Hi Peter, Thanks for testing. I've just been hacking away with alternative places to put the xml:lang attribute, and finally hit on a setup that makes the PDF. But the build still fails. :-( I can't see why -- I get a "working" PDF. (By which I mean it's complete with all pages, although as has been the case for years, stuff is broken all over the place.) I can't figure out why the xelatex build is returning an error when it finishes OK. There are myriad warnings throughout the tex log, but they're all normal as far as I can see. What's happening, do you think? The PDF output is messed up in all sorts of ways, and I can't believe anyone is using it seriously; could we consider dropping it in favour of paying more attention to the ePubs? Cheers, Martin On 2021-02-08 7:52 a.m., Peter Stadler wrote:

Hi Martin,

I was just having a look but it seems you were quicker :) Only thing left for me is to confirm that it now (with your latest commit 5ea4d84) builds for me locally, too.

Best Peter

...

Am 08.02.2021 um 02:53 schrieb Martin Holmes :

Hi all,

The ruby stuff is now merged into the repo, and I've spent a good few hours now trying to get the PDF to build properly. It dies with this error:

! LaTeX3 Error: The key 'fontspec-opentype/...' is unknown and is being (LaTeX3) ignored.

which is something to do with xml:lang='ja' appearing in the source. At one point I got it to build, but without the Japanese language attributes, the characters just come out as square boxes. Can anyone who knows more about LaTeX than me suggest anything? I don't really want to go round removing xml:lang attributes just to get past the PDF problem, but I don't know what else to do at this point.

Cheers, Martin -- ------------------------------------- Humanities Computing and Media Centre University of Victoria mholmes@uvic.ca _______________________________________________ Tei-council mailing list Tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

-- ------------------------------------- Humanities Computing and Media Centre University of Victoria mholmes@uvic.ca

Hi Martin, As Peter mentions, there may be other issues with the PDF build, but I *think* what's going on is that the GLs building processes are trying to italicise fonts (Noto CJK fonts) that don't have an Italic form. I guess that's not a thing—I wasn't aware of it. Trying to troubleshoot a bit... Hugh On Mon, Feb 8, 2021 at 11:05 AM Martin Holmes wrote:

Hi Peter,

Thanks for testing. I've just been hacking away with alternative places to put the xml:lang attribute, and finally hit on a setup that makes the PDF.

But the build still fails. :-( I can't see why -- I get a "working" PDF. (By which I mean it's complete with all pages, although as has been the case for years, stuff is broken all over the place.)

I can't figure out why the xelatex build is returning an error when it finishes OK. There are myriad warnings throughout the tex log, but they're all normal as far as I can see. What's happening, do you think?

The PDF output is messed up in all sorts of ways, and I can't believe anyone is using it seriously; could we consider dropping it in favour of paying more attention to the ePubs?

Cheers, Martin

...

Hi Martin,

I was just having a look but it seems you were quicker :) Only thing left for me is to confirm that it now (with your latest commit 5ea4d84) builds for me locally, too.

Best Peter

...

Am 08.02.2021 um 02:53 schrieb Martin Holmes :

Hi all,

The ruby stuff is now merged into the repo, and I've spent a good few hours now trying to get the PDF to build properly. It dies with this error:

! LaTeX3 Error: The key 'fontspec-opentype/...' is unknown and is being (LaTeX3) ignored.

which is something to do with xml:lang='ja' appearing in the source. At one point I got it to build, but without the Japanese language attributes,

On 2021-02-08 7:52 a.m., Peter Stadler wrote: the characters just come out as square boxes. Can anyone who knows more about LaTeX than me suggest anything? I don't really want to go round removing xml:lang attributes just to get past the PDF problem, but I don't know what else to do at this point.

...

...

Cheers, Martin -- ------------------------------------- Humanities Computing and Media Centre University of Victoria mholmes@uvic.ca _______________________________________________ Tei-council mailing list Tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

-- ------------------------------------- Humanities Computing and Media Centre University of Victoria mholmes@uvic.ca _______________________________________________ Tei-council mailing list Tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

Dear All, Having experimented with PDFs and print outputs for a number of projects now, I am very much against having the print output "tail" wagging the digital "dog." I also wonder how many people use PDFs of the Guidelines? Even as a daily user of the Guidelines, I didn't even know we had a PDF output. Trying to hack the encoding so that the PDF will build seems wrong in all kinds of ways. J. Janelle Jenstad, Associate Professor, Department of English, University of Victoria Director, Map of Early Modern London; Director, Linked Early Modern Drama Online Coordinating Platform Editor, Digital Renaissance Editions Email: jenstad@uvic.ca, lemdo@uvic.ca, or london@uvic.ca -----Original Message----- From: Tei-council On Behalf Of Martin Holmes Sent: February 8, 2021 9:13 AM To: Hugh Cayless Cc: TEI Council Subject: Re: [Tei-council] PDF build with Japanese Notice: This message was sent from outside the University of Victoria email system but is claiming to be from UVic. Please be cautious with links, attachments, and sensitive information. Notice: This message was sent from outside the University of Victoria email system. Please be cautious with links and sensitive information. Hi Hugh, Yes, I think that was definitely part of it; where I was including Japanese text inside e.g. a <mentioned> element, for example, the build would fail. The result of working around this is that the Guidelines lack tagging that they should actually have, which is a bit sad. The other thing that caused a problem, I think, was the fact that I had embedded Japanese examples inside an egXML with @xml:lang="ja", but hadn't included any wrapper element inside the example. So if you do this: <egXML xmlns="http://www.tei-c.org/ns/Examples" xml:lang="ja"> <!--...--> <ruby> <rb>大学</rb> <rt place="above">だいがく</rt> </ruby> <!--...--> </p> </egXML> then the build will fail; instead, you have to do this: <egXML xmlns="http://www.tei-c.org/ns/Examples"> <p xml:lang="ja"> <!--...--> <ruby> <rb>大学</rb> <rt place="above">だいがく</rt> </ruby> <!--...--> </p> </egXML> This is quite misleading because it suggests that the host document for the example is not Japanese; there's just a para of Japanese embedded in it. It may also encourage the perception that the language must be specified not just on the root element but throughout the document. I think these things are fixable, but frankly I don't think the PDF has sufficient value at this point that it's worth anyone fixing them, and I think it's unlikely that anyone will ever get around to it -- note that this ticket: https://github.com/TEIC/Stylesheets/issues/151 which I raised the last time I was doing multilingual stuff in the Glines remains unaddressed after five years. As long as we include the PDF in our build process, we'll continue to compromise the quality of the encoding in the Guidelines to work around bugs in it. That's not good. Cheers, Martin On 2021-02-08 8:56 a.m., Hugh Cayless wrote:

Hi Martin,

As Peter mentions, there may be other issues with the PDF build, but I *think* what's going on is that the GLs building processes are trying to italicise fonts (Noto CJK fonts) that don't have an Italic form. I guess that's not a thing—I wasn't aware of it. Trying to troubleshoot a bit...

Hugh

On Mon, Feb 8, 2021 at 11:05 AM Martin Holmes mailto:mholmes@uvic.ca> wrote:

Hi Peter,

Thanks for testing. I've just been hacking away with alternative places to put the xml:lang attribute, and finally hit on a setup that makes the PDF.

But the build still fails. :-( I can't see why -- I get a "working" PDF. (By which I mean it's complete with all pages, although as has been the case for years, stuff is broken all over the place.)

I can't figure out why the xelatex build is returning an error when it finishes OK. There are myriad warnings throughout the tex log, but they're all normal as far as I can see. What's happening, do you think?

The PDF output is messed up in all sorts of ways, and I can't believe anyone is using it seriously; could we consider dropping it in favour of paying more attention to the ePubs?

Cheers, Martin

On 2021-02-08 7:52 a.m., Peter Stadler wrote: > Hi Martin, > > I was just having a look but it seems you were quicker :) > Only thing left for me is to confirm that it now (with your latest commit 5ea4d84) builds for me locally, too. > > Best > Peter > >> Am 08.02.2021 um 02:53 schrieb Martin Holmes mailto:mholmes@uvic.ca>: >> >> Hi all, >> >> The ruby stuff is now merged into the repo, and I've spent a good few hours now trying to get the PDF to build properly. It dies with this error: >> >> ! LaTeX3 Error: The key 'fontspec-opentype/...' is unknown and is being >> (LaTeX3) ignored. >> >> which is something to do with xml:lang='ja' appearing in the source. At one point I got it to build, but without the Japanese language attributes, the characters just come out as square boxes. Can anyone who knows more about LaTeX than me suggest anything? I don't really want to go round removing xml:lang attributes just to get past the PDF problem, but I don't know what else to do at this point. >> >> Cheers, >> Martin >> -- >> ------------------------------------- >> Humanities Computing and Media Centre >> University of Victoria >> mholmes@uvic.ca mailto:mholmes@uvic.ca >> _______________________________________________ >> Tei-council mailing list >> Tei-council@lists.tei-c.org mailto:Tei-council@lists.tei-c.org >> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >

-- ------------------------------------- Humanities Computing and Media Centre University of Victoria mholmes@uvic.ca mailto:mholmes@uvic.ca _______________________________________________ Tei-council mailing list Tei-council@lists.tei-c.org mailto:Tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

-- ------------------------------------------ Martin Holmes UVic Humanities Computing and Media Centre _______________________________________________ Tei-council mailing list Tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

I'd be delighted if you've fixed it! How? I fixed the corpus/DTD error, so P5-dev is now building and the source files with ruby appearing in them now actually validate in Oxygen against the last-successful schemas. Cheers, Martin On 2021-02-08 11:16 a.m., Hugh Cayless wrote:

You may not want to hear then that I think I’ve fixed it. At an appointment right now but will try to check everything in when I get back.

Hugh

...

On Feb 8, 2021, at 12:13, Martin Holmes wrote:

Hi Hugh,

Yes, I think that was definitely part of it; where I was including Japanese text inside e.g. a <mentioned> element, for example, the build would fail. The result of working around this is that the Guidelines lack tagging that they should actually have, which is a bit sad.

The other thing that caused a problem, I think, was the fact that I had embedded Japanese examples inside an egXML with @xml:lang="ja", but hadn't included any wrapper element inside the example. So if you do this:

<egXML xmlns="http://www.tei-c.org/ns/Examples" xml:lang="ja"> <!--...--> <ruby> <rb>大学</rb> <rt place="above">だいがく</rt> </ruby> <!--...--> </p> </egXML>

then the build will fail; instead, you have to do this:

<egXML xmlns="http://www.tei-c.org/ns/Examples"> <p xml:lang="ja"> <!--...--> <ruby> <rb>大学</rb> <rt place="above">だいがく</rt> </ruby> <!--...--> </p> </egXML>

This is quite misleading because it suggests that the host document for the example is not Japanese; there's just a para of Japanese embedded in it. It may also encourage the perception that the language must be specified not just on the root element but throughout the document.

I think these things are fixable, but frankly I don't think the PDF has sufficient value at this point that it's worth anyone fixing them, and I think it's unlikely that anyone will ever get around to it -- note that this ticket:

https://github.com/TEIC/Stylesheets/issues/151

which I raised the last time I was doing multilingual stuff in the Glines remains unaddressed after five years.

As long as we include the PDF in our build process, we'll continue to compromise the quality of the encoding in the Guidelines to work around bugs in it. That's not good.

Cheers, Martin

...

On 2021-02-08 8:56 a.m., Hugh Cayless wrote: Hi Martin, As Peter mentions, there may be other issues with the PDF build, but I *think* what's going on is that the GLs building processes are trying to italicise fonts (Noto CJK fonts) that don't have an Italic form. I guess that's not a thing—I wasn't aware of it. Trying to troubleshoot a bit... Hugh On Mon, Feb 8, 2021 at 11:05 AM Martin Holmes mailto:mholmes@uvic.ca> wrote: Hi Peter, Thanks for testing. I've just been hacking away with alternative places to put the xml:lang attribute, and finally hit on a setup that makes the PDF. But the build still fails. :-( I can't see why -- I get a "working" PDF. (By which I mean it's complete with all pages, although as has been the case for years, stuff is broken all over the place.) I can't figure out why the xelatex build is returning an error when it finishes OK. There are myriad warnings throughout the tex log, but they're all normal as far as I can see. What's happening, do you think? The PDF output is messed up in all sorts of ways, and I can't believe anyone is using it seriously; could we consider dropping it in favour of paying more attention to the ePubs? Cheers, Martin On 2021-02-08 7:52 a.m., Peter Stadler wrote: > Hi Martin, > > I was just having a look but it seems you were quicker :) > Only thing left for me is to confirm that it now (with your latest commit 5ea4d84) builds for me locally, too. > > Best > Peter > >> Am 08.02.2021 um 02:53 schrieb Martin Holmes mailto:mholmes@uvic.ca>: >> >> Hi all, >> >> The ruby stuff is now merged into the repo, and I've spent a good few hours now trying to get the PDF to build properly. It dies with this error: >> >> ! LaTeX3 Error: The key 'fontspec-opentype/...' is unknown and is being >> (LaTeX3) ignored. >> >> which is something to do with xml:lang='ja' appearing in the source. At one point I got it to build, but without the Japanese language attributes, the characters just come out as square boxes. Can anyone who knows more about LaTeX than me suggest anything? I don't really want to go round removing xml:lang attributes just to get past the PDF problem, but I don't know what else to do at this point. >> >> Cheers, >> Martin >> -- >> ------------------------------------- >> Humanities Computing and Media Centre >> University of Victoria >> mholmes@uvic.ca mailto:mholmes@uvic.ca >> _______________________________________________ >> Tei-council mailing list >> Tei-council@lists.tei-c.org mailto:Tei-council@lists.tei-c.org >> http://lists.lists.tei-c.org/mailman/listinfo/tei-council > -- ------------------------------------- Humanities Computing and Media Centre University of Victoria mholmes@uvic.ca mailto:mholmes@uvic.ca _______________________________________________ Tei-council mailing list Tei-council@lists.tei-c.org mailto:Tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

-- ------------------------------------------ Martin Holmes UVic Humanities Computing and Media Centre

-- ------------------------------------------ Martin Holmes UVic Humanities Computing and Media Centre