These exports are failing on an updated TeX Live 2020. It is not
high priority to look into the underlying cause of these two
failures since TeX fonts are not commonly used with LuaTeX and the
ectaart class is not in TeX Live.
Includes a fix when wrappers are used with empty paragraphs. The logic is probably wrong if two legit wrapped paragraphs have one empty paragraph in the middle. This is only for slightly convoluted cases... Wrapping decisions should rather be taken for a series of paragraphs at once, but it would be more complex to implement and would only solve very peculiar cases like these ones.
The export inputenc-luatex-utf8_pdf5_texF has been succeeding for a
while and was likely fixed by a TeX Live update. Compiling manually
and looking at the output looks good.
The unicode tests would often fail when tested in parallel because
we were not exporting to unique file names. From what I understand,
a variant similar to the following race condition occurred:
1. Thread A exports to file blah.pdf.
2. Thread B exports to file blah.pdf.
3. Thread A confirms file blah.pdf exists.
4. Thread A deletes exported file blah.pdf to clean up.
5. Thread B fails to find file blah.pdf and reports a failure.
Runs the compare via the command line, and then compares the output to the
expected result. Required adding a script to do the comparison, so that
the timestamps on changes in the lyx file are ignored.
The new JSS cls has been failing on an updated TeX Live 2020 for a
while, both with the template they provide and with LyX's template.
I expect this will be fixed eventually but for now it's best to
invert it as a "texissue".
With an updated TeX Live 2020 an assertion is given. The assertion
was reported on the LuaTeX ML [1], and will be converted into the
following error as of LuaTeX svn commit r7385:
! the word doesn't start with a character
The error could likely be worked around with a change to the
document but since the document compiles without error with XeTeX I
did not change anything since I do not understand the issue.
[1]
https://www.mail-archive.com/search?l=mid&q=20201030030430.4y4i6bdpwx633qs7%40vbox-VirtualBox
Export to all formats seems to work well after removing the "ps2pdf"
option to the hyperref package.
Accordingly we uninvert the tests for the other formats. All ctests
pass on an updated TeX Live 2020.
This is mostly for shapepar support, in a rare situation. Fixing this would create a lot of special cases in output_docbook.cpp, i.e. fixing the issue (which will barely happen in real life) would make maintenance much harder.
The debug mode is set with the environment LYX_DEBUG_LATEX
$ export LYX_DEBUG_LATEX=1
The downside: From time to time the need to remove the superfluous dirs
$ cd build-dir
$ find autotests/out-home -name AbC_\* | xargs rm -rf
More generally, ensures that paragraphs in abstracts do not have something else configured.
A major problem in making the layout more useful is that article titles are not supposed to be in TOC.
Compilation of our Seminar example file fails on updated TL20. The
maintainer of "Seminar" is not planning to fix the core issue and
states the following (in a private email with permission to quote):
it is a problem with the new hook management of the current latex.ltx
seminar is a quite old package and there is no reason to use it with a
new LaTeX format. It won't be fixed, so the usual way is to use the
package latexrealease to get the old hook management.
This commit adds a note to the example files explaining the
workaround of exporting to a .tex file and prepending the following
line:
\RequirePackage[2020-02-02]{latexrelease}
We now invert the relevant tests.
1.) If building with external iconv, remove remnants of libiconv data
created if previousy used internal iconv
2.) Remove the variable HELP from cache to enable subsequent builds.
(Different variable-handling on new cmake (version >= 3.18))
3.) New macro to extract cmake-known-cxx-sts features
4.) Check for CXX11-features only for a subset of possibilities
known by the used cmake
LyX should still avoid crashing on those, and if they start to pass, then something is wrong (most likely, a very large part of the files is being ignored).
This is a minimal implementation, as DocBook lacks a serious way of encoding all of this. Maybe a <formalpara> could do the trick, but I'd need to find a way to shoehorn a title through the styles (i.e. a first complete tag):
Theorem: Bla bla
<formalpara>
<title>Theorem</title>
<para>Bla bla</para>
</formalpara>
This would also only be a solution for single-paragraph things, as formalpara only allows one paragraph. Or a sidebar, but it's semantically very remote.
That's mostly generating DocBook tags at an inappropriate place with some metadata, rather than outputting whatever you have at your disposal. Far from satisfying, but good enough for a generic tool (see details in the new TODO). Doesn't trigger assertions. Is valid XML.
These exports now pass, and the output looks reasonable to me
(although I do not know Hebrew). I believe they work now because of
Jürgen's fixes at a7ad0747 and d7b64b8e.
Previously there are po-entries like
#: lib/examples/Articles:0 src/TocBackend.cpp:296
#:src/frontends/qt/GuiExternal.cpp:87
msgid "External Material"
msgstr ""
now:
#: src/TocBackend.cpp:296 src/frontends/qt/GuiExternal.cpp:87
#: lib/examples/Articles:0
msgid "External Material"
msgstr ""
(The string at 'lib/examples/Articles:0' proceeds also from the directory name 'lib/examples/External_Material')
This makes it easier to use some po-editors like 'linguist'
1.) Math-editor seems to use system fonts, so install it there.
2.) We use fonts from the support-dir, so install them there too.
This is something automake may take int account too.
Thanks to Jürgen, who mentions the following:
luaotfload does not find "DavidCLM". In fact, at least on my system,
there is no such font, only "DavidCLM Medium" (and other shapes). This
one is found. Apparently, luaotfload cannot infer from the one to the
other.
As opposed to LuaTEX, XeTeX also queries TEXMF so maybe it just finds
its font there.
In many cases, round trip with older formats involves exporting ERT
or preamble code in the backwards conversion. In the forwards
conversion, if this code is not parsed, often errors can result.
However, in many cases, especially for older formats, it might not
be worth the time or code complexity to address these cases. Such
tests are labled "ertroundtrip".
This commit also inverts a currently failing lyx22x test under the
label "ertroundtrip" since the above paragraph is my best guess as
to why that test is failing. It is likely not worth the time to fix
it, especially since the APA7 layout wasn't even shipped for LyX
2.2.x.
