##PRB
CLI "National Library of the Czech Republic"
MRA Iva.Pribramska@nkp.cz
DAT 19990309
TY QUERY
COD NKP/99002
REF Iva Pribramska
UR Normal
OS Digital Unix
ALE 33_03
SH Problems of UTIL 64 for ordera
PB UTIL 64 for ordera has crashed with a message:
A line of the input file contains more than 20480 characters.
for order1.scr file. This resulted in:

I/O error : file 'TP1'
error code: 3/5 (ANS85), pc=0, call=1, seg=0
13 File not found

Looking at order1.scr in vi editor we do not see anything abnormal.
The message about the size of a line of the output file comes from a command from proc64_b:
$aleph_sort -o $dd_TP2 +0.0 -d $dd_TP1.
When we try to run this command either in the form:
$aleph_sort -o $dd_TP2 +0.0 $dd_TP1
or in the form
$aleph_sort -o $dd_TP2 -d $dd_TP1, there is no message like this produced and file seems to be sorted all right (but accented characters).

It seems that there is a relation to locale parameters of the Unix system:
if we use setenv LANG en_US.ISO8859-2 instead of setenv LANG en_US.ISO8859-1 (which we were using until now), the sort command in the original form does not crash but, anyway, the data are not sorted correctly from the point of view of accented characters. But this value of LANG parameter is not defined properly and the system uses some default value (what causes other troubles, e.g. in vi editor).