The docs for JSONPath Assertion state, that it will fail the assertion,
when no element is found with the given JSON path. This was currently
not followed, when an indefinite path was used. In such a case, the
JSONPath library would return an empty list, which the assertion logic
would log as OK.
With this change we let the assertion fail, when
* an indefinite path was given
* and an empty list is extracted
* and no assertion value is given
Bugzilla Id: 65794
Make sure, that we scale away from zero (which is in our case mostly up).
Otherwise we might end up with a scaling factor of zero, which would
result in really small images.
Bugzilla Id: 65784
* Switch in LazyLayeredConnectionSocketFactory https.socket.ciphers to https.cipherSuites as the latter is documented
* Use https.cipherSuites, https.socket.protocols in Proxy (HTTP(S) Test Script Recorder)
* Introduce JMeterUtils.getArrayPropDefault
Rename based on Vladimir proposals
This resolves https://bz.apache.org/bugzilla/show_bug.cgi?id=65692
It looks like there's an open IDEA issue regarding the use of continuation indent size.
Kotlin official style guide recomments using *single* indent for continuations,
however, IDEA wants to use double indent for some reason.
So far ij_kotlin_continuation_indent_size = 4 workaround seems to be sufficient
to make IDEA and ktlint format in the same way.
https://youtrack.jetbrains.com/issue/KTIJ-17907https://github.com/pinterest/ktlint/issues/805
When we extract a null value for a given expression, we now replace that
null with the given default value for the expression.
That leads to slight change of the meaning of a missing value.
Consider the JSON structure: {"context": null} and the JSON Path: "$.context"
The new implementation will replace the `null` with the default value, while
the old would have stringified it to the empty string.
Cleanup of parameter handling of default values. We should use the default
value for the current handled expression and not pass all default values
for all expression. That leads to confusion, only. (Well, I was confused)
Bugzilla Id: 65681
There are two fixes:
1) undoManager.setLimit(0) -> setLimit(1) since (0) means "unlimited undo"
See https://github.com/bobbylight/RSyntaxTextArea/issues/99
2) By default, JMeter adds undoManager to ALL text fields via Swing updateUI method,
so we need to explicitly uninstall it for the case when undo is not needed
Now it should be possible to specify an IPv6 address with the bracketed notation:
remote_host=[::1]:2001
Based on a patch by Peter Wong
Closes#673
Bugzilla Id: 65611
Currently a user can not differentiate between a wrong password and missing or
wrong keys, algorithms, etc. So try do handle the excpetions, that the API
gives us and log different error messages to better inform our users.
Part of #654
The reported problem was fixed with the patch (that is reverted here), but the Tree Results View
was broken with it. So revert it for now.
This reverts commit 2e7c471734.
This reverts commit a7bdcb24a3.
Bugzilla Id: 65545
Try to re-combine the parameters to regard spaces in file names
Java command line has no knowledge of the 'real' parameters and
we have to do a bit of guessing to re-assemble the parameters with
spaces and the drop the spaces, that should split the parameters.
So we guess, that each parameter starts with a dash (-) and
everything else are values, that should be stitched together.
Bugzilla Id: 65522
Compare JSON objects and not their string representations.
When using our stringifier, the order of the entries in maps
is not guaranteed and can lead to wrong results.
In the old days we made no difference between a string or int
when asserting a result. Jackson JSON Parser differentiates
between 'foo' and '"foo"' (former is invalid) and '1' and '"1"'
(former is an int, latter a string).
To enable both (complex, simple and edge cases), we now have to do
more work.
Bugzilla Id: 65299
Felix Schumacher
Sampath Kumar Krishnasamy
Piotr Smietana
Vladimir Sitnikov
Magnus Spångdal
Woonsan Ko
Chromico Rek
Philippe M
Stefan Seide
Rithvik Patibandla