ODT Emitter always uses text:anchor-type="paragraph" for images regardless of display setting

[SR-Cochise]

Description

ODT knows three different ways to anchor image within a document:

1. "paragraph"
2. "at-character"
3. "as-char"

Regardless of display type in the layout of a rptdesign the ODT emitter always creates xml like this:

<text:p>
    <draw:frame draw:style-name="Draw1" draw:name="Image75" text:anchor-type="paragraph" svg:width="0.10416666666666667in" svg:height="0.10416666666666667in" draw:z-index="0">
    <draw:image xlink:href="Pictures/Image1.png"/><svg:title/></draw:frame>
</text:p>

This is no problem for MS Word or Open/Libre Office but Google Docs ignores / removes images that aren't anchored as "as-char".
Minimal solution would switch between

• text:anchor-type="paragraph" for display type "block"
• text:anchor-type="at-char" for display type "inline"

Optimal solution would probably have to use an advanced / user-property that knows all three types of anchoring.

Environment

• OS: any
• Browser: any
• BIRT Version: 4.15+ (oldest tested)

[speckyspooky]

@SR-Cochise
Could you please add a short demo-report with the problem based on an embedded image and without external dependencies. My first try was ok for me. But I think I used a completely different structure like you.

And a screen of the missing/wrong image would be good and what get with the changed attribute. So we woul have a good compare option after the change.

[SR-Cochise]

@speckyspooky
Here is a ZIP file with

• Report file: ExampleReportForODT.rptdesign
• ODT file 1: ExampleReportForODT.odt (created with BIRT 4.15 from the report file)
• ODT file 2: ExampleReportForODT_hacked.odt (manually extracted the content.xml from the previous file, replaced text:anchor-type="paragraph" with text:anchor-type="as-char" and re-zipped
• screenshot 1: ODT_inGoogleDocs_noHack.png -> how file 1 looked in Google Docs (after "minor" manual corrections for page / line breaks)
• screenshot 2: ODT_inGoogleDocs_hacked.png -> how file 2 looked in Google Docs (after "minor" manual corrections for page / line breaks)

When unzipping the file 1 you'll note that all pictures are written into content.xml with text:anchor-type="paragraph" regardless of how the images are positioned in the body of the report. This anchor type causes Google Docs' importer to omit entirely or remove the image from the visible areas.

(along with other stupid things it shouldn't do based on the report defintion and the odt file layout but that's a different story).

ODT_Examples.zip