Gawk Bug

[Richard Patchet]

Invoking gawk as follows:

 

    gawk --re-interval -f \dos\gawk-bug.awk 4010_segments.txt > gawk_bug.txt

 

produces repeatable errors.

 

First instance:

 

    <xs:complexType name="sxC2">
    <xs:sequence>
    00 104 110|00 1|4 |10
    <xs:element name="C200 1" type="dx004 " minOccurs="0" otl:name=""/>

Second instance:

 

    <xs:complexType name="sxE30">
    <xs:sequence>
    16 Code List Reference AN 1|16 C|de List|Reference 
    <xs:element name="E3016 C" type="dxList" minOccurs="0" otl:name=""/>

Third instance:

 

    <xs:complexType name="sxSV1">
    <xs:sequence>
    10 1340 Multiple Procedu|10 1340 Mult|ple Pro|edu
    <xs:element name="SV110 1340 Mult" type="dx Pro" minOccurs="0" otl:name=""/>

Reviewing the gawk code, the program produces XML (those beginning with "<") with "<xs:element" lines prededed by printing $0, $1, $2, and $3 separated by vertical bars.

 

Reviewing the source file, 4010_segments.txt, will show that the errant lines actually appear some 8-10 lines _further_ into the file than the lines printed immediately thereafter. Very anomalous behavior.

 

 

Richard Patchet

A4 Networks

101 W McDermott Dr

Allen, TX 75013

214.547.8020

www.a4networks.com

 

gawk_bug.ZIP
Description: Binary data