On Jan 11, 2008 3:51 PM, Ted Herman <[EMAIL PROTECTED]> wrote:
>    I'd like to know the correct interpretation of TEP 103 (which came up
> because of a "bug" in my application).  TEP 103 says that
> LogWrite.append writes records, which might have a maximum length of 255
> bytes, depending on the platform/implementation.  LogRead.read can read
> a stream of bytes at a position not aligned to a record, and also the
> number of bytes requested could apparently be different from an existing
> record length.  The type on LogRead.read's length parameter is
> abstractly defined as storage_len_t, which is currently typedef'ed to
> uint32_t.  OK so far.
>
>    Now, when I try something like LogRead.read(&buf,256), the response
> discovered in the LogRead.readDone event is that zero bytes are always
> returned.  This is logical, because the storage_len_t parameter is saved
> to a uint8_t in Stm25pLogP.nc, which I suppose to be processing the request.
>
>    So my question is, should I suppose that TEP 103 implicitly bounds
> the maximum read length to be the same as the maximum record length?

I think I would argue that it's a bug in the ST M25P implementation (a
cursory look through the AT45DB implementation suggests that it
handles this correctly...)

David Gay
_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to