xmlgraphics-fop-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeremias Maerki <dev.jerem...@greenmail.ch>
Subject Re: DO NOT REPLY [Bug 35937] - [PATCH] Support for FormattingResults ported from 0.20.5 to 1.0
Date Tue, 02 Aug 2005 08:22:40 GMT

On 31.07.2005 11:55:00 Simon Pepping wrote:
> On Sat, Jul 30, 2005 at 03:35:07PM +0200, bugzilla@apache.org wrote:
> > ------- Additional Comments From jeremias@apache.org  2005-07-30 15:35 -------
> > Patch applied with some little modifications. Thanks a lot, Manuel!
> > 
> > http://svn.apache.org/viewcvs?rev=226511&view=rev
> I changed this to a notify mechanism, similarly to that already used
> in PSLM.activateLayout to notify the Root FO of the same. This gives a
> more direct access to the formatting results.


> It would be a nice solution to let embedding apps register themselves
> as a FormattingResultsListener in the user agent. apps.Fop could do
> the same to present the user with the formatting results if he
> requires verbose output.

I agree with the actual concept but I actually want something more
general in the long-term. There are various aspects that I'd like to
provide to embedders:

- Number of pages over all and Number of pages for every page-sequence 
- Callbacks on layout problems such as overflows, missing resources etc. (on
which the listener can veto and therefore cause an exception.) [1]
- Possibly additional hooks to insert images (via BufferedImage or
providing a Graphics2D). We already have a stream insertion (or rather
custom URI resolution) hook in FOUserAgent. Users have repeatedly asked
for easy ways to do stuff like that.
- there may be more.

And! We need to split debugging output from verbose user information!

[1] Could be very interesting particularly for editor creators and
people who want to be informed about every irrgularity in server-side
mass document production. Every kind of problem should have an ID
(machine-interpretation) and the error which go onto the log should also
be i18n-ized (human-interpretation).

Jeremias Maerki

View raw message