Open martindholmes opened 8 years ago
@smjwsk already has a patch (locally) in place, I believe.
if @smjwsk did have a patch, we never got it. This is still a problem, four years on.
My dirty hack for that caption would be to insert a <seg xml:lang='ja'>
into the <title>
at https://github.com/TEIC/TEI/blob/513c0b14a94a157ddede60dea902d8cd00208224/P5/Source/Guidelines/en/WD-NonStandardCharacters.xml#L992
As noted at https://github.com/TEIC/Stylesheets/issues/202#issuecomment-610989143 the processing of @xml:lang
is very limited and only works for some elements.
As of now, it's p.211 Figure 3, which has this caption:
5.7 Examples of Different Writing ModesFigure 3:Detail from p.62 of”. . 1985. II. 11
Note that this is not fixed by https://github.com/TEIC/TEI/commit/66144683a4a210a8ea6a5e23dffc74ad80d142b6 which switches to Noto fonts; I believe that the @xml:lang
attribute is not being acted on here:
<figure>
<graphic width="500px" height="624px" url="Images/ja_vertical_indonesian_frag.jpg"/>
<head>Detail from p.62 of <title xml:lang="ja">インドネシア語". 崎山理. 1985. 外国語との対照 II. 講座日本語学 11.</title></head>
</figure>
although I haven't confirmed that yet.
Exactly @martindholmes, as I wrote in https://github.com/TEIC/Stylesheets/issues/202#issuecomment-610981840
It turned out it's not a font issue but a processing issue – which is way bigger, sigh. There is actually almost no support for
@xml:lang
except for occurrences on<bibl>
and<seg>
.
Hence the proposed hack to to insert a <seg xml:lang='ja'>
into the <title>
.
@peterstadler Hmm. Should we hide the problem in this case and pretend it's not there, or try and fix it?
I ve seen FOP do this when either a) only essential Linux font support libraries exist on a VM , or b) the JVM running fop was not properly set to Unicode encoding for all its locale and encoding settings. Some other corner cases, but either or both of these would be my best guess.
Japanese fonts appear to be working fine in the latest build with Noto; I think this is just a case where @xml:lang
is being ignored in processing, as @peterstadler suggests.
@pstadler Hmm. Should we hide the problem in this case and pretend it's not there, or try and fix it?
Just for the record: My Github handle is @peterstadler . I got confused myself about this "Doppelgänger"
Yes, I got confused as well. Hey Peter. All good?
Sorry @pstadler @peterstadler. @peterstalder is @pstadler
in Slack, no?
The current build of the Guidelines PDF, p.203, Figure 3:
http://jenkins-paderborn.tei-c.org/job/TEIP5/lastSuccessfulBuild/artifact/P5/release/doc/tei-p5-doc/en/Guidelines.pdf
shows a clear example of this. Compare the same part of the HTML, where it's described as Figure 5.2 (subject of a different ticket):
http://jenkins.tei-c.org/job/TEIP5/lastSuccessfulBuild/artifact/P5/release/doc/tei-p5-doc/en/html/WD.html#WDWMEG2