Allow more control over HTML export
Submitted by Matti Aarnio
Gnumeric version 1.1.16 on Debian stable branch.
In spreadsheet I have a bit of text with special characters, like: µ ° ½ (micro, degree, half) Occasionally I do use also various umlauts (ä Ä ö Ö) as is the style in my native language.
When a spreadsheet is exported with "Save As -- HTML fragment" to be then included in a web-page, for example 'µ' becomes exported as: Âµ which may be correct two-byte sequence for UTF-8, but is not in general correct HTML entity, especially in web-page containing ISO-8859-15 encoded data.
Entity reference document: http://www.w3.org/TR/html4/sgml/entities.html#h-24.2
Another matter is the recent trend to start adding style= attribute with explicite font size specification:7`` which (again) in included HTML fragment is dubious in extreme, as it breaks the overall web-page style definitions, which do expect the HTML content to use whatever top-level definitions are.
Third nuisance thing is Gnumeric's long-standing "helpfull" style to span multiple columns of box text on screen happens to overflow the box. In exported HTML automatic folding within the box would handle the issue just fine, but alas, this exporting requires that I carefully manipulate the screen table to have enough column width to contain all text in it. Explicitely merged cell ranges are to be exported as defined, of course.
I would be happy, if there are multiple HTML-fragment exporting mechanisms. Some which make as WYSIWYG as possible of the result (for those who want such things), but always also one which makes as plain as possible. What are ok, are: cell background/font color setting, font style (normal, italic, bold), text alignments within the cell. Leaving away all font-size tricks, and automatic column spans.
My aim is to calculate things in Gnumeric, and produce HTML 4.01 compatible HTML to be included as a part of a larger page.