Changes to build-toc to address issue #628

[drgrigg]

I've altered where we check for a language tag in build-toc which seems to fix the problem in issue #628.

I've run a comparison between the results of new and old tocs and the only productions which don't match are either ones that always required a manual adjustment to the ToC or a handful of others which have unrelated issues (though some of these mean I should recheck my code in other areas.

For reference, the non-matching ToCs after this change to build-toc are as follows. The line numbers are in the ToC for each production.

• anatole-france_penguin-island_a-w-evans : line 132
• arthur-w-pinero_the-second-mrs-tanqueray : line 20
• friedrich-nietzsche_beyond-good-and-evil_helen-zimmern : line 236
• george-macdonald_phantastes : line 17
• honore-de-balzac_shorts-from-scenes-from-private-life_clara-bell_ellen-marriage : line 27
• james-branch-cabell_domnei : line 17
• jean-toomer_cane : line 27
• john-locke_some-thoughts-concerning-education : line 20
• ludwig-wittgenstein_tractatus-logico-philosophicus_c-k-ogden : line 32
• p-g-wodehouse_short-fiction : line 82
• t-e-lawrence_seven-pillars-of-wisdom : line 453
• wilfred-owen_poetry : line 76

[acabal]

To clarify, which ones had "unrelated issues"? I know many of those always required manual fixes, but are there some now that have some kinds of issues?

[drgrigg]

I’ll spell them all out later today, but one source of difference is just changes in our practice since the book was originally produced, such as using z998:name-title in abbreviation tags. There’s a couple I want to check my code for, but they don’t seem at all related to the changes I made in the PR. Still, I should have checked that first! I’ll have time later today to do that.

…

On 29 Dec 2023 at 3:09 AM +1100, Alex Cabal ***@***.***>, wrote: To clarify, which ones had "unrelated issues"? I know many of those always required manual fixes, but are there some now that have some kinds of issues? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>

[drgrigg]

To clarify my clarification (!) I have been comparing ToCs created by my new code to the ToCs in the released productions because I wanted to pick up the ones I knew had manual alterations.

…

On 29 Dec 2023 at 9:38 AM +1100, ***@***.***, wrote: I’ll spell them all out later today, but one source of difference is just changes in our practice since the book was originally produced, such as using z998:name-title in abbreviation tags. There’s a couple I want to check my code for, but they don’t seem at all related to the changes I made in the PR. Still, I should have checked that first! I’ll have time later today to do that. On 29 Dec 2023 at 3:09 AM +1100, Alex Cabal ***@***.***>, wrote: > To clarify, which ones had "unrelated issues"? I know many of those always required manual fixes, but are there some now that have some kinds of issues? > — > Reply to this email directly, view it on GitHub, or unsubscribe. > You are receiving this because you authored the thread.Message ID: ***@***.***>

[drgrigg]

OK, first point: there are zero differences in the ToCs generated with the old master branch build-toc and those generated with the new code in the fixtoc branch. So I think that's OK to merge.

The ToCs in the released versions which differ are as follows:

• anatole-france_penguin-island_a-w-evans : a known issue where the naming of the Book parts needs to use the subtitles rather than the titles in the hgroup, eg "Trinco" rather than "Modern Times" as the latter is repeated for several of the book parts.
• arthur-w-pinero_the-second-mrs-tanqueray : the halftitlepage entry in the release ToC is missing abbreviation tags around "Mrs." Do you want me to raise a pull request to fix that?
• friedrich-nietzsche_beyond-good-and-evil_helen-zimmern : the two entries for chapter-4-65a and chapter-4-73a show up with lower-case "a" in the ToC of the release version, whereas the source has an uppercase "A", which is what build-toc therefore uses.
• george-macdonald_phantastes : the release ToC correctly has tags for the roman numerals in Epigraph I, Epigraph II and Epigraph III. But build-toc is reduced to using the <title> to get the text for the ToC and of course that doesn't include the "roman" tag. Rather than trying tricky code on this, it's rare enough for it to be left as a manual task, I think.
• honore-de-balzac_shorts-from-scenes-from-private-life_clara-bell_ellen-marriage : definitely something odd going on here: when a short story starts with a letter or other material in a separate section before the actual story begins in a new section, build-toc is indenting it as though it were a sub-chapter in the previous story. I'll need to look at this closely in the code.
• james-branch-cabell_domnei : same issue as for george-macdonald_phantastes.
• jean-toomer_cane : build-toc isn't correctly indenting the ToC entries in each Part, even though data-parent is correctly used. This definitely needs to be looked at in the code, I think it's due to the fact that the short stories are in articles rather than sections.
• ludwig-wittgenstein_tractatus-logico-philosophicus_c-k-ogden : this has a complex structure, as you know, and is one of our known exceptions to automatic build-toc.
• p-g-wodehouse_short-fiction : it looks as though the order and number of stories in the collection has changed without build-toc having been run again. Do you want me to raise a pull request to fix that?
• t-e-lawrence_seven-pillars-of-wisdom : another odd problem to do with indenting in the ToC, this time a spurious indent of the first Appendix, as though it were a subchapter of the Endnotes. I'll need to have a look at the code to see what's going on.
• wilfred-owen_poetry : the release ToC shows a subtitle for the poem "Strange Meeting", that subtitle being "Another Version". This definitely has to be a manual exception to the build-toc rules.

So, to sum up, I think the changes in the pull request should be applied, but there are several new issues thrown up by my comparison between the released ToCs and what build-toc is generating. But I think these should be the subject of new Issues, which I'll create.

[acabal]

OK great, thanks. Some of those can't be fixed as the ebooks are exceptions, like Phantastes. If you have time though you can look at the ones that aren't to see if we can get things into shape.