vanishing .xac and .log files

[hendrik at topoi.pooq.com]

On Fri, Aug 17, 2007 at 09:47:32AM -0400, Josh Sled wrote:
> hendrik at topoi.pooq.com writes:
> > Now I need a tool that will tell me all the differences between 
> > reconciled splits between those two .xac files.  Anyone know of one?
> > output from diff itself (after gunzipping, of course) contains too much 
> > irrelevant junk, and consists of uncontexted fragments of 
> > transactions.
> 
> (I've seen you on another couple of mailing lists, so I know you're a good
> bit technical...)
> 
> I'd extract the relevant Splits via XSLT, though not to XML .. just
> line-oriented plain text.  Probably something like:
> 
>     Split1GUID Split1Date Split1Desc Split1Acct Split1Amt
>     Split2GUID Split2Date Split2Desc Split2Acct Split2Amt
> 
> Then, /usr/bin/sort it so the splits are ordered by guid.  Then do the
> line-oriented diff.
> 
> Something like...
> 
>     <xsl:output method="text"/>
>     <xsl:template match="//trn:split[split:reconciled-state='y']">
>       <xsl:value-of select="./split:id/text()"/><xsl:text> </xsl:text>
>       <xsl:value-of select="./split:memo/text()"/><xsl:text> </xsl:text>
>       <!-- ... -->
>     </xsl:template>
> 
> Should do the trick, I think.

Thanks, I should really learn xsl/xslt sometime.  

But by the time I got your reply, I had hacked together another 
way.  Sleeping on a problem does wonders!

I took an old program I had debugged into existence a few years ago that 
prints all the transactions sorted by account tree in a somewhat 
readable form -- just plain text, one per line.

I modified it so it would not filter output by year, but would instead 
pay attention only to reconciled splits.

After maiking these changes, it didn't work, so I inserted some trace 
statemente to see what it was doing instead.  I never got it to work 
(programs debugged into existence are not of the best), but I realised 
that the trace output I was getting was what I really needed!  So I took 
the trace output files from processing the two gunzipped .xac files, 
sorted them, and diffed the results.

Eureka!

-- hendrik