Need some help with a query (uniq -c)

Hello!

I have a table (think of it as a table of log messages)

time |  message
-----------------------
1      |   a
2      |   b
3      |   b
4      |  b
5      |  a

the three 'b' are the same message, so I would like to write a query
that would give me a result that is similar to what the unix command
"uniq -c" would give:

first |  message | last | count
--------------------------------------
1     |     a              |   1   |     1
2     |     b              |   4   |     3     <--- here it squeezes
similar consecutive messages into a single row
5     |     a              |   5   |     1

How do I write such a command?

I could of course write a plpgsql function that loops but that is not
as interesting as finding out if this can be done in a single simple
command.

Perhaps it would be best to regularly delete neighbouring similar
rows and keeping a "count" value would reduce the number of rows and
make it more efficient if the query would be run many times and the
number of duplicate messages would be large.