Re: [netrexx-pipelines] Migrate OS/2 pipelines to NetRexx Pipes

[<hp...@we...>]

Hi Jeff!

Am 26.06.2022 um 01:14 schrieb Jeff Hennick:
> [...]
> This is the first I have heard of a problem with XLATE/TRANSLATE, except that
> the various code pages are not supported "yet."  (I don't see much use, in the
> NetRexx world, for all the EBCDIC pages that CMS supports.)

You certainly remember, recently you pointed me to M. Cowlishaw's
'The REXX Language', section 3: 'Fundamental Language Concepts'.
In paragraph 'Be adaptable' he says:

> A language [...] certainly will be used for applications not foreseen by the designer.

Therefore, your "don't see much use" is no criterion. For example,
rarely but it happens that "normal" FTP transfer fails so I take a
runaround by binary transfer and xlate/deblock afterwards. By FTP
transfer I mean between PC and z/VM. (It fails for Intel load
files with lines terminated either by lf or cr instead of crlf.)

The failure xlate does not specify any codepage:

> xlate 1 00-FF 3A 20-3F 20-3F 40-5F 00-1f 60-66 40-46

The arguments are:
1           = inputRange
00-FF 3A    = xrange-xchar, the 'basic table'
20-3F 20-3F = xrange-xrange, 1st mod. of table
40-5F 00-1f = xrange-xrange, 2nd mod. of table
60-66 40-46 = xrange-xrange, 3rd mod. of table

It should transliterate from ASCII to HP41 display coded
characters as used in its firmware, see attached table from ZENROM
users' manual.

> These are the tests it currently passes:
>
>> -- xlate xrange char ?
>> -- xlate xrange char char char?
>> -- xlate ?
>> -- xlate xrange char char char (overlap)?
>> -- xlate xrange (multiple overlap)?
>> -- xlate pairs reversed ?
>> -- xlate xrange short-xrange ?
>> -- xlate short-xrange xrange ?
> What am I missing?

See a. m. details.

BTW, what is 'short-xrange', the "Pipelines Guide and Reference"
does not mention 'short' in relation of xlate.

Best,
M.