import csv containing html link errors

[flamerohr]

Hi, I've got a csv file which contain html content. Most import fine, except for when the html contain <a href="http://example.com">Click here</a>
When that happens, I get returned the error File extension does not match known MIME type

[tractorcow]

The problem this is caused by is HTML content within CSV causing the file to be mis-classified by the mime validator.

A possible fix for this is to allow file extensions to declare they allow wildcard mime types, so that text/* can be assigned to extensions for known plain-text types (e.g. CSV). This is necessary because specific matching of plain text file types to mimes is always going to be inprecise, and in most cases any mis-match would be within reasonably acceptable limits.

[NightJar]

Is there an example CSV document that could be used for testing?
What kind of ratio of CSV:HTML are we talking here?
How 'correctly' formatted is the CSV? I mean, I know it's a bit of a 'flakey' non-standard as it is, but generally speaking...

[flamerohr]

As long as the CSV contains any HTML, it would be considered an HTML file - rather than CSV.
If a small amount doesn't work, try push it to a 50% ratio of HTML

[fspringveldt]

I encountered this when a DataObject I exported to CSV using ModelAdmin contained a DBHTML field with HTML content.
Trying to re-import that file then silently fails and closes the Dialog. It's not until you click Import CSV again and the error is revealed

After digging I noticed that yes, indeed the mime-type was text/html and not text/plain as expected because of the presence of HTML.

Anyway, my workaround was to just allow it in config.

SilverStripe\MimeValidator\MimeUploadValidator:
  MimeTypes:
    csv:
      - 'text/html'
      - 'text/plain'