Re: chunked output
[email protected] Tue, 19 Aug 2003 20:10:40 -0400
| Newsgroups | gmane.comp.programming.tools.fit.devel |
|---|---|
| Message-ID | <[email protected]> |
----- Original Message ----- From: "Ward Cunningham" <[email protected]> To: "FIT Developers" <[email protected]> Sent: Tuesday, August 19, 2003 4:16 PM Subject: Re: [Fit-dev] chunked output > > On Tuesday, August 19, 2003, at 12:52 PM, Micah Martin wrote: > > > FitFilter could be a much better solution to running suites. Currently > > FitFilter works like this: > > 1. html is recieved from stdin > > 2. read from stdin until EOF > > 3. process > > 4. write results to stdout > > 5. exit > > Bjorn Freeman-Benson and I cooked up something along these lines for > the Eclipse Fitster plugin. Using Java serialization we would send > tables to a custom runner and read tables and counts back. Our approach > to progress indication was to say that the custom runner could send > count objects at its convenience and we would update the display. Sending counts back sounds like an excellent idea. > We also communicated over a socket, not standard in/out. I was > following Bjorn's lead with this decision and don't know what the > tradeoffs are. I presume that most firewalls know to stay out of the way of sockets on localhost, but I'm not all that certain. I don't think that sockets bring anything to the party that you can't do with stdin/stdout, and my impression is that sockets would be more complex. > > (Is FitFilter better named FilterRunner? Would it be the same as > FileRunner if FileRunner took "-" to indicate standard in/out as file > names?) The word filter doesn't do anything for me, although I like xxxRunner. Maybe ProcessRunner? I'm not sure I want the two different processes in the same program, though. It strikes me that the entire process of passing a callback to Fixture so it can pass back a table at a time simply complicates the business of file in, file out; I'd rather have FileRunner as a separate, much simpler program so there's one program with the core, and another program with the embellishments. On the other hand, I certainly don't feel strongly enough about it to argue the point with any vigor. John Roth > > -- > Ward Cunningham > 503-245-5633 v mailto:[email protected] > 503-246-5587 f http://c2.com > > > _______________________________________________ > Fit-dev mailing list > [email protected] > http://lists.freepan.org/mailman/listinfo.cgi/fit-dev