Tagging support for array, tabularx, and longtable #1209

FrankMittelbach · 2023-12-16T13:32:40Z

Internal housekeeping

Status of pull request

Ready for merging (further development and cleanup later)

Checklist of required changes before merge will be approved

Test file(s) added
Version and date string updated in changed source files
Relevant \changes entries in source included
Relevant changes.txt updated
Rollback provided (if necessary)?
ltnewsX.tex (and/or latexchanges.tex) updated

array: put sockets into array ; array: updated \insert@column to use sockets

us separate \insert@pcolumn with its own sockets

change begindocument hook chunk to "kernel"

base/lttagging.dtx

required/tools/array.dtx

u-fischer · 2023-12-16T15:49:23Z

required/tools/array.dtx

@@ -2486,11 +2766,41 @@
 %    =\@array=. To save \textsf{token} space we then redefine
 %    =\@preamble=
 %    because its replacement text isn't longer needed.
+%


The $ around 20 lines above should be surrounded by something like \FakeMathStart$\FakeMathStop to avoid that it triggers math tagging. The commands can be defined as dummies (in lttagging.dtx?) and the math code can then redefine them.

We have to talk about that suggested fakemath interface, I think it is wrong, but that isn't for this pull.

It would be really good if there were directly an interface (can be internal) to suppress the math grabbing here. Otherwise it is hard to tag math inside tabulars (one has to disable it outside the tabular and reenable it in every cell) and I need that for the larger math example ... (also it would resolve an open issue with tabularx).

Again, I can't put anything in that is only in your private version of tagpdf, because if that then goes out as latex-dev it will blow up. Anway, in my opinion the interface should just say \FakeMathStart hiding the $ inside (and whatever is necessary to ensure that inner math is still grabbed and tagged --- which may be nothing if the command is protected)

Sorry? There is nothing about this in my private tagpdf. We simply need here something that allows us to say "this $ should not do math grabbing" and does not affect following math.

sorry not in your tagpdf but in your branch for math coding which is not released

but why should it blow up? I only asked that there is something around the dollar which does nothing by default but can be redefined by the math code to suppress the math grabbing in this place.

# Conflicts: # required/latex-lab/testfiles-table-luatex/table-000.tlg # required/latex-lab/testfiles-table-pdftex/table-000.tlg # required/latex-lab/testfiles/documentmetadata-support-000.luatex.tlg # required/latex-lab/testfiles/documentmetadata-support-000.tlg # required/latex-lab/testfiles/documentmetadata-support-001.luatex.tlg # required/latex-lab/testfiles/documentmetadata-support-001.tlg # required/latex-lab/testfiles/documentmetadata-support-002.luatex.tlg # required/latex-lab/testfiles/documentmetadata-support-002.tlg

# Conflicts: # required/tools/changes.txt

# Conflicts: # required/latex-lab/testfiles-table-luatex/table-004-tabularx.tlg # required/latex-lab/testfiles-table-luatex/table-005.tlg # required/latex-lab/testfiles-table-luatex/table-006-longtable.tlg # required/latex-lab/testfiles-table-luatex/table-007-longtable.tlg # required/latex-lab/testfiles-table-luatex/table-008-multi.tlg # required/latex-lab/testfiles-table-luatex/table-009.tlg # required/latex-lab/testfiles-table-pdftex/table-004-tabularx.tlg # required/latex-lab/testfiles-table-pdftex/table-005.tlg # required/latex-lab/testfiles-table-pdftex/table-006-longtable.tlg # required/latex-lab/testfiles-table-pdftex/table-007-longtable.tlg # required/latex-lab/testfiles-table-pdftex/table-008-multi.tlg # required/latex-lab/testfiles-table-pdftex/table-009.tlg

# Conflicts: # base/changes.txt # required/tools/changes.txt

davidcarlisle

got to end of longtable.dtx (only made doc comments) out of time this morning but will pick up from here, but pushing these comments now

base/doc/ltnews39.tex

davidcarlisle · 2024-02-08T09:12:21Z

base/doc/source2e.tex

@@ -342,6 +342,8 @@

 \DocInclude{ltoutput} % Output routine

+ \DocInclude{lttagging}% Tagging support


not an issue for this PR but since we went past a-z-A-Z for source2e files, the sorting of files in the code index is off as a-zA-Zaa-zz doesn't work with the current makeindex style.

ahh, good point. Could you open a separate pr that this should be looked at at some point?

required/latex-lab/latex-lab-table.dtx

required/tools/longtable.dtx

davidcarlisle · 2024-02-08T09:25:54Z

required/tools/longtable.dtx

+%    them up as artifact with the exception of the head at the begin and
+%    the foot box at the end.
+%
+%    TODO: When a line is killed the structure subtree is there already


if this is hard to do in this release should we make \LT@kill warn that tagging may (will?) have duplicated structure?

yes, maybe , but it is probably not too hard for @u-fischer ... we'll see. I added a further note.

hm, not easy. Currently \kill (with table-tagging) errors, and if I avoid the error it doesn't kill the row (don't know why). And if that problem can be resolved: getting tagging right is not trivial either, it would probably require some label and an additional counter to keep track on the rows that should be ignored. I'm inclined to make that a simply error with "sorry not supported with tagging".

I guess part of the problem is that it acts "after the fact" so to speak. That means one would basically either wait with a lot of things to see what comes at the end of a row or maybe record the fact that there was a kill in row X and next time around do the killing properly.
Instead if this we could perhaps have a \killstart command to be added at the beginning of the row which would tell the tagging what's up. That would still mean some manual work with such documents but at least the functionality could be preserved.

Perhaps @davidcarlisle has some thoughts?

In a first version, just stating that this doesn't work with tagging is certainly ok.

I was going to suggest same, whether \xkill ..... \\ would be easier so that knew up front that this was a killed row. Can you handle \kill in the kernel tabbing ? (on which the longtable version was based)

No we can't do tabbing yet, but the problem with tabbing is also to figure out first how that should be tagged in general.

required/tools/longtable.dtx

@array

* take latex-lab-table.dtx from uf-review-05 * latex-lab-table: move \@@_show_curr_cell_data: into socket ; array: put sockets into array ; array: updated \insert@column to use sockets * first stab at tabularx (WIP) * getting there (tabularx) * fix docu * fix so that the code runs when tagging is not used * leftover \end{macrocode} * add verst version of lttag.dtx * simplify tabularx * provide \ProvideDocElement * change lttag -> lttagging; us separate \insert@pcolumn with its own sockets * WIP (broken) * WIP * WIP (working again) * WIP * medium rewrite/restructure (WIP) * would have been too easy * and a doc issue * and one more * move multicolumn changed into array; change begindocument hook chunk to "kernel" * better also add the socket declaration in the right file * starting with L3 layer in array (WIP) * added missing \MaintainedBy defs * l3doc needs amsmath :-( * updated \@array in array * move \ar@ialign into array * tracing/debugging now optional * move \@addamp * move \@Arraycr * move \endarry ; add \ar@crcr * move cell data handling to lttaging; start with longtable (WIP) * stray macrocode * use \tbl_update_multicolumn_cell_data:n * more work on longtable (WIP) * update \endlongtable and \LT@t@bularcr (WIP) * hopefully fixed logic for \endhead and friends; also fixed \hline issue (guess \cline needs also attention but not tonight) * trying to get around l3bitset issue * using newer kernel with l3bitset included * l3bibset vanished * sigh * and a docu fix * cleanup docu (WIP) * docu changes (fail table-14-pbox) * doc updates for "array" * renamed a couple of sockets; but more docu * more updates and documentation * getting ready for first draft * implemented most of Ulrike's suggestions * more suggestions * cleanup after merge * drop definitions now in tagpdf * ltnews entry * fix tests * updated with David's comments

FrankMittelbach added 30 commits November 28, 2023 10:34

take latex-lab-table.dtx from uf-review-05

88daedc

latex-lab-table: move \@@_show_curr_cell_data: into socket ;

9702527

array: put sockets into array ; array: updated \insert@column to use sockets

first stab at tabularx (WIP)

4d06e3e

getting there (tabularx)

fd92251

fix docu

a0ddc7d

fix so that the code runs when tagging is not used

2a39266

leftover \end{macrocode}

15c3a73

add verst version of lttag.dtx

3633632

Merge branch 'develop' into array

9fb3bce

simplify tabularx

ee8c964

provide \ProvideDocElement

e4d3f8d

change lttag -> lttagging;

1bb7ae7

us separate \insert@pcolumn with its own sockets

Merge commit 'e4d3f8dc1fa6cdfeed6a32908e984a2dcc29dd33' into array

f71035a

WIP (broken)

f259df8

WIP

c33f05a

WIP (working again)

f7ea62c

WIP

cc868f9

medium rewrite/restructure (WIP)

9e75a8e

would have been too easy

e491d6c

and a doc issue

150198a

and one more

e9b55a9

move multicolumn changed into array;

083f9d9

change begindocument hook chunk to "kernel"

better also add the socket declaration in the right file

866476b

Merge branch 'develop' into array

8d160cf

starting with L3 layer in array (WIP)

69ee815

added missing \MaintainedBy defs

a3fd27e

l3doc needs amsmath :-(

171f08e

updated \@array in array

f9154ab

move \ar@ialign into array

5cfbdf3

tracing/debugging now optional

f37eb89

FrankMittelbach added 6 commits December 11, 2023 19:17

cleanup docu (WIP)

e9f8fb8

docu changes (fail table-14-pbox)

2184096

doc updates for "array"

417c9d6

renamed a couple of sockets; but more docu

938bd73

more updates and documentation

b86ce59

getting ready for first draft

216037d

FrankMittelbach requested review from josephwright, davidcarlisle, u-fischer and car222222 December 16, 2023 13:33

u-fischer reviewed Dec 16, 2023

View reviewed changes

FrankMittelbach added 10 commits December 17, 2023 00:25

implemented most of Ulrike's suggestions

cb59e5a

more suggestions

5d382cf

cleanup after merge

e6d3bbd

Merge branch 'develop' into array

7cc23b6

# Conflicts: # required/tools/changes.txt

drop definitions now in tagpdf

39a0ce1

Merge branch 'develop' into array

1c18437

# Conflicts: # base/changes.txt # required/tools/changes.txt

ltnews entry

5513e8b

fix tests

81dc33b

FrankMittelbach requested a review from u-fischer February 8, 2024 08:42

davidcarlisle reviewed Feb 8, 2024

View reviewed changes

updated with David's comments

4671ea4

davidcarlisle self-requested a review February 8, 2024 22:00

davidcarlisle approved these changes Feb 8, 2024

View reviewed changes

u-fischer approved these changes Feb 8, 2024

View reviewed changes

FrankMittelbach merged commit c1eedbc into develop Feb 9, 2024
78 checks passed

FrankMittelbach deleted the array branch March 11, 2024 14:19

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Tagging support for array, tabularx, and longtable #1209

Tagging support for array, tabularx, and longtable #1209

FrankMittelbach commented Dec 16, 2023 •

edited

Loading

u-fischer Dec 16, 2023

FrankMittelbach Dec 16, 2023

u-fischer Dec 16, 2023

FrankMittelbach Dec 16, 2023

u-fischer Dec 16, 2023

FrankMittelbach Dec 16, 2023

u-fischer Dec 16, 2023

davidcarlisle left a comment

davidcarlisle Feb 8, 2024

FrankMittelbach Feb 8, 2024

davidcarlisle Feb 8, 2024

FrankMittelbach Feb 8, 2024

u-fischer Feb 8, 2024

FrankMittelbach Feb 8, 2024

davidcarlisle Feb 8, 2024

FrankMittelbach Feb 9, 2024

		@@ -342,6 +342,8 @@

		\DocInclude{ltoutput} % Output routine

		\DocInclude{lttagging}% Tagging support

Tagging support for array, tabularx, and longtable #1209

Tagging support for array, tabularx, and longtable #1209

Conversation

FrankMittelbach commented Dec 16, 2023 • edited Loading

Internal housekeeping

Status of pull request

Checklist of required changes before merge will be approved

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

davidcarlisle left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

FrankMittelbach commented Dec 16, 2023 •

edited

Loading