home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Collection of Hack-Phreak Scene Programs
/
cleanhpvac.zip
/
cleanhpvac
/
TIERRA40.ZIP
/
ALMOND
/
ALCOMM
/
SRC
/
TODO
< prev
next >
Wrap
Text File
|
1992-07-06
|
2KB
|
39 lines
need some way of breaking out of the waiting state when waiting for a reply
if for some reason the reply doesn't happen... is there any possibility of
this at all?? I don't *think* so, but verify.
|
\---> this should have been fixed by requiring the query handler to
call MGenReply. if it doesn't, the mlayer issues a bogus (error)
reply which satisfies the blocking end, so there should never be
a problem.
MAY WANT TO PROVIDE a function to do messages, to do queries, etc... so that
the user doesn't have to use MGenRequest and MGenRequestWithReply and get
himself into trouble! Think about it - it's just a few lines!
use MrERROR to signal out-of-range messages, queries, etc.
when enabling a new dataflow, should we automatically call the Initialise
procedure (if one is registered) from within MDataflowControl(), or should
we enter a request in the queue and deal with it separately?
|
\---> right now, it's handled inside MDataflowControl().
look over AL and M layers wherever handles are used; make sure there's enough
verification for valid handles... don't want to crash something.
wherever status return values are used, make sure variable holding them
has value either ALtStatus or MtStatus (depending on package)
=============================================================================
tierra-end TODO:
replace instances of ALGripe in tierra with whatever error reporting
scheme is to be used.
provide a mechanism for rerouting errors (but still through ALGripe) wherever
user code requires - alerror.c will have to change slightly. maybe a
callback-type approach would work?