Age | Commit message (Collapse) | Author | Files | Lines |
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
This allows for easier testing.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
This allows the tests to run somewhat parallel (still limited by the
database) and simplifies running single testcases without the
./run-tests.sh wrapper.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Also don't print the return data of the writer. In the case of clover
it's the XML content, with the HTML writer it's nothing.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
According to the RFC this is only useful for services that use HTTP's
built in authentication schemes. We don't so we can't use this code.
References: https://tools.ietf.org/html/rfc7235
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Chromium apparently requires a <br> or some text to be present for a
line to be copied to the clipboard.
Firefox interprets the \n at the end as line break and copies that if
there is no content. Adding an explicit <br> makes chromium work
correctly, but firefox still converts the \n. Fix this by putting the
\n in an HTML comment. That way the output is still somewhat readable,
but both browsers do the correct thing.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Apparently firefox doesn't like html code in <pre> and adds incorrect
linebreaks when copying the code. This does not happen with chromium.
Fix it by switching to <code> instead of <pre>. Also change the CSS to
refer to the element by class instead of by type so future changes are
simpler.
Not sure why, but in chroium changing the container element type breaks
the ordering of the lines. Adding display: table to the .code-container
fixes this.
References:
http://stackoverflow.com/questions/24851173/why-does-text-copied-from-a-set-of-pre-tags-have-double-line-breaks#comment41552816_26408656
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
We actually don't need to hide this from the user. The error should be
shown rather than a blank page being returned.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Already in \libraries\Exif
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
This bypasses require.js for asciinema because apparently require.js
can't deal with the embedded React library. The result is that React is
not loaded properly and the asciinema player is unable to find it.
Just load it like upstream tells you to.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Keep this field private for now. I may be exposed at a later time.
The returned values are already normalized so they can be safely
exposed.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
The documentation already refers to 2.1.0 and lists a new feature of
the file/history endpoint, but requests for 2.1.0 are not yet accepted
because the server doesn't know that it actually supports this version.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Previously the login box in the navigation would redirect to the
current page, but this page will throw an error in the case of the
registration page since that's the page with the invition key and that
key is no longer valid.
Fix this by redirecting to the $redirect_uri and ensure that this value
is set for all requests.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Clients are only hosted on paste.xinu.at and everywhere else the links
will point to missing files so they become useless.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
The - from m-ID was missing in the regex.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Also adjust test cases to check for the new value. API v1 does not
change.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Since this is a breaking change bump the api version to 2.
The private fields are user_id and multipaste_id which where leaked via
the multipaste_items field. This commit also adds a test case to both
api versions that checks the returned fields.
NOTE: Most of this commit is copied from the files of api v1 so when
viewing the diff use --find-copies-harder for an easy to read diff.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
These are necessary for clients that want to send lots of fields e.g.
in a delete request or upload multiple files in one upload request.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
If the user has multiple tabs open, but is not logged in this will be
called multiple times (unless he logs in after the first upload batch)
and earlier uploads would be lost.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
If we store only the last called URI in the session we can't support
multiple browser tabs that all need to log in again. Fix this by
storing the URI in the URL.
Also change a trim() to ltrim() so that the URI string we store keeps
it's trailing slash.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Also try to clean up when files are deleted since 1 month is a rather
long time. Granted, thumbnails are small, but whatever
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
- Use the filedata we already have in
c/file->upload_history_thumbnails() rather than fetching it per id in
m/mfile->valid_id
- Construct the config array for s/f::valid_id only once and not for
every validation.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
The user might click the wrong link or might change their mind and it
is really not necessary to clean them up early. Let the cron job take
care of that.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
This happens when an invalid lexer is used so until we use the exceptions
this should not be forbidden since we won't fall back to plain text
then.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
No need to specify the interpreter twice.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|