CLIM mail archive
Output recording and presentations
> Date: Mon, 11 Oct 1993 13:28 -0400
> From: Scott McKay <SWM@stony-brook.scrc.symbolics.com>
> Date: Mon, 11 Oct 1993 12:16 EDT
> From: Brian Anderson 234-0878 <firstname.lastname@example.org>
> I have noticed that even though output recording is shutdown (using
> clim:with-output-recording-options to turn it off) calls to
> clim:with-output-as-presentation (and also
> clim:with-new-output-record) seem to still take affect. Is this the
> way its supposed to be? Or do I have to check with
> clim:stream-recording-p? Isn't a presentation considered an output
> record and therefore should be affected by the status of output
> recording on the stream?
> WITH-OUTPUT-AS-PRESENTATION turns output recording back on for you,
> but I can't recall the reason why it does this. I think it's probably
> a bug.
Note that on page 23-17 of the spec, it says that the presentation is
*not* inserted in the output history if output recording is off. I
was counting on this to use the idiom of creating a bunch of output
records with output recording and drawing off, insert them into the
appropriate place in the output history, and replay them myself. This
almost worked except for the fact that the presentation output record
was inserted at the top level of the history (when I ran things to
create an output record) and in another place in the history when I
inserted it there myself. I suppose that this is the cause of some
wierd behavior that I have experienced with ghost presentation
highlighting appearing after I have tried to hand manipulate the
output record tree. I have resorted to setting
clim:stream-current-output-record appropriately to "open" the
appropriate record in the history and then just doing normal output.
But, this prevents any analysis and manipulation of the output records
before they are actually inserted into the output history and/or drawn.
> Note that higher level formatted output frobs, like FORMATTING-TABLE,
> do turn output recording on. I believe that this is probably correct,
> since you can't actually format the output correctly otherwise. But I
> wonder if, when such things are called with output recording off, the
> resulting output record should not be added to the output history.
Hmmm. Can't you still use the idiom of turning output recording off
to prevent things from being inserted into the output *history*, while
still creating output records for you to manipulate? Or is output
recording supposed to prevent *any* output records from being created?
If so, maybe there should be a distinction between output recording
history (inserting output records in the streams output history while
still creating output records) and output recording (creating output
records as a result of doing output).
Brian H. Anderson (206) 234-0881
Boeing Commercial Airplane Co. email@example.com
P.O. Box 3707 M.S. 6A-PX
Seattle, Wa. 98124-2207
Main Index |