry(jf=JarFile,je=JarEntry,scriptdir=String)
>>>>>> -- method to copy a jarfile entry to a directory
>>>>>>
>>>>>> if df then trace results nrscript=scriptdir||fs||je.getName
>>>>>>
>>>>>> nrs=File(nrscript)
>>>>>>
>>>>>> if \nrs.exists then do
>>>>>>
>>>>>> rc=copyjarentrytofile(jf,je,nrs)
>>>>>>
>>>>>> if rc=1 then say je.getName "copied to:" scriptdir
>>>>>>
>>>>>> end
>>>>>>
>>>>>> method copyjarentrytofile(jf=JarFile,je=JarEntry,nrs=File)
>>>>>>
>>>>>> if df then trace results
>>>>>>
>>>>>> do
>>>>>>
>>>>>> scriptstream=jf.getInputStream(je)
>>>>>>
>>>>>> if Rexx(je.getName).pos(".class")=0 then do
>>>>>>
>>>>>>
>>>>>> scriptreader=BufferedReader(InputStreamReader(scriptstream))
>>>>>>
>>>>>> --
>>>>>> scriptwriter=BufferedWriter(OutputStreamWriter(outstream))
>>>>>>
>>>>>> scriptwriter=BufferedWriter(FileWriter(nrs))
>>>>>>
>>>>>> copyfile(scriptreader,scriptwriter)
>>>>>>
>>>>>> scriptwriter.close
>>>>>>
>>>>>> end
>>>>>>
>>>>>> else do
>>>>>>
>>>>>> outstream=FileOutputStream(nrs)
>>>>>>
>>>>>> copybin(scriptstream,outstream)
>>>>>>
>>>>>> outstream.close
>>>>>>
>>>>>> end
>>>>>>
>>>>>> catch badguy=Exception
>>>>>>
>>>>>> say je.getName "copy error =" badguy
>>>>>>
>>>>>> return 0
>>>>>>
>>>>>> end
>>>>>>
>>>>>> return 1
>>>>>>
>>>>>> /* methods to copy a file: */
>>>>>>
>>>>>> method
>>>>>> copyfile(ifq=java.io.BufferedReader,ofq=java.io.BufferedWriter)
>>>>>> signals IOException
>>>>>>
>>>>>> if df then trace results
>>>>>>
>>>>>> line=ifq.readline
>>>>>>
>>>>>> if line = null then return
>>>>>>
>>>>>> ofq.write(string line,0,line.length)
>>>>>>
>>>>>> trace off
>>>>>>
>>>>>> loop forever
>>>>>>
>>>>>> line=ifq.readline
>>>>>>
>>>>>> if line = null then leave
>>>>>>
>>>>>> ofq.newline()
>>>>>>
>>>>>> ofq.write(string line,0,line.length)
>>>>>>
>>>>>> end method copybin(ifq=InputStream,ofq=OutputStream)
>>>>>> binary signals IOException
>>>>>>
>>>>>> bite=ifq.read
>>>>>>
>>>>>> loop while bite \= -1
>>>>>>
>>>>>> ofq.write(bite)
>>>>>>
>>>>>> bite=ifq.read
>>>>>>
>>>>>> end
>>>>>> ---------------------------------------------------------------------------------------------------------------------------------------------------
>>>>>>
>>>>>>
>>>>>>
>>>>>> Kinda crude, but maybe it helps.
>>>>>>
>>>>>> note: fs=File.separator
>>>>>>
>>>>>> -- Kermit
>>>>>>
>>>>>>
>>>>>> Thomas Schneider wrote:
>>>>>>> Hi Divid,
>>>>>>> .. many thanks for the Details.
>>>>>>> ... MFC already said me the same thins!
>>>>>>>
>>>>>>> Isn't that (OUR) Universe ...
>>>>>>>
>>>>>>> ..... a UNIQUE thing ??
>>>>>>> Thomas.
>>>>>>> ... will read your message (below) tomorrow in more detail.
>>>>>>> = Did archive it now, I'm gooing a'sleeeeep ...
>>>>>>> =================================================================
>>>>>>> David Requena schrieb:
>>>>>>>> Thomas,
>>>>>>>>
>>>>>>>> ZIP and JAR are the same thing.
>>>>>>>>
>>>>>>>> There is standard stuff in the java library for manipulating
>>>>>>>> this format.
>>>>>>>> Look at the package java.util.zip classes.
>>>>>>>>
>>>>>>>> David
>>>>>>>>
>>>>>>>> 2010/2/12 Thomas Schneider <[hidden email]
>>>>>>>> <mailto:[hidden email]>>
>>>>>>>>
>>>>>>>> Hi there again,
>>>>>>>> first, I must admit that I am *NO* Java *NOR* NetRexx Expert.
>>>>>>>> second, I would like to enhance my FILE-Access routines
>>>>>>>> (current
>>>>>>>> doc in \Rexx2RT\RexxFile
>>>>>>>> (in the Run-Time package of www.Rexx2Nrx.com
>>>>>>>> <http://www.Rexx2Nrx.com> since years) to do the following:
>>>>>>>>
>>>>>>>> -- using a special (computer independent) Syntax to read 'the
>>>>>>>> members of:
>>>>>>>>
>>>>>>>> -- a JAR File
>>>>>>>> -- a ZIP File
>>>>>>>> -- an IBM zOS PDS file
>>>>>>>>
>>>>>>>> ON demand (e.g. to use linein, lineout, and all the other
>>>>>>>> things
>>>>>>>> already implemented
>>>>>>>> in RexxFile (without the need to CHANGE the 'calling program')
>>>>>>>>
>>>>>>>> Also, the '/' may now be used as a GENRIC File-separator
>>>>>>>> (translated to '\' in the
>>>>>>>> Windows case...)
>>>>>>>>
>>>>>>>> Questions:
>>>>>>>>
>>>>>>>> 1.) What do you suggest as the proper syntax for the
>>>>>>>> LIBRARY Name
>>>>>>>> preceeding
>>>>>>>> the filename (may it be a JAR, a ZIP, or an IBM z-OS PDS)
>>>>>>>>
>>>>>>>> 2.) Does anyone of you have a (small) example how I can
>>>>>>>> read/write
>>>>>>>> the LINES
>>>>>>>> of a specified MEMBER of a JAR/ZIP file in Java/NetRexx..
>>>>>>>>
>>>>>>>> Tom.
>>>>>>>>
>>>>>>>> PS: I already got from Walter Pachl some (very old & very
>>>>>>>> good)
>>>>>>>> examples
>>>>>>>> how to read an IBM PDS File thru EXECIO.
>>>>>>>>
>>>>>>>> I am currently testing my own implementation of EXECIO for
>>>>>>>> Rexx2Nrx.
>>>>>>>> Thomas Schneider.
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> Ibm-netrexx mailing list
>>>>>>>> [hidden email]
>>>>>>>> <mailto:[hidden email]>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> Saludos / Regards,
>>>>>>>> David Requena
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------
>>>>>>>>
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> Ibm-netrexx mailing list
>>>>>>>> [hidden email]
>>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Ibm-netrexx mailing list
>>>>>>> [hidden email]
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Ibm-netrexx mailing list
>>>>>> [hidden email]
>>>>>>
>>>>> ------------------------------------------------------------------------
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Ibm-netrexx mailing list
>>>>> [hidden email]
>>>>>
>>>>
>>>> _______________________________________________
>>>> Ibm-netrexx mailing list
>>>> [hidden email]
>>>>
>>> _______________________________________________
>>> Ibm-netrexx mailing list
>>> [hidden email]
>>>
>>>
>>>
>>
>> _______________________________________________
>> Ibm-netrexx mailing list
>> [hidden email]
>>
>>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>

Tom. (ths@db-123.com)

Thomas.Schneider.Wien

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Filesin Java &/ NetRexx (5.4)

In reply to this post by Mike Cowlishaw

Hello there,

just my 2 % :

The current implementation of NetRexx does copy all SOURCE-Lines of the
source in question
into the TARGET (Java) Program ....

... At least when I did read the generated Java Code right (with the
past releases).

I am (personally) following another PHILOSOPHY:

But I do NOT like to discuss this, unless other members of this list are
interested....

Thomas.
================================================================

Mike Cowlishaw schrieb:

> According to the language definition 1.00:
> "In some implementations, the use of trace instructions may
> substantially increase the size of classes and the execution time of
> methods affected by tracing."
>
> I wonder how much that increase might be. To be honest I always
> assumed that if it was big enough to be mentioned in the docs, it must
> be quite significant.
>
> Yes .. the code to generate the traces has to be in place (unless
> turned off by the options/compile-time flag). To decide whether
> that's significant or not, in your context, choose a reasonably large
> class, and compile it. Observe the size of the class file.
>
> Next, add a trace statement at the top of the class -- again compile
> and observe the resulting size. This will vary depening on the trace
> option, of course, so several variations would be interesting.
> (Report back here?)
>
> Size of class might not matter too much nowadays, although it will
> slow down loading. But worth being aware of it.
>
> Mike
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>

Tom. (ths@db-123.com)

Mike Cowlishaw

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR FilesinJava &/ NetRexx (5.4)

> The current implementation of NetRexx does copy all
> SOURCE-Lines of the source in question into the TARGET (Java)
> Program ....

That's nonsense Tom. The source is only there if the programmer requires it
(for tracing, for example). A NetRexx class is [almost] identical to a Java
class for the same program if one wants it to be (the binary applet
examples, for example). I seem to recall that the NetRexx version is one
byte shorter, in fact, because the name of the source file is .nrx instead
of .java ...

Mike

Thomas.Schneider.Wien

TRACE in NetRexx vs. 2.05

NO, Mike,

*it's* ***NO *** NON-Sense***

what you are actually doing, If I am reading your Generated Java Code,
(unless somebody did change it recently):

When TRACE is used, you are putting the WHOLE source into the COMPILED
program(as a COPY...)

I am currently NOT sure, if you do this with all OPETIONS of TRACE (like
TRACE VAR ::)
But I would have means on my fingers that I sould intantly RE-TRY this
again, as I did some 7 years ago ...)

Kind regards,
Tom.

PS: See you at SpeleoTrove

PPS: And thank you so much for the (personal) explanation what
SpeleoTrove stands for.

Wouldn't it be the right time to explain it to this list, Mike ???
====================================================================
Mike Cowlishaw schrieb:

>> The current implementation of NetRexx does copy all
>> SOURCE-Lines of the source in question into the TARGET (Java)
>> Program ....
>>
>
> That's nonsense Tom. The source is only there if the programmer requires it
> (for tracing, for example). A NetRexx class is [almost] identical to a Java
> class for the same program if one wants it to be (the binary applet
> examples, for example). I seem to recall that the NetRexx version is one
> byte shorter, in fact, because the name of the source file is .nrx instead
> of .java ...
>
> Mike
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>
>

Tom. (ths@db-123.com)

David Requena

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Filesin Java &/ NetRexx (5.4)

In reply to this post by Mike Cowlishaw

Mike,

I'll be checking out of curiosity. Should have been done this in the first place.

I agree class size is not that much of a concern nowadays. OTH execution time could definitely be.

I've a class which, if so configured, will apply a complex regex to a whole nrx buffer*on every user key press*. Of course it's doing a lot of other things. I wonder if it will be noticeable.

Unfortunately my timings are bound to be inexact but surely a trend could be seen.

I this was running standalone I would keep generated java files and instrument them in NetBeans for some proper profiling but, being run inside jEdit, I'm not sure at all this can be done.

Be back to the list the having some numbers.

---
Saludos / Kind regards.
David Requena

El 15/02/2010 21:41, Mike Cowlishaw escribi?:

> According to the language definition 1.00:
> "In some implementations, the use of trace instructions may
> substantially increase the size of classes and the execution time of
> methods affected by tracing."
>
> I wonder how much that increase might be. To be honest I always
> assumed that if it was big enough to be mentioned in the docs, it must
> be quite significant.
> Yes .. the code to generate the traces has to be in place (unless
> turned off by the options/compile-time flag). To decide whether
> that's significant or not, in your context, choose a reasonably large
> class, and compile it. Observe the size of the class file.
> Next, add a trace statement at the top of the class -- again compile
> and observe the resulting size. This will vary depening on the trace
> option, of course, so several variations would be interesting.
> (Report back here?)
> Size of class might not matter too much nowadays, although it will
> slow down loading. But worth being aware of it.
> Mike
>
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ns.hursley.ibm.com/pipermail/ibm-netrexx/attachments/20100216/d253b1ea/attachment.html

Reading ZIP/JAR Files in Java &/ NetRexx

Reading ZIP/JAR Files in Java &/ NetRexx

Reading ZIP/JAR Files in Java &/ NetRexx

Reading ZIP/JAR Files in Java &/ NetRexx

Reading ZIP/JAR Files in Java &/ NetRexx

Reading ZIP/JAR Files in Java &/ NetRexx

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx

**** SPAM **** Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4)

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4)

Rey OPTIONS.

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4)

**** SPAM **** Re: Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4) (5.9)

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4)

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Filesin Java &/ NetRexx (5.4)

**** SPAM **** Re: Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4) (6.2)

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Filesin Java &/ NetRexx (5.4)

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR FilesinJava &/ NetRexx (5.4)

TRACE in NetRexx vs. 2.05

Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Filesin Java &/ NetRexx (5.4)

SPAM Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4)

SPAM Re: Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4) (5.9)

SPAM Re: Trace construct - was: Re: [Ibm-netrexx] Reading ZIP/JAR Files in Java &/ NetRexx (5.4) (6.2)