Entering the entities you wish to preserve in the Preserve Entities prefernces certainly works for quite a few people (myself included). There are some rare entities that it doesn't work for, but emdash is certainly one that works for me and many others.
For the record, the gumbo parser does not speak "entity," so unless the entities are properly entered in the Preserve Entities list first, they will all be converted to their unicode characters equivalents. This is just how Sigil works now.
But this is this first I've ever heard of entities being re-"entitied." The only time I've ever seen anything like that is if someone tries to type (or paste) an html entity/tag into Book View. If that's the case, then Sigil is correct in changing — to &mdash; in Code View. Just like a typing an html tag (like <h2>) in Book view will result in <h;2;> in Code View.
Can you provide a small sample (non-copyrighted) epub and the documented steps to reproduce what you're seeing? That would be very helpful. Thanks.
Quote:
Originally Posted by sinclair
Hi everyone,
Sorry if this has already been reported elsewhere, but I seem to have found a serious bug that is making Sigil pretty much unusable for producing a clean epub. I'm using version 0.8.901 on Windows 10 and it keeps messing up the html escape characters in my html documents. For example, when I open my html file in Sigil it changes an emdash ( —) into &amp;mdash; Unfortunately it's also doing the same to single double quotes and basically every html escape I have in my html file.
If I go into 'Preferences' and then 'Clean Source' and untick both the 'Open' and 'Save' boxes for automatically cleaning the html source, it makes no difference. If I correct the html escape codes and then save and close the epub, when I reopen it, it once again messes up all the html escape characters. Similarly, if I add all the html escape characters to the 'Preserve Entities' box that doesn't seem to make any difference either. Hope this information is helpful.
Paul
|