TESTING

how i test

1) every now and then i run the testall.sh script over the incuded doc
files and a few selected trouble makers that ive received.

2) i have a set of supported features documents that display what
mswordview should be able to do, i regularly check these to see if theres
any feature rot.

3) occasionally i repeat the above test while using purify and identify
out of bounds and mem leaks.

4) once a week i run the huge test of (1) over the 300 megs of uploaded
documents that i have received on poor old skynet.csn.ul.ie, many of these
are not word 8 docs as people often upload rtf & word 6/7 & excel docs as
well, but its a good brute force test for crashes, but no good to see if
the conversion makes sense.

5) i scan through the comments tha uploaders make of their files to identify 
trouble docs and visually scan the output, making changes if necessary.

some tools

i still use laola's lls to help me find errors, its a great piece of kit
to have. i also gives me other ole code to compare against the one thats
included in mswordview, if both of them cant decode a file, then theres 
pretty good evidence that somethings buggered in the ole tables, when word
hangs on them as well, thats probably becomes definite.

never having gotten the hang of hexdump i use the little jhex script that
was posted in the linux-gazette at one stage as my hex reader. And theres
nothing like doing a diff on the text output of a hexdump to find what
changes between two versions of the same file, but with a different style
chosen.

purify is also the biz for tracking down mem leaks and uninitilized memory
elements.