source blocks: syntax highlighting for JSON does not always work

jechterhoff commented 1 year ago

When converting an OGC draft Best Practice document with JSON source blocks, some of these blocks end up in all green text, while others get what looks like proper syntax highlighting. When trying the CodeRay syntax highlighting online, both cases work as expected.

@ronaldtse: Do you have any idea why this could be the case?

Here are some screenshots:

Example 1: definitions schema

Defined in adoc as:

[[example_req_core_jsonschema_definitions_schema]]
.JSON Schema example of a definitions schema
[source,json,linenumbers]
----
{
  "$schema": "http://json-schema.org/draft/2019-09/schema",
  "$id": "http://example.org/some-definitions-schema.json",
  "$defs": {
    "Class1": {
      "type": "object",
      "properties": {
        "prop1": {"type": "string"}
      },
      "required": ["prop1"]
    },
    "Class2": {
      "type": "object",
      "properties": {
        "prop2": {"type": "number"}
      },
      "required": ["prop2"]
    }
  }
}
----

HTML compiled by metanorma looks like this:

definitions_schema_source_example

The CodeRay online example shows proper syntax highlighting: http://coderay.rubychan.de/rays/9369

Example 2: definitions schema

Defined in adoc as:

[[example_req_core_jsonschema_types_inheritance_generalization]]
.JSON Schema example for realizing generalization using "allOf"
[source,json,linenumbers]
----
{
  "$schema": "http://json-schema.org/draft/2019-09/schema",
  "$defs": {
    "TypeA": {
      "properties": {
        "propertyA": {
          "type": "number"
        }
      },
      "required": [
        "propertyA"
      ]
    },
    "TypeB": {
      "allOf": [
        {
          "$ref": "#/$defs/TypeA"
        },
        {
          "type": "object",
          "properties": {
            "propertyB": {
              "type": "string"
            }
          },
          "required": [
            "propertyB"
          ]
        }
      ]
    }
  },
  "$ref": "#/$defs/TypeB"
}
----

HTML compiled by metanorma looks like this:

generalization_source_example

CodeRay online example: http://coderay.rubychan.de/rays/9370

metanorma --version:

Metanorma 1.5.3 Metanorma::Cli 1.6.7 Metanorma::Standoc 2.2.8/IsoDoc 2.3.6 Metanorma::ISO 2.2.4 Metanorma::Iec 2.1.13 Metanorma::IEEE 0.1.3 Metanorma::Ietf 3.0.14 Metanorma::Generic 2.2.5 Metanorma::BIPM 2.1.13 Metanorma::CC 2.1.13 Metanorma::Csa 2.1.13 Metanorma::IHO 0.6.13 Metanorma::M3AAWG 2.1.13 Metanorma::UN 0.9.13 Metanorma::Ogc 2.2.7 Metanorma::ITU 2.1.13

coderay --version: CodeRay 1.1.3

ronaldtse commented 1 year ago

Thanks @jechterhoff for reporting this. On first look I do not know why this is happening.

I tried these two snippets locally and I get syntax highlighting for both, in the highlight pattern of the second image.

By any chance can you share the source for reproduction? Thanks!

jechterhoff commented 1 year ago

@ronaldtse The source is actually available in the following GitHub repository: https://github.com/Geonovum/uml2json

opoudjis commented 1 year ago

Our code prettifier to date has been https://github.com/googlearchive/code-prettify

We constrained the languages we recognised to a small number that we knew (at the time) that Google code-prettify supported; JSON was not in the list. (In fact, it still isn't.)

Because JSON is not in the list, whatever code-prettify is doing, it is doing without any guidance from us: it is inferring the span CSS classes to apply on its own. It is applying spans in the second instances that it doesn't on the first, but I'm not convinced that it actually knows what it's doing on the second span. I don't know why Ronald is getting different results from me, but I do know that whatever code-prettify is doing, I don't trust it.

As it turns out, the code-prettify repo was archived a month ago, and the URI we are invoking was supposed to have been turned off by Google 3 years ago. It hasn't, and this is clearly still a heavily used script, but it makes sense for us to move on from it, to an actively maintained syntax highlighter.

After looking at the 4 solutions Asciidoctor natively supports, highlight.js, rouge, pygments, and coderay, I'm going to go with rouge. I would rather go with highlight.js frankly, because that's a lot less work for me, but we do want our HTML to be self-contained as much as possible, and server-side also allows us to extend highlighting to DOC. (PDF in principle too, but that might prove more fiddly.)

I'm avoiding pygments because we'd rather not drag python into the compilation chain if we can avoid it, and coderay seems a lot more restricted than rouge, in both language support, and pre-fab skins (which is going to be an issue for HTML styling specific to SDOs.)

All of rouge, pygments, and coderay support line numbering, even though highlights.js refuses to (and there is a very popular plugin that does implement line numbering.) I've been reluctant to add line numbering as an attribute to source code, but I see that it is a recurring enough requirement to support, and that Asciidoctor does support it in its integrations (though its rouge support is lagging).

So, I will be switching syntax highlighting in HTML (and Word, and potentially PDF) to Rouge. This will take a few days, but unless some of the issues stuck in queue become more urgent, I should have this done in time for next Metanorma release next week.

jechterhoff commented 1 year ago

@opoudjis Sounds good to me. That may also solve #476.

opoudjis commented 1 year ago

It won't, @jechterhoff, because the quotes are being made smart at an earlier stage of processing. I have not been able to replicate a duplicate issue instance of this behaviour, and I'll see if I can replicate it in your case.

opoudjis commented 1 year ago

@Intelligent2013 I am about to move Syntax Highlighting for code snippets from client-side, and HTML-only, to server-side, on compilation of Metanorma.

The syntax highlighting generates HTML with CSS styling in spans. If I do the styling in Presentation XML instead of my HTML output, and convert the output from HTML back to XML (so removing the <br/>), can you process the CSS-styled spans into something sensible in PDF?

You would be getting something like:

<sourcecode>require <span class="s2">"uri"</span> <span class="k">if</span> /^2<span class="se">\.</span>/.match?<span class="o">(</span>....</sourcecode>

or, if we enable line numbers,

<table type="sourcecode">
<tbody>
<tr><td>1<br/>2<br/>3...</td></tr>
<tr><td><sourcecode>...</sourcecode></td></tr>
</tbody>
</table>

I'd end up putting the CSS definitions in the misc-container element for you to access:

".highlight table td { padding: 5px; }\n.highlight table pre { margin: 0; }\n.highlight, .highlight .w {\n  color: #303030;\n}\n.highlight .err {\n  color: #151515;\n  background-color: #ac4142;\n}\n.highlight .c, .highlight .ch, .highlight .cd, .highlight .cm, .highlight .cpf, .highlight .c1, .highlight .cs {\n  color: #505050;\n}\n.highlight .cp {\n  color: #f4bf75;\n}\n.highlight .nt {\n  color: #f4bf75;\n}\n.highlight .o, .highlight .ow {\n  color: #d0d0d0;\n}\n.highlight .p, .highlight .pi {\n  color: #d0d0d0;\n}\n.highlight .gi {\n  color: #90a959;\n}\n.highlight .gd {\n  color: #ac4142;\n}\n.highlight .gh {\n  color: #6a9fb5;\n  background-color: #151515;\n  font-weight: bold;\n}\n.highlight .k, .highlight .kn, .highlight .kp, .highlight .kr, .highlight .kv {\n  color: #aa759f;\n}\n.highlight .kc {\n  color: #d28445;\n}\n.highlight .kt {\n  color: #d28445;\n}\n.highlight .kd {\n  color: #d28445;\n}\n.highlight .s, .highlight .sb, .highlight .sc, .highlight .dl, .highlight .sd, .highlight .s2, .highlight .sh, .highlight .sx, .highlight .s1 {\n  color: #90a959;\n}\n.highlight .sa {\n  color: #aa759f;\n}\n.highlight .sr {\n  color: #75b5aa;\n}\n.highlight .si {\n  color: #8f5536;\n}\n.highlight .se {\n  color: #8f5536;\n}\n.highlight .nn {\n  color: #f4bf75;\n}\n.highlight .nc {\n  color: #f4bf75;\n}\n.highlight .no {\n  color: #f4bf75;\n}\n.highlight .na {\n  color: #6a9fb5;\n}\n.highlight .m, .highlight .mb, .highlight .mf, .highlight .mh, .highlight .mi, .highlight .il, .highlight .mo, .highlight .mx {\n  color: #90a959;\n}\n.highlight .ss {\n  color: #90a959;\n}"

If this is doable, let me know, and I'll create a ticket. If this is not doable, let me know, and I'll limit myself to HTML and DOC.

You are already doing syntax highlighting, and while I do want to eliminate redundant processing from the PDF, I can let you continue doing so if you're comfortable to. There is no clear efficiency dividend in this case, the way there is in biblio-tag.

Intelligent2013 commented 1 year ago

@opoudjis I agree with your proposal. I would be better to use common approach/styles in the syntax highlighting. And yes, I can apply CSS to span in PDF. But one restriction - CSS should be simple, without browser-specific properties like moz-..., etc.

if we enable line numbers,

ok for me too.

opoudjis commented 1 year ago

Having to modify Metanorma default assumption that all tables must have borders: the tables of Rouge linenumber-display code must not.

The introduction into code snippets of callouts is disruptive (since the code snippets are lexed, and the callouts need to be ignored by the lexer, and then restored as real XML markup.) Asciidoctor skips callouts in processing snippets, and I need to follow suite.

opoudjis commented 1 year ago

P.S. I'm using the Igor Pro theme of Rouge for CSS, which IMO is the best of a bad bunch—the Rouge styles are either dark background, which is a non-starter for print-like publication, or irritating pastels. https://spsarolkar.github.io/rouge-theme-preview/

We don't want to overdo colouring, but I don't think Igor Pro colours enough: it doesn't differentiate XML tags (class "nt") from content, for example. (See https://github.com/rouge-ruby/rouge/wiki/List-of-tokens for the classes defined by Rouge.)

We can make our own stylesheet, I'm just not convinced it's worthwhile.

We can import a stylesheet from Pygments, which Rouge has ported its CSS from, and Pygments has more themes: https://pygments.org/styles/ . But I'm not convinced that's worth it either.

See how you go with Igor Pro for now, at any rate.

jechterhoff commented 1 year ago

@opoudjis Can you give a hint as to when the recent changes (in this issue, as well as on the line numbering), are going to be available? Just checking to see if I can test today (which I'd like to do, but no rush) - two weeks of holidays ahead of me.

Attempting to upgrade metanorma via Chocolatey just now tells me I already got the latest version available there (1.6.7).

metanorma --version results:

Metanorma 1.5.3 Metanorma::Cli 1.6.7 Metanorma::Standoc 2.2.8/IsoDoc 2.3.6 Metanorma::ISO 2.2.4 Metanorma::Iec 2.1.13 Metanorma::IEEE 0.1.3 Metanorma::Ietf 3.0.14 Metanorma::Generic 2.2.5 Metanorma::BIPM 2.1.13 Metanorma::CC 2.1.13 Metanorma::Csa 2.1.13 Metanorma::IHO 0.6.13 Metanorma::M3AAWG 2.1.13 Metanorma::UN 0.9.13 Metanorma::Ogc 2.2.7 Metanorma::ITU 2.1.13

opoudjis commented 1 year ago

Tuesday. I do releases Monday night Australian time every fortnight, and next release is in 3 days.

opoudjis commented 1 year ago

You can test ahead of time if you set up your Gemfile to pull all of metanorma, metanorma-standoc and isodoc from github, but bad things might happen: i hold off the integration testing until release.

jechterhoff commented 1 year ago

Your last comment is scary enough for me to wait until January. :-) Thanks!

Intelligent2013 commented 1 year ago

@opoudjis do you have a real example of the source code with line numbers?

This example has wrong structure - first tr contains line number, and second one contains source code:

<table type="sourcecode">
<tbody>
<tr><td>1<br/>2<br/>3...</td></tr>
<tr><td><sourcecode>...</sourcecode></td></tr>
</tbody>
</table>

The source code sometimes doesn't fit to the page width and automatically carries on the next line, for instance: . In PDF there isn't the horizontal scrollbar for long lines.

Therefore we need to split each source code line into individual tr with line number, i.e.:

<table type="sourcecode">
<tbody>
...
<tr><td>3</td><td><sourcecode>&lt;title language="en" format="text/plain"&gt;Testbed-12 OWS Context User Guide&lt;/title&gt;</sourcecode></td></tr>
...

Intelligent2013 commented 1 year ago

Rendered PDF example with wrong line numbers:

opoudjis commented 1 year ago

As I've mentioned, there are two options for Rouge to do line numbers:

A table with one row for the entire snippet, with one cell for the line numbers, and <br/> to separate lines
A table with one row for each line of code

We have used the former currently.

The disadvantage of the former is that lines overflow and throw alignment off; Rouge actually warns about this in their readme.

The disadvantage of the latter is that you cannot cut and paste the snippet in isolation from the line numbers, something @jechterhoff explicitly said in #460 was desirable about the current Asciidoc pygments solution (as illustrated in https://docs.ogc.org/per/20-012.html#jsonschema_schemaconversionrules_types_classname_typeidentification :

<tbody><tr><td class="linenos"><div class="linenodiv"><pre> 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19</pre></div></td><td class="code"><pre><span></span><span class="tok-p">{</span>
  <span class="tok-nt">"$schema"</span><span class="tok-p">:</span> <span class="tok-s2">"http://json-schema.org/draft-07/schema#"</span><span class="tok-p">,</span>
  <span class="tok-nt">"definitions"</span><span class="tok-p">:</span> <span class="tok-p">{</span>
    <span class="tok-nt">"Type"</span><span class="tok-p">:</span> <span class="tok-p">{</span>
      <span class="tok-nt">"properties"</span><span class="tok-p">:</span> <span class="tok-p">{</span>
        <span class="tok-nt">"entityType"</span><span class="tok-p">:</span> <span class="tok-p">{</span>
          <span class="tok-nt">"type"</span><span class="tok-p">:</span> <span class="tok-s2">"string"</span>
        <span class="tok-p">},</span>
        <span class="tok-nt">"property"</span><span class="tok-p">:</span> <span class="tok-p">{</span>
          <span class="tok-nt">"type"</span><span class="tok-p">:</span> <span class="tok-s2">"string"</span>
        <span class="tok-p">}</span>
      <span class="tok-p">},</span>
      <span class="tok-nt">"required"</span><span class="tok-p">:</span> <span class="tok-p">[</span>
        <span class="tok-s2">"entityType"</span><span class="tok-p">,</span> <span class="tok-s2">"property"</span>
      <span class="tok-p">]</span>
    <span class="tok-p">}</span>
  <span class="tok-p">},</span>
  <span class="tok-nt">"$ref"</span><span class="tok-p">:</span> <span class="tok-s2">"#/definitions/Type"</span>
<span class="tok-p">}</span>
</pre></td></tr></tbody>

opoudjis commented 1 year ago

If we switch to the (clearly more correct) approach of a table row per code line, there are one workaround for in HTML, that would allow users to select only the code snippet, as shown in https://stackoverflow.com/questions/35738049/select-only-a-single-column-in-a-html-table:

The first solution suggested, https://stackoverflow.com/a/35815198, of nesting tables within tables, will not work, because the rows of the linenumber table and the rows of the codesnippet table will no longer be aligned.
The second solution is CSS user-select: none vs user-select: all.

The second solution https://stackoverflow.com/a/35738248 will indeed work, and is easy for us to insert; but it will only work on HTML, not in DOC. (And @Intelligent2013, I suspect there is no equivalent in PDF; I've certainly found mouse-select behaviour with tables in PDF to be so capricious, I never bother to do it.)

There are Javascript solutions that do this too, e.g. https://stackoverflow.com/a/6619995, but the CSS option is simpler, and we're avoiding JS when we can.

@jechterhoff Are you ok to proceed with a solution that works only for HTML in making copy-paste easy, but still keeps line numbers aligned?

Intelligent2013 commented 1 year ago

I've tried manually split each source code line into tr:

<table type="sourcecode">
<tbody>

<tr>
<td>1</td>
<td><sourcecode lang="ruby"><span class="nb">require</span> <span class="s2">"isodoc/ogc/html_convert"</span></sourcecode></td>
</tr>

<tr>
<td>2</td>
<td><sourcecode lang="ruby"><span class="nb">require</span> <span class="s2">"isodoc/ogc/pdf_convert"</span></sourcecode></td>
</tr>

<tr>
<td>3</td>
<td><sourcecode lang="ruby"><span class="nb">require</span> <span class="s2">"isodoc/ogc/word_convert"</span></sourcecode></td>
</tr>
...

and tried to set fox:header="true" for 1st column with line numbers (found here https://xmlgraphics.apache.org/fop/2.0/accessibility.html#fox:header):

<fo:table-column column-width="8%" fox:header="true"/>
<fo:table-column column-width="92%"/>

but it does not affect on the selection order - both columns selected together:

I don't an ideas how to solve it. May be. again, Apache FOP Intermediate Format post-processing, but if this task is actually a problem...

opoudjis commented 1 year ago

Let's not make life unnecessarily difficult for you.

People truly do expect weird stuff to happen if they copy paste a table from PDF. If there isn't a quick and easy solution for PDF, we'll ignore it. It's a nice to have, given that expectation. There is at least a workable solution for HTML, which does not involve too much effort.

opoudjis commented 1 year ago

Needing to change behaviour of converting carriage return to <br/> in sourcecode; we don't do so if the sourcecode has already been converted into a table with a row per line.

opoudjis commented 1 year ago

Will finalise and upload tomorrow; there was more to fix here than I'd expected, both with tabular representation of code lines, and with callouts processing, and especially with CSS manipulation to smooth over the introduction of the table.

opoudjis commented 1 year ago

Callouts and annotations should be styled in comment style.

Word code tables need styling and debugging.

I'm really not in love with Igor Pro, and will export it as an import in isodoc, which can be overruled locally.

opoudjis commented 1 year ago

Localised IgorPro to add .nt: XML tags

https://coolors.co/444444-cc00a3-ff0000-c34e00-0000ff-007575-009c00 for list of colours, https://github.com/rouge-ruby/rouge/wiki/List-of-tokens for list of classes

Onyx = Grey #444444
Byzantine = Magenta #CC00A3
Skobeloff = Teal #007575

Token name	Token shortname	Description	Current stylesheet
Text		Any type of text data	Onyx
Text.Whitespace	w	Specially highlighted whitespace	Onyx
Error	err	Lexer errors
Other	x	Token for data not matched by a parser (e.g. HTML markup in PHP code)
Keyword	k	Any keyword	Blue
Keyword.Constant	kc	Keywords that are constants	Burnt Orange
Keyword.Declaration	kd	Keywords used for variable declaration (e.g. var in javascript)	Blue
Keyword.Namespace	kn	Keywords used for namespace declarations	Blue
Keyword.Pseudo	kp	Keywords that aren't really keywords	Blue
Keyword.Reserved	kr	Keywords which are reserved (such as end in Ruby)	Skobeloff
Keyword.Type	kt	Keywords wich refer to a type id (such as int in C)	Blue
Name	n	Variable/function names
Name.Attribute	na	Attributes (in HTML for instance)
Name.Builtin	nb	Builtin names which are available in the global namespace	Burnt Orange
Name.Builtin.Pseudo	bp	Builtin names that are implicit (such as self in Ruby)	Burnt Orange
Name.Class	nc	For class declaration
Name.Constant	no	For constants
Name.Decorator	nd	For decorators in languages such as Python or Java
Name.Entity	ni	Token for entitites such as in HTML
Name.Exception	ne	Exceptions and errors (e.g. ArgumentError in Ruby)
Name.Function	nf	Function names
Name.Property	py	Token for properties
Name.Label	nl	For label names
Name.Namespace	nn	Token for namespaces
Name.Other	nx	For other names
Name.Tag	nt	Tag mainly for markup such as XML or HTML	Blue added
Name.Variable	nv	Token for variables
Name.Variable.Class	vc	Token for class variables (e.g. @@var in Ruby)
Name.Variable.Global	vg	For global variables (such as $LOAD_PATH in Ruby)
Name.Variable.Instance	vi	Token for instance variables (such as @var in Ruby)
Literal	l	Any literal (if not further defined)
Literal.Date	ld	Date literals	Slimy Green added
Literal.String	s	String literals	Slimy Green
Literal.String.Backtick	sb	String enclosed in backticks	Slimy Green
Literal.String.Char	sc	Token type for single characters	Slimy Green
Literal.String.Doc	sd	Documentation strings (such as in Python)	Slimy Green
Literal.String.Double	s2	Double quoted strings	Slimy Green
Literal.String.Escape	se	Escaped sequences in strings	Slimy Green
Literal.String.Heredoc	sh	For "heredoc" strings (e.g. in Ruby)	Slimy Green
Literal.String.Interpol	si	For interpoled part in strings (e.g. in Ruby)	Slimy Green
Literal.String.Other	sx	Token type for any other strings (for example %q{foo} string constructs in Ruby)	Slimy Green
Literal.String.Regex	sr	Regular expressions literals	Slimy Green
Literal.String.Single	s1	Single quoted strings	Slimy Green
Literal.String.Symbol	ss	Symbols (such as :foo in Ruby)	Slimy Green
Literal.Number	m	Any number literal (if not further defined)
Literal.Number.Float	mf	Float numbers
Literal.Number.Hex	mh	Hexadecimal numbers
Literal.Number.Integer	mi	Integer literals
Literal.Number.Integer.Long	il	Long interger literals
Literal.Number.Oct	mo	Octal literals
Literal.Number.Hex	mx	Hexadecimal literals
Literal.Number.Bin	mb	Binary literals
Operator	o	Operators (commonly +, -, /, *)
Operator.Word	ow	Word operators (e.g. and)
Punctuation	p	Punctuation which is not an operator
Comment	c	Single ligne comments	Red
Comment.Multiline	cm	Mutliline comments	Red
Comment.Preproc	cp	Preprocessor comments such as <% %> in ERb	Byzantine
Comment.Single	c1	Comments that end at the end of the line	Red
Comment.Special	cs	Special data in comments such as @license in Javadoc	Byzantine
Generic	g	Unstyled token
Generic.Deleted	gd	Token value as deleted
Generic.Emph	ge	Token value as emphasized
Generic.Error	gr	Token value as an error message
Generic.Heading	gh	Token value as a headline
Generic.Inserted	gi	Token value as inserted
Generic.Output	go	Marked as a program output
Generic.Prompt	gp	Marked as a command prompt
Generic.Strong	gs	Mark the token value as bold (for rst lexer)
Generic.Subheading	gu	Marked as a subheadline
Generic.Traceback	gt	Mark the token as a part of an error traceback
Generic.Lineno	gl	Line numbers

opoudjis commented 1 year ago

@Intelligent2013 just to ensure that we're aligned, please find enclosed updated OGC Presentation XML.

Αρχειοθήκη 5.zip

ronaldtse commented 1 year ago

There is a problem with using a <table> for source code: when people try to copy paste it, the result on the clipboard will include those line numbers.

If you have used GitHub's file compare feature, you notice that a copy and paste does not include the line numbers. Which is the correct behavior. This desired behavior applies to all output formats, including HTML and PDF.

If the line numbers do exist, they should be purely presentational -- they are NOT part of the content, and we should not pollute source code content with tables.

but it will only work on HTML, not in DOC

OGC only cares about HTML and PDF. Technically, OGC should not care about DOC output.

opoudjis commented 1 year ago

We've discussed this offline. The line numbers are indeed limited to the Presentation XML, as is the syntax colouring. The Semantic XML contains just the source code. The HTML is preventing line numbers being copied, without resorting to Javascript. I am not optimistic there is any solution that will work with PDF.

Intelligent2013 commented 1 year ago

@Intelligent2013 just to ensure that we're aligned, please find enclosed updated OGC Presentation XML.

Αρχειοθήκη 5.zip

@opoudjis thanks, it's ok for me:

Intelligent2013 commented 1 year ago

I have an idea how to omit line numbers from copy-pasting from PDF. I'll try to render the line numbers as mathml in SVG (or just SVG).

Intelligent2013 commented 1 year ago

Line numbers on-fly in XSLT converted to mathml:

<math xmlns="http://www.w3.org/1998/Math/MathML">
    <mtext>1</mtext>
</math>

and text selection is working for source code only now:

Mathml text renders a bit bolded than main text, so I will investigate it...

Example: a2.presentation.pdf

opoudjis commented 1 year ago

... I am in complete awe of you @Intelligent2013

Releasing tomorrow, am going on holiday for a week.

Intelligent2013 commented 1 year ago

Mathml text renders a bit bolded than main text, so I will investigate it...

I think the issue in jEuclid and Batik. I'll try to investigate it in a few hours, but if can't then keep SVG source code line numbers in a bit bolded.

Intelligent2013 commented 1 year ago

I think the issue in jEuclid and Batik. I'll try to investigate it in a few hours, but if can't then keep SVG source code line numbers in a bit bolded.

I'll solve this issue in https://github.com/metanorma/mn2pdf/issues/117, need more time. Current result is https://github.com/metanorma/metanorma-ogc/issues/465#issuecomment-1366775899.

opoudjis commented 1 year ago

I'm doing release now; those results are acceptable for the time being.

jechterhoff commented 1 year ago

@opoudjis You asked:

@jechterhoff Are you ok to proceed with a solution that works only for HTML in making copy-paste easy, but still keeps line numbers aligned?

Yes, that would be ok for me. If I understand the conversation correctly, a solution for PDF format may have been found in the meantime. :-) Sorry for my late reply. I've been on holiday for two weeks.

jechterhoff commented 1 year ago

@opoudjis May I ask when / if this enhancement is released for windows users that installed metanorma using Chocolatey?

Executing choco upgrade metanorma -y did not update any package.

Result of metanorma --version is:

Metanorma 1.5.3 Metanorma::Cli 1.6.7 Metanorma::Standoc 2.2.8/IsoDoc 2.3.6 Metanorma::ISO 2.2.4 Metanorma::Iec 2.1.13 Metanorma::IEEE 0.1.3 Metanorma::Ietf 3.0.14 Metanorma::Generic 2.2.5 Metanorma::BIPM 2.1.13 Metanorma::CC 2.1.13 Metanorma::Csa 2.1.13 Metanorma::IHO 0.6.13 Metanorma::M3AAWG 2.1.13 Metanorma::UN 0.9.13 Metanorma::Ogc 2.2.7 Metanorma::ITU 2.1.13

opoudjis commented 1 year ago

isodoc is currently on 2.4.2, and metanorma-standoc on 2.3.6. The version you've just updated to dates from November.

@ronaldtse Is choco still being kept up to date?

ronaldtse commented 1 year ago

@opoudjis the Windows package is failing to build because of the recent addition of Rouge, which is broken on Windows.

@CAMOBAP is fixing it at:

CAMOBAP commented 1 year ago

FYI missing 1.6.11 release for windows is in progress now https://github.com/metanorma/chocolatey-metanorma/actions/runs/3862089946

jechterhoff commented 1 year ago

Thanks. The upgrade to 1.6.11 was successful.

The weird behavior, that syntax highlighting for JSON did not always work, no longer occurs in my document.

metanorma / metanorma-ogc

source blocks: syntax highlighting for JSON does not always work #465