In article <47909b514e.harriet@xxxxxxxxxx>, Harriet Bazley <lists@xxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote: > On 4 Aug 2006 as I do recall, Frank de Bruijn wrote: [snip] > After making the changes suggested below > ( L%+=sb% : sb%=L% ) > I'm now seeing a couple of messages with the extra blank line you > mention. Unfortunate in practice, since it pushes the relevant text > below the testing range of the rule Hmmmmm... Maybe an option to ignore blank lines *inside* the body text would be useful? > - but it still wasn't getting called for those messages anyway. :-( Strange. > I'm also seeing a different couple (out of about eight containing the > offending text) of messages where the 'twoticks' body text Rule *is* now > logged as being called... once only, whereas AntiSpam is configured to > test the first three lines. These messages do not seem to contain the > extra blank line. > Still something wrong here, I think. I must admit I didn't look any further after I discovered the bit of code that was eating blank lines [1]. I'll check. Unfortunately, I won't be near a RISC OS computer from later today until at least August 20. [snip] > > Now you've lost me. I thought you wrote trialling (i.e. using the Trial > > window) of these messages worked properly while testing (checking before > > downloading - or not) didn't? > Terminology confusion - I assumed 'testing' referred to simulating > download via the Trial window, as opposed to actually deleting > anything! :-) Understandable. Maybe it's time to change the 'testing' state to 'checking' (I did think about this when I started working on AntiSpam but decided against it at the time). Regards, Frank [1] Because that *was* the root cause: no blank line, no switch to checking body lines. However, with two adjacent blank lines, only the first was swallowed. So in those messages, the switch *was* made.