Opened 16 years ago

Closed 16 years ago

Last modified 13 years ago

#8 closed defect (invalid)

Rexx script filled with all zeroes and chkdsk errors.

Reported by: Rob Hamerling Owned by: somebody
Priority: minor Milestone:
Component: IFS Version:
Severity: Keywords:
Cc: KO Myung-Hun

Description

Problem with a Rexx script on a FAT32 volume.

After a number of cycles of updating and execution a Rexx script the script file was 'suddenly' completely filled with hex zeroes. For editing of this type of files I usually use Tim Baldwin's T2 editor (now called Tedit) which never failed on me. I experienced the problem for the first time after I migrated to Fat32.ifs 0.9.13, but when I went back to 0.9.12 (of the eCS 2.0 RC5 CD) it appeared again. I must say that I seldomly use a FAT32 volume to directy edit and execute Rexx scripts. So the cause of the problem may have been there all the time but I never hit it. After I moved the file to a JFS volume where it did not appear anymore. During xcopy (/H/O/T/S/E/R/V) of the directory with the all-zero Rexx script xcopy terminated on this file. Chkdsk /F reported lost clusters and recoverd three files (each the size of the allocation uniut: 32K), and in each I found back the origical Rexx script. Although it occurred three times I'm not sure if I can reproduce the problem. But even if I can what does it prove? I suspect Fat32.ifs is the failing component, but how can be sure? Which other relevant information could be collected? Since I have a work-around (moving the file to a JFS volume) I have marked it as 'minor' priority.

Regards, Rob.

Change History (6)

comment:1 by dwgras, 16 years ago

Rob, where can this editor be downloaded form? I initially thought you were referring to the tedit that comes with OS/2 and eComstation.

David

comment:2 by dwgras, 16 years ago

Rob sent me the following information.

When I talk about Tedit I mean indeed the Tedit which comes with eCS or Warp. T2 is (was) an IBM 'Internal Use Only' tool developed by Tim Baldwin at IBM UK (under the EWS flag, Employee Written Software). Tedit is probably a subset of T2: the help files are identical apart from some minor details. T2 is only about 1000 bytes larger than Tedit (maybe the customizability).

You can find it at http://os2site.com/sw/ews

Look for tinyed220b.

in reply to:  2 comment:3 by KO Myung-Hun, 16 years ago

Hi/2.

I'm sorry I cannot reproduce this.

We need some more detailed procedures to reproduce this.

KO Myung-Hun

Replying to dwgras:

Rob sent me the following information.

When I talk about Tedit I mean indeed the Tedit which comes with eCS or Warp. T2 is (was) an IBM 'Internal Use Only' tool developed by Tim Baldwin at IBM UK (under the EWS flag, Employee Written Software). Tedit is probably a subset of T2: the help files are identical apart from some minor details. T2 is only about 1000 bytes larger than Tedit (maybe the customizability).

You can find it at http://os2site.com/sw/ews

Look for tinyed220b.

comment:4 by KO Myung-Hun, 16 years ago

Cc: KO Myung-Hun added

comment:5 by KO Myung-Hun, 16 years ago

Resolution: invalid
Status: newclosed

comment:6 by ataylor, 13 years ago

I've seen this, or a variation of it, for a long time. It only occurs when you try to execute a REXX script on a FAT32 drive and you do NOT have EAs enabled in the FAT32 driver. The REXX script will often end up replaced by a 0-byte file, and the actual REXX code may show up later as a "recovered" file following the next CHKDSK.

I assume this is caused by OS/2 REXX trying to tokenize the script into the EAs and failing due to the lack of support.

(I have seen this with classic REXX, I haven't tried with OREXX.)

Note: See TracTickets for help on using tickets.