Subject:Build 287g A step backwoards Author:(Deleted User) Date:07 Feb 2005 06:11 PM Originally Posted: 07 Feb 2005 06:06 PM
Build 287g is the worst I have ever seen from you guys for stability.<br>
<br>
It has blownup up at least three time today alone.<br>
<br>
Case 1.<br>
Process a large file. Works OK.<br>
Apply same work to same file a second time, Boom.<br>
Feels like memory leak.<br>
<br>
Case 2.<br>
Process two large files A and B. A is source document B <br>
loaded from document() Works.<br>
B as source A loaded from Document Boom. Related to Document loaded via
document() when it is invalid.
Subject:Build 287g A step backwoards Author:(Deleted User) Date:08 Feb 2005 10:13 AM
It has happened both in xquery and XSLT.
Case 1.
The source document is 8M. Repeated XQUERY ends up in a crash.
Source is a W3C Schema representing all the structures of a Mortgage
Loan Origination system ( or at least the ones that get exported to
other organizations). It only takes three of four searches to crash.
Case 2. XSLT of the same file as Case 1 to produce, say a report or some
SQL. thrid or fourth time causes a crash.
Case 3. XSLT using one large source file (12 M) and another large
document via
<xsl:variable name="schema" select="document('../GFX_LOAN_included.xsd')"/>
(This is our 8 M friend from case 1 and 2). This may have nothing to
do with size. I discover I can make a crash repeatedly if the document
accessed via the document() is invlaid.
Cases 1 and 2 fee like a memory leak to me. Case 3 is an unhandled
exception of some sort.
Does struzzo.exe make a log file anyplace I can send you?