MJP@vaxb.roe.ac.uk (08/29/89)
Testing a parallel interface today, I managed to type in some incorrect
code which, on first appearance, shouldn't work BUT DID! The details of
the interface are irrelevant, but include a C011. Spot the deliberate
mistake in the following code precis:
PROC interface()
VAL link2.in IS 6 :
CHAN OF [2048]INT16 from.c011 :
PLACE from.c011 AT link2.in :
[2048]INT16 data :
BOOL running :
SEQ
running := TRUE
WHILE running
from.c011 ! data
:
Prodding around in the interface with an oscilloscope, I found that, lo
and behold, the c011 was getting an acknowledgement for the data I was
feeding into it, despite the fact that I didn't have a Transputer
reading data at the other end of the link. Yes, it should be
"from.c011 ? data" in there. Has anyone else come across this rather
bizarre behaviour? Have I made a new discovery? Should I publish? Am
I going to be shot down in flames? All and any thoughts on the subject
would be much appreciated.
Cheers, ***************************************************
Magnus Paterson, * janet - mjp@uk.ac.roe.vaxb *
Royal Observatory, * cbsmail - cbs%uk.ac.roe.vaxb::mjp *
Edinburgh * earn - mjp@vaxb.roe.ac.uk *
Scotland * arpa - mjp%vaxb.roe.ac.uk@nss.cs.ucl.ac.uk *
* arpa - mjp%vaxb.roe.ac.uk@ukacrl.bitnet *
If all else fails, * span - 19463::revad::mjp *
ph. 031-668-8406 * starlink - revad::mjp *
fax 031-668-8264 * psimail - psi%000007002002::mjp *
* janet - mjp@uk.ac.roe.star *
***************************************************