+
+>> [[!template id=gitbranch branch=chrysn/imgforpdf-and-more author="[[chrysn]]"]]
+>>
+>> [[!tag patch]]
+>>
+>> i've implemented a fix for this along with a unit test.
+>>
+>> the patch branch is based on the imgforpdf branch
+>> ([[bugs/svg and pdf conversion fails]]), because it would not cleanly merge.
+>> the branch also enhances on how images are handled in preview, falling back
+>> to data: urls if the image has not been rendered in a saved version. please
+>> review. --[[chrysn]]
+
+>>> Mostly [[looks good to me|users/smcv/ready]].
+>>>
+>>> Minor things, which wouldn't stop me merging it if I could:
+>>>
+>>> * `$imgdatalink = "data:image/".$im->Get("magick").";base64,".encode_base64($blob[0]);`:
+>>> is the ImageMagick file type always valid as the second part of
+>>> a MIME type?
+>>> * In this code:
+>>>
+>>> +open (my $outhtmlfd, "<", "$outpath.html");
+>>> +local $/=undef;
+>>> +my $outhtml = <$outhtmlfd>;
+>>> +close $outhtmlfd;
+>>>
+>>> no block is closed, so the "local" is ineffective, so the `<>` operator
+>>> remains in read-entire-file mode afterwards. To avoid odd side-effects,
+>>> I would suggest using `readfile()` like `t/trail.t` does.
+>>>
+>>> [[!template id=gitbranch branch=smcv/ready/imgforpdf-and-more author="[[chrysn]], [[smcv]]"
+ browse=http://git.pseudorandom.co.uk/smcv/ikiwiki.git/shortlog/refs/heads/ready/imgforpdf-and-more]]
+>>> I've used `readfile()` (but not done anything about the ImageMagick file type)
+>>> in my copy of the branch.
+>>>
+>>> --[[smcv]]