Add image scaling attributes to block image in DocBook backend

[mojavelinux]

Add the following image scaling attributes to the block image output in the DocBook backend:

• scaled
• scaledwidth (assumed to be a %)

When the scaledwidth attribute is specified, the scalefit="1" attribute should be added to the <imagedata> tag as well.

Example:

image::sunset.jpg[Sunset, scaledwidth="75%"]

Output:

<informalfigure>
<mediaobject>
<imageobject>
<imagedata fileref="sunset.jpg" width="75%" scalefit="1"/>
</imageobject>
<textobject><phrase>Sunset</phrase></textobject>
</mediaobject>
</informalfigure>

Note that in Python AsciiDoc, it's necessary to define the alt attribute explicitly when any other attribute is present. For example:

image::sunset.jpg[alt="Sunset", scaledwidth="75%"]

[mojavelinux]

Open Question: Should we assume the scaledwidth value is a percentage? Or should we leave it open for the writer to set the units?

[lordofthejars]

I think that % should be the default and it is a good practice to use
percentatges, but supporting other units Can be useful for some people, if
it is easy I think we should support it

El divendres 1 de novembre de 2013, Dan Allen ha escrit:

Open Question: Should we assume the scaledwidth value is a percentage?
Or should we leave it open for the writer to set the units?

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/763#issuecomment-27540066
.

Enviat amb Gmail Mobile

[mojavelinux]

Super. I added that feature.

[nerk]

There still seems to be a compatibility issue with original asciidoc. If I remember correctly, original asciidoc allows to use both attributes, width and scaledwidth at the same time. width will only apply to HTML output, scaledwidth to docbook output.

I am able to write

image::someimage.png[width="500", scaledwidth="70%"]

This is different in Asciidoctor. scaledwidth only applies to docbook output, but width applies to both, HTML and docbook. This causes such images not being scaled properly in generated PDF.

Apparently, the backend sets contentwidth and contentheight attributes based on width and height parameters.

As a workaround, I could use some conditional macros, but I am just wondering if this is how it should be.

[mojavelinux]

I'm pretty sure Asciidoctor and AsciiDoc Python have the same behavior. However, it's possible that in both cases it's incorrect.

You can see that AsciiDoc Python uses width and height if present, even if scaledwidth is present.

https://github.com/asciidoc/asciidoc/blob/master/docbook45.conf#L88

If you think a change is warranted, let's move forward with a separate issue that proposes the change.

(not that this may be an Asciidoctor PDF issue because I know that scaling isn't working quite as we expect there).

[nerk]

Yes, you are correct! I compared the docbook outputs asciidoc and asciidoctor and they are the same. Must have been something else, maybe related to #1059 and the FOP backend I am using.

Scaling to different media is quite a bit of PITA...

[mojavelinux]

Scaling to different media is quite a bit of PITA...

I agree. And PDF doesn't make reasoning about it easy in my experience.