#110 Poetry tags cause pagenums to be seriously mispositioned

[Charlie Howard]

When a poem extends to the bottom of a page or crosses a page boundary, Guiguts is moving the page markers to the beginning of the poem. Sometimes, this means the marker is just a few lines earlier than the page break, but if a poem is several pages long, then the markers for most of the pages are just lumped together at the beginning of the poem.

This makes it harder to find and see the correct page images while doing the pp work, and in the HTML version, makes ToC and Index page references incorrect, sometimes by several pages.

[Anonymous]

I think it would help if you provided a demo file for this...

[Anonymous]

The attached zip (418KB) contains two html files and their corresponding .bin files. They are complete files, so they're pretty large, but I'll identify a precise example of where this happens. Although the eBook's been posted to PG, it's still available here at DP if you want to look at the page images: "Stories of Old Greece and Rome", url: http://www.pgdp.net/c/project.php?id=projectID52d5bc7a9efbe&

The first file is h01-pregen.html. Line 4727 begins a poem that is near the bottom of page 157, which is 201.png url: http://www.pgdp.net/c/tools/project_manager/displayimage.php?project=projectID52d5bc7a9efbe&imagefile=201.png and the first two lines of the poem are on that page; the rest of the poem is on the next page.

The second file is h02-gen.html and is what GG produced with AutoGen from the first file. If you use GG to look at the poem, you will see that it's now on page 158, with the pagenum span to the right of the first line of the poem.

<tb>

This also happens when more of the poem is on the first page, but one example should suffice. If you need more information, just let me know.

[Anonymous]

The page marker in h01-pregen.html is at the beginning of the poem, so h02-gen.html has the page number placed correctly. If the page marker should've been in the middle of the poem it must have been moved earlier. There doesn't seem to be any problem with HTML generation here.

[Anonymous]

Ah! The problem occurs MUCH earlier, when I change /*...*/ to /p...p/ and the */ is at the bottom of the page: after the regex, the page marker has moved from just after the [closing] */ to just before the [opening] /p

Modifying the regex should solve the problem, so this bug report can be closed. Sorry for all the extra work, and thank you for helping me find the true cause of the problem.

EDIT: Modifying the regex will not solve the problem, as any REPLACE that crosses a page boundary seems to move the page marker. But that's a different problem than the reported one.

[Anonymous]

Just use the regex before you run Fixup Page Separators. I always do that as late as possible, because the invisible separators are quite fragile.