Re: a form with unknown input names?

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 454-5900

Richard Michael <rmi...@fi...> writes:

> Anything else useful not yet documented?  I've written a bunch of stuff
> to handle vars as alists..  (and I noticed they seemed to be handled
> that way behind the scenes anyways), now I wonder if some of that is
> exposed but undocumented.

Yes.  The stuff in brl-context has not been documented because I had
vague thoughts of revising it.  However, looking at what's there it
seems unlikely that backward compatibility would have to be broken, so
feel free to look in gnu/brl/progfun.scm for the brl-context- stuff.

(Next paragraph is why I'm cc'ing Per)

In the next few months I plan to redo a lot of things to fit into Per's
framework for BRL now in the Kawa distribution.  One of those will be to
make brl-context be a subclass of ServletCallContext.  (A total rewrite
of brlsv.java will likely also be necessary).  I'll do at least one more
release using the current framework, to get stuff out into an official
release that's been waiting in CVS.

Now, more about "a bunch of stuff to handle vars as alists".  There's
some more undocumented stuff that's quite useful.  After BRL has better
support for multiple similar-named inputs (e.g. foo1, foo2, etc.), there
will be a cleaner way to do this, but code already in there makes it
easy to do an update of *only* changed fields, reducing the chances of
simultaneous updates clobbering each other.

Here's some actual code that takes input from a signup sheet for a
Toastmasters club I'm involved with.  The syntax brl-get-update and
procedure sql-partial-update are the pertinent undocumented features.

For each input, say "grammarian", there's a hidden input "ogrammarian"
that contains the old value.  The brl-get-update syntax returns an alist
just for inputs that have been changed.  You can use it now if you don't
mind changing your code when I make an incompatible change (maybe even
in the next release), to go with "grammarian_" instead of "ogrammarian",
to be more consistent with how a new way of doing this will work.

[
 (inputs m tmod_name otmod_name othought general_evaluator ogeneral_evaluator
      evaluator1 evaluator2 speaker1 speaker2 timer otimer topicmaster
      oevaluator1 tmod_email oevaluator2 ospeaker1 ospeaker2 thought
      otmod_email otopicmaster grammarian ogrammarian other_business
      oother_business)

 ; Get a JDBC statement object that will be closed automatically
 (define st (brlewis-st brl-context))

 (define update-info
   (brl-get-update brl-context
    (ss tmod_name)
    (ss tmod_email)
    (ss topicmaster)
    (ss grammarian)
    (ss thought)
    (ss speaker1)
    (ss evaluator1)
    (ss speaker2)
    (ss evaluator2)
    (ss timer)
    (ss other_business)
    (ss general_evaluator)))
]
    <title>Sign-Up Sheet [m]</title>
  </head>

    <body>
      <h1>Sign-Up Sheet [m]</h1>

<p>Updated info for
[(sql-partial-update st "oa_meeting" update-info
      "where date_trunc('day', mtg_time) = "
      (ss m))
]
<a href="./">Oratory Authority meeting</a>.</p>

-- 
<brlewis@[(if (brl-related? message)    ; Bruce R. Lewis
              "users.sourceforge.net"   ; http://brl.sourceforge.net/
              "alum.mit.edu")]>