TMG-L Archives

Archiver > TMG > 1999-09 > 0936337011


From: Robin Lamacraft <>
Subject: Re: TMG-L: Cannot print to file - Error #3
Date: Fri, 03 Sep 1999 15:06:51 +0930


At 22:25 2/09/99 -0500, Steven Chall wrote:
>The following message was sent to Tech Support but, so far, they have not
>found the gremlin. Since so many of you are able to print reports to a
>file, I know the problem is on my system. Does anyone know what
>situation will generate this error message?
>
>Whenever I try to print *any* report to file, I get:
> Error #3
> Could not open CONV_LOG.TXT on scratch drive
>
>The file, CONV_LOG.TXT, is created each time with zero bytes in my TMGW
>directory but nothing more happens. There is always an indication that
>the desired file is being written to the proper directory, (with the grey
>progress bar but not the red one) but that window is always followed by
>the error message and nothing else exists.
>I even tried to print a simple blank Family Group Sheet to use as an
>e-mail attachment. I have about 20 MB free on my TMGW drive and 50 MB
>free on "C".
>This has happened whether I try to create an ASCII, WordPerfect or Word
>document. On the assumption that somehow I lost a required file, I
>re-installed ver4.0 Gold from the CD but even that did not help. I have
>no trouble sending any of these reports to screen or printer so I assume
>something *is* missing.

Steven,
This sounds to me like the old "file is in use" problem that has plagued
many other users. My analysis of all of these problems so far is as follows.

When you run a report in TMG there are a number of phases:
(1) you call up the CRW control windows and populate them with the
requirements of your report.

(2) on clicking the "Generate" button, the Foxpro code behind this CRW
control window set constructs a series of database queries and creates a
number of temporary databases that contain just the information required by
the report type given your selected focus.

(3) then control passes to R&R report writer to compose the formatted
content of the report. This is just like Crystal Reports (but more
powerful). This process again generates a series of temporary files - If
you render the report to the screen, then R&R stays in control and it
constructs each page of the report "on demand" - hence output to the screen
takes a different path.

(4) if the output is directed straight at a printer, it is likely that the
intermediate file is processed according to the properties of the printer
and travels through the code in another path (and uses different temporary
files) than to the screen, especially if the report could have footnotes
and endnotes.

(5) if the output is directed to a word processor, a third module comes
into play - this converts the generic intermediate file into one that
should be understood by the target wordprocessor. Here the endnote or
footnote problem arises again as each wordprocessor needs these stored and
cross-referenced diffferently.

Now to the point of all this - in this chain of events each of the
intermediate files is usually created/opened for *exclusive* by the module
that fills its contents. When that stage finishes that module should change
the status of the file (ie. unlock it and release it for use by the next
stage). At the end of this whole report generation cycle and before you can
specify another report CRW should cleanup all these intermediate files.
This is why at the end of generating a narrative report you are asked
wherther you want keep (and hence accumulate) the endnote/footnotes or
clean them out before the next report.

If, for some reason, either the change of status and release of the
exclusive use does not happen or at the end of the report generation cycle
the temporary files (usually they have fixed names) are left around as
garbage in the TMGW folder, then you may will get the problem that you are
describing.

Unfortunately, if this is what happening to you, there is no quick or clean
answer. And TMG is not the only program that suffers from this syndrome.
The most probable cause is that something has disturbed the environment and
the settings/libararies/permissions and other dependencies within your
system. That is, some installed software or some upgrade to previously
installed software has made your system unstable for this CRW chain to work.

At work, I downloaded a demo of a Lucent Technologies product. This used
the latest technology to some very fancy things. Unfortunately, to make it
run it required a whole suite of the latest Microsoft dlls, ocxs, etc to be
installed. Once this was installed the demo worked fine. However, for other
reasons (a project) our machines had Microsoft Visual Studio 97 installed.
But now many of the tools folded, could not open databases, etc. There is
an incompatibility between versions of dlls and disaster ensued. I now
having to upgrade my Visual Studio to the latest version that uses the same
dlls, ocxs, etc as that demo.

I am searching the Internet for tools that will maintain a database of
which instal installed which version of what. Comparision of system maps
between good and freezing systems is the only way getting the evidence. And
don't believe it will be better with a 32-bit TMG when it comes along. One
set of problems will be replaced by another. And W2000 just adds another
kind of installation hassle for all of us that have programs of many
different vintages on our machines. It ain't going to get better - so we
had better find better diagnostic tools.

Finally, download the 30-day/30-try demo of filemon v4.2 from
http://www.sysinternals.com/ - it will run on 95/98/NT
Boot your machine, stop everything that you can (no internet or virus scan).
Set filemon to capture file system requests just before you click the
"Generate" button and stop the capture once you have the error. You will
get several thousand lines of output which you should save to a text file
so that can look and search easily. You may want to send it to Tech support
or myself if you don't understand what it tells you.

I hope that this helps,
Robin

- ---------------------------------------------------------------------------
TMG-L The Internet Mailing List for The Master Genealogist
To unsubscribe: Send an e-mail to with 'UNSUBSCRIBE TMG-L'
or... if you get the digest version: 'UNSUBSCRIBE TMG-L-DIGEST'
- ---------------------------------------------------------------------------

This thread: