[m-dev.] deep profiling issues

[Fergus Henderson]

On 26-Jun-2001, Zoltan Somogyi <zs at cs.mu.OZ.AU> wrote:
> On 26-Jun-2001, Fergus Henderson <fjh at cs.mu.OZ.AU> wrote:
> > 	(2) I think the binary distribution doesn't install the deep profiling
> > 	    scripts or executables
> 
> Right. It is not intended to, yet. It needs to wait on two other things.
> First, for the deep profiler's functionality to stabilize, so people who
> download releases of the day don't get used to functionality that will change.

It's already in the source releases of the day.
I think it's confusing if the binary release doesn't match the source release.

> Second, I need to find out *how* to find out whether a machine has a web server
> installed and enabled, since without a web server, there is no point in
> installing the deep profiler. I have not yet found any documentation relevant
> to the latter problem.

The simplest way that I can think of would be to try connecting to
it, e.g. by downloading the root page.  You could write a small test
program to connect to the server's http port and send the appropriate
HTTP request.

But for now I think it would be fine to install it always, or only if a
certain configure option was enabled.  Currently the *.profdeep grade of
the libraries is getting installed on any system with <unistd.h>, and so
you might as well install the other stuff needed for deep profiling too
(e.g. the mdprof script and the mdprof_server and mdprof_cgi executables.)

Getting it to install correctly is more important than getting it to
autoconfigure whether or not a web server is installed.  After all,
the user might not realize that the deep profiler requires a web server,
and so might install Mercury before installing a web server.  When they
discover that it needs a web server, they'll have to install one, but
it would be a pain if they then have to reinstall Mercury.

> > 	(3) on hg, using egcs-1.1, building the library in the profdeep
> > 	    grade results in a gcc error:
> 
> Do you know what part of the file it is complaining about?

The definition of the perform_sampling procedure (all modes).
I spent some time trying to isolate it further but without success.

> How old is egcs-1.1?

egcs-1.1 is a little over two years old.

> Can we say that it is not supported for deep profiling?

I think that would be OK.

P.S. There's one more that I forgot about:

	(4) the *.profdeep grade is being installed, but it is
	    not being tested by our nightly tests script

-- 
Fergus Henderson <fjh at cs.mu.oz.au>  |  "I have always known that the pursuit
The University of Melbourne         |  of excellence is a lethal habit"
WWW: <http://www.cs.mu.oz.au/~fjh>  |     -- the last words of T. S. Garp.
--------------------------------------------------------------------------
mercury-developers mailing list
Post messages to:       mercury-developers at cs.mu.oz.au
Administrative Queries: owner-mercury-developers at cs.mu.oz.au
Subscriptions:          mercury-developers-request at cs.mu.oz.au
--------------------------------------------------------------------------