XML CDATA minify issue

[amoledwatchfaces]

<![CDATA[ %d ]]> where empty space is before and after %d results in " %d" - space after %d is stripped.

Can this be improved in the future so CDATA maintains white spaces more properly?

[tdewolff]

In the general case, XML whitespace should be preserved as it is up to the parser/application to decide if it's significant (e.g. whitespace at the beginning or end of a tag). In most cases it is not significant, much like HTML, which has been the default for the XML minifier. You should enable the KeepWhitespace option to keep all whitespace at the start/end of a tag's content.

[amoledwatchfaces]

@tdewolff

I've tried KeepWhitespace but it still removes empty spaces at the end of the CDATA content.

Can I somehow force this inside of the CDATA?

Unfortunately, my parser accepts empty spaces at the end and this is viable for my usecase.

[tdewolff]

Do you have a small but complete XML example? This shouldn't happen inside a tag.

[amoledwatchfaces]

Do you have a small but complete XML example? This shouldn't happen inside a tag.

@tdewolff Here, test.xml, test-minify.xml & test-minify-keepwhitespace
test-xml.zip

test.xml

test-minify.xml (removes spaces completely)

test-keep-whitespace-minify (removes space at the end)

[amoledwatchfaces]

I would expect from simple minify to output this?

[tdewolff]

I believe that test-minify.xml outputs as expected. We can debate if the KeepWhitespace option should be on or off by default though.

The test-keep-whitespace-minify.xml is also as I expect. There is a whitespace before and after the %s part: before it is a space 0x20, and after it is a newline 0x0A. Both are valid whitespace. Perhaps we shouldn't collapse multiple whitespace as we do in HTML?

The original content between <Template> and <Parameter ... is: [space]%d[space][newline][tab][tab][tab][tab][tab][tab]. You say you only expect the spaces to be kept but all other whitespace to be removed, or do you wish to remove all whitespace? Then what is the purpose of minifying?

[amoledwatchfaces]

You say you only expect the spaces to be kept but all other whitespace to be removed

This. Application that parses this xml needs to maintain all spaces correctly inside CDATA. This is used in a watch face renderer.
So after minifying CDATA should keep all spaces correctly [0x20][%d][0x20].

Newline renders wrong result unfortunately. Perhaps there could be --xml-keep-cdata ?

minify --xml-keep-cdata which will remove all whitespaces as single 'minify' but keeps CDATA and its spaces correctly:

[tdewolff]

I've added a fix that keeps all whitespace inside CDATA as-is.

[amoledwatchfaces]

Thanks! Two coffees on me ;)

[tdewolff]

Awesome, thank you very much!!

[tdewolff]

I've released v2.20.36 with this fix!

[amoledwatchfaces]

I've released v2.20.36 with this fix!

Thanks! 🙂