Re: FOP 0.95b observations
On 09.04.2008 02:32:32 Christopher R. Maden wrote: 1) FOP now complains about empty fo:table-cells. This is strictly correct, I suppose, but annoying. More troubling, it fails with an exception and a stack trace when it happens, rather than a simpler error message. If you want to feed FOP invalid XSL-FO, you can enable relaxed validation (either on FopFactory or through the configuration) as described on the website. The thing with the stacktrace was a temporary measure as some exceptions were swallowed due to suboptimal exception handling. This will be improved with the next release as we add an processing feedback system that allows more fine-grained event handling. 2) Line-wrapping doesnt seem to happen automatically in table cells: Apr 8, 2008 4:09:58 PM org.apache.fop.layoutmgr.inline.LineLayoutManager $LineBreakingAlgorithm updateData2 WARNING: Line 1 of a paragraph overflows the available area by 11431mpt. (fo:block, location: 2/56170) is the message, now. These same paragraphs used to wrap in 0.94 and in 0.20.5. I assume you have a keep-together=always somewhere upstream of the table-cell. Please note that some support for inline keeps has been added. If you do keep-together=always you're indirectly setting keep-together.within-line=always which forbids FOP to break the line. Just change that to keep-together.within-column=always. 3) While this doesnt produce the strange page-breaking behavior seen in 0.94 (thanks!) it still does not appear to honor keep-together=always on table-body. By comparison 0.20.5 kept table-bodies together by default. My document consists of a series of these: fo:table-body border=solid 1pt black keep-together=always Dummy output is available on request, if it would be helpful. Confirmed. There is indeed a bug there. But keep in mind, it should be keep-together.within-column=always! Thanks for the feedback! Thanks, Chris -- Chris Maden, text nerd URL: http://crism.maden.org/ Weve replaced American airports with incompetent mini-police- states. Lets see if they notice! GnuPG Fingerprint: C6E4 E2A9 C9F8 71AC 9724 CAA3 19F8 6677 0077 C319 Jeremias Maerki - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: FOP 0.95b observations
On Wed, 2008-04-09 at 09:39 +0200, Jeremias Maerki wrote: The thing with the stacktrace was a temporary measure as some exceptions were swallowed due to suboptimal exception handling. This will be improved with the next release as we add an processing feedback system that allows more fine-grained event handling. Good to know; thanks. I assume you have a keep-together=always somewhere upstream of the table-cell. Please note that some support for inline keeps has been added. If you do keep-together=always you're indirectly setting keep-together.within-line=always which forbids FOP to break the line. Just change that to keep-together.within-column=always. Ah! I helped write XSL, but it was now many years ago, and little things like that occasionally jump up and surprise me, especially when the same stylesheet has been working acceptably for years. (-: Confirmed. There is indeed a bug there. But keep in mind, it should be keep-together.within-column=always! Thanks for the feedback! Gotcha. An anonymized test, if it helps, is at URL: http://crism.maden.org/table-test.fo and URL: http://crism.maden.org/table-test.pdf . Thanks, Chris -- Chris Maden, text nerd URL: http://crism.maden.org/ We’ve replaced American airports with incompetent mini-police- states. Let’s see if they notice! GnuPG Fingerprint: C6E4 E2A9 C9F8 71AC 9724 CAA3 19F8 6677 0077 C319 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: FOP 0.95b observations
On 09.04.2008 10:19:09 Christopher R. Maden wrote: snip/ Confirmed. There is indeed a bug there. But keep in mind, it should be keep-together.within-column=always! Thanks for the feedback! Gotcha. Actually, I have to correct myself. I just noticed that keep-together does not apply to fo:table.body: http://www.w3.org/TR/xsl11/#fo_table-body It is an inherited property so a keep is imposed on all table-rows in the body but not the whole body is kept together. So there's actually no bug. So you'll have to work with keep-with-next on all table-rows in the body to keep them together, or you specify the keep on the table itself. An anonymized test, if it helps, is at URL: http://crism.maden.org/table-test.fo and URL: http://crism.maden.org/table-test.pdf . Thanks, Chris -- Chris Maden, text nerd URL: http://crism.maden.org/ Weve replaced American airports with incompetent mini-police- states. Lets see if they notice! GnuPG Fingerprint: C6E4 E2A9 C9F8 71AC 9724 CAA3 19F8 6677 0077 C319 Jeremias Maerki - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: FOP 0.95b observations
On Wed, 2008-04-09 at 11:17 +0200, Jeremias Maerki wrote: Actually, I have to correct myself. I just noticed that keep-together does not apply to fo:table.body: http://www.w3.org/TR/xsl11/#fo_table-body Wow... thanks for catching that. It’s working perfectly now. I can’t explain or understand the rationale for that decision, though... this is one of the things that using multiple table-body elements is great for. Not being able to keep them together doesn’t gain anything. Thanks again, Chris -- Chris Maden, text nerd URL: http://crism.maden.org/ We’ve replaced American airports with incompetent mini-police- states. Let’s see if they notice! GnuPG Fingerprint: C6E4 E2A9 C9F8 71AC 9724 CAA3 19F8 6677 0077 C319 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
FOP 0.95b observations
1) FOP now complains about empty fo:table-cells. This is strictly correct, I suppose, but annoying. More troubling, it fails with an exception and a stack trace when it happens, rather than a simpler error message. 2) Line-wrapping doesn’t seem to happen automatically in table cells: Apr 8, 2008 4:09:58 PM org.apache.fop.layoutmgr.inline.LineLayoutManager $LineBreakingAlgorithm updateData2 WARNING: Line 1 of a paragraph overflows the available area by 11431mpt. (fo:block, location: 2/56170) is the message, now. These same paragraphs used to wrap in 0.94 and in 0.20.5. 3) While this doesn’t produce the strange page-breaking behavior seen in 0.94 (thanks!) it still does not appear to honor keep-together=always on table-body. By comparison 0.20.5 kept table-bodies together by default. My document consists of a series of these: fo:table-body border=solid 1pt black keep-together=always Dummy output is available on request, if it would be helpful. Thanks, Chris -- Chris Maden, text nerd URL: http://crism.maden.org/ We’ve replaced American airports with incompetent mini-police- states. Let’s see if they notice! GnuPG Fingerprint: C6E4 E2A9 C9F8 71AC 9724 CAA3 19F8 6677 0077 C319 signature.asc Description: This is a digitally signed message part
