This is a very hacky way of doing it: a base path of /seance will
lead to rendering in,
public/seance/seance/css/styles.css
but it seems to be the only way out for now 🙁
Hopefully in future we'll come up with a more elegant solution.
This gives us more flexibility: after all how often does one have
a WebDAV server handy (unless you're smart enough to use HelioHost
or some other awesome server)?
These options make Seance go through the motions of processing a
post without actually uploading images and/or pushing them to
Ghost. This makes it easier to test out stuff for new features
(or, more accurately, it reduces the need to clean up afterwards!)
The long-standing "make title proper" bug has finally been fixed.
Yayy! So now, basically, the title and subtitle gets set properly
even if Medium messed it up by showing some of them twice.
Now provides a reminder to fix the title/subtitle/excerpt, and a
helpful URL of the final post destiation to publish on Buffer!
@manasakashi12 hope you find this useful!
For some reason, loading /fetch runs the fetch page, but then
ends up rendering the / page instead. Can't figure it out, but
this solves it for now so let's go with it till we come up with a
proper solution :P
If you're on localhost, your protocol (http vs https) takes
precedence. If not, the other website's takes precedence, because
otherwise you could end up being blocked by security
Earlier, it used to print out the stuff to the screen, which you'd
then have to copy to a file manually - tedious!
By the way, the other fixes were typos and omissions while changing
Seance from a simple file to an object.
We're using the Promise/async version of file read/write commands,
so the script doesn't keep blocking or hanging around but can
keep doing other stuff till the file reading or writing is over
This way, if you have the JSON already there locally, you don't
have to bother downloading it from online. Especially useful if
the JSON is passed to you some other way, like through a browser
for example
Now, the Seance object is what everyone operates with. That way,
we can do fancy stuff like emitting events, etc. if the need so
arises (as it will, very soon; trust me).
It was looking for a (nonexistent, obsolete) variable to check if
the "config" command is running. Now, it just looks at the
command-line options directly and does it.
This is still ungraceful, in the sense it'll fail spectacularly
later, but at least it gives us the opportunity to run the new
`setup` command first.
Where's that command, you say? Just wait for the next commit! ;-)