[Demexp-dev] Re: get_timestamps() and get_timestamps_decoded()

[David MENTRE]

Augustin,

Augustin <address@hidden> writes:

> But it seems that I eventually will HAVE TO implement get_timestamps() on the 
> web client side, won't I?

Yes but the timeframe is so long that there is no hurry, see below.

> There are two aspects to the problem: the compression, and the integers.

> About your gremlins ("32 bits unsigned big endian integers, one
> integer"), we 

In fact, I was wrong, the 32 bits big endian integers are *signed*, see
start of Chapter 8 (Timestamp) in demexp-book (I'm not reading correctly
my own doc :-).

> may have more of a problem:
> http://php.net/manual/en/language.types.integer.php
> ""The size of an integer is platform-dependent, although a maximum value of 
> about two billion is the usual value (that's 32 bits signed). PHP does not 
> support unsigned integers.""

So there is no issue related to integer sign. 

In fact, the only issue would be a performance issue (loading about
1,000,000 question timestamps would be too long) but we can fix it at
that time (or a little earlier ;-).

> Do you think that it would take too much time to deal with this now, or you 
> can help me make an attempt at using the right protocol straight away.
>
> If it is too troublesome, or if it would take too much time to do this, given 
> our time constraints (those I have just announced on the fr list!!), maybe I 
> should go forward and use the  get_timestamps_decoded() you kindly provided, 
> and come back to this problem only much later...

Yes, use get_timestamps_decoded(). And put a strong warning in your
code, stating about the potential perfomance issue. If you maintain a
bug list for your code, add a TODO item for this issue.


Best wishes,
d.
-- 
GPG/PGP key: A3AD7A2A David MENTRE <address@hidden>
 5996 CC46 4612 9CA4 3562  D7AC 6C67 9E96 A3AD 7A2A