> per the configuration ) and saves into .tmp files.
> If there is no connection active (or at the end of
> the crawl), it converts the .tmp pages into   normal
> .html pages with some additional header information.
> My requirement is that the crawl should not convert
> the .tmp pages into .html pages.

.tmp are created in the 3.4x series internal processing. The .tmp file is
binary data.  I don't understand why you wouldn't want that. 3.32 does not
create them (it kept the data in memory and had to recreate it on restart.)

The .html contains an additional header (options -> Browser ID -> HTML footer)
and internal links on the page converted to relative filenames (options ->
Expert -> rewrite links = Relative/Absolute.)