[m-rev.] for review: deep profiler redesign

[Zoltan Somogyi]

On 11-Nov-2002, Michael Day <mikeday at yeslogic.com> wrote:
> Or what about using FastCGI? That supports persistent CGI applications, 
> and works with Apache.

It didn't exist when the deep profiler was originally designed. We looked
for a standard solution of the initialization time problem, at that time
in vain.

FastCGI still has two problems.

- The only implementation I know of is in Perl. Integrating that with Mercury
  code probably won't be easy. Integrating with C code would be a lot easier.
  Anyone know of a C FastCGI library?

- It assumes that a given service needs only one server. The natural design
  for mdprof wants one server per profiling data file, which is significantly
  different. One could of course build one server for all files, but then
  you would have to solve the problem of discarding the state of one data file
  that hasn't been accessed in a long time while keeping the state of other,
  recently accessed data files. Not conceptually difficult, but annoying.

Until both these problems are solved, there is no point in switching.

Zoltan.
--------------------------------------------------------------------------
mercury-reviews mailing list
post:  mercury-reviews at cs.mu.oz.au
administrative address: owner-mercury-reviews at cs.mu.oz.au
unsubscribe: Address: mercury-reviews-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-reviews-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------