| Age | Commit message (Collapse) | Author |
|---|
|
|
|
take it into account
|
|
crude heuristic
What we need to do is two things:
1. when a new link is supplied, go download it and let the browser tell
us if it is an image and if so what are it's dimensions
2. cache that data in the db (on the to-be-added view node).
Actually
3. cache the image itself. This will also allow uploading images with no
link.
Where to cache the images - on disk with a hash, plus breakdown into
multiple directories to avoid filesystem limits.
|
|
|
|
|
|
and no interaction possible yet
|
|
|
|
|
|
the suggestions from textanalysis now come as a dictionary with
{nodes: {name: id},
links: {name: id}}
(before it was a single dictionary, {name:id})
And the options bus for the textanalyser and search completion have been
updated accordingly. textanalyser also takes into account the cursor
position to determine which of the option dictionaries to use.
|
|
required for completion updates)
|
|
|
|
third-internship-proposal removal
|
|
when shift pressed
|
|
|
|
|
|
|
|
is ugly
Together with addition of new attributes fixes #357
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
from declaration
|
|
|
|
|
|
At the same time lets completion work with spaces in the graph entry
bar. In the search bar it still doesn't work, will require some thinking
- how to let a space be both and/or and part of a name. It's better to
ditch it as an operator and use |/& or AND/OR.
|
|
needs more complex logic to account for user changes that have not
round-tripped yet, so for now just do not update the dialog.
Fixes #349
|
|
|
|
|
|
|
|
completer, plus some cleanup
|
|
|
|
|
|
|
|
Killed some dead code in the process.
|
|
|
|
|
|
dfc93cf003e8b4484e98ead7f7eac63ad10182e7
|
|
Source for inspiration / copying:
http://jsfiddle.net/brx3xm59/
http://stackoverflow.com/questions/9308938/inline-text-editing-in-svg
comment 2 by John Ktejik
|
|
|
|
before Bacon
|
