joao-carloto / firerobot Goto Github PK

1. Click somewhere to select a resource file and automatically add it's path to the Resources section.
2. Every time we add steps to the script, FireRobot would check if part of the script corresponds exactly to some existing keyword and automatically replace the steps (if such option is enabled).

This might be difficult to implement, not too reliable and possibly on the heavy side in terms of processing. But it's worth looking into it.

Not working properly in OSx.
Namely context menu does not show up.

Reasons:

• The Sel2lib-java (with all dependencies packed) is very big (54MB) and bloats the extension a lot.
• It does not support the ff_profile_dir parameter in the Open Browser keyword, and I would like to build a FireRobot feature on top of this.
• The Python Sel2Lib project seems to be bigger and with possibly a better long term support.

Options:

• Give up on the "Out of the box" concept and rely on people having Python, RF and Sel2Lib already installed (make link to RF quick start guide). This option would also imply giving up on RF jar and make the extension small enough to apply for the Mozilla repository. This would also make tests faster to start running (known Jython issue).
• Pack everything into one jar and rely on Jython 2.7-beta3 to run the Python Sel2Lib. I made some preliminary tests. It kinda works, but is slow to start running and there are some error messages. Not sure if I could trust the compatibility with Jython 2.7-beta3 without extensive testing (which I do not have time for).
• Make executables for each main OS using Pyinstaller. Not sure, at this time, if it's feasible. Would require a considerable effort and testing.

I'll probably go for the first option.

The extract keyword poup does not allow editing.
This was due to a implementation limitation that is already solved.

Similar to what is being done when filing out text boxes or verifying text, when we add a step of type "Click Link " or "Page Should Contain Link ", we should check if a variable exists with the desired value, and use it (if option is enabled).

Test run button does not preserve status if window shifts mode from sidebar to float and vice versa, while there's a test suite running at the same time.

This error will then break the loading of the Test Cases section. But it might not be obvious if that section already has some content.
If the Variables section is not defined at all, there will be no problem (just the expected warning).

The "Check Foram As Is" option is using existing variables but not creating new ones.
Currently this is just working for "Form Form As Is".

Not sure if it's a good idea. The available space is limited, specially when using the sidebar mode.
Collapse this table by using a splitter element.

In utils.jsm:

function _replaceHTMLEntities(text) {
text = text.
replace("<", "<").(...)

does not replace globally

The "Fill Foprm Elements" and "Check Form Elements" options are producing test steps for hidden elements

Under the "Browser" menu

It will fallback to "Select From List By Label"

Add the ability to indicate path of local .jar files of RF and Selenium2.

When doing "Page Should Contain", or other type of text verification, remove duplicate test steps.

E.g. if the step is "${Replace} Remove String ${Values}".
The [Arguments] line will have "${Replace} Remove String ${Values}" instead of "${Replace} ${Values}".
Basically the regex used is not appropriate.
This might be relevant with manually edited steps, not steps added automatically with FireRobot (considering the currently supported Sel2lib keywords on the context menu).
Spotted on v1.0.7.

This is relevant for test files edited outside FR.
See aliases list at http://robotframework.org/robotframework/latest/RobotFrameworkUserGuide.html#test-data-tables

The main use case for FR would be creating new test scripts, exporting them in the form of keywords and then integrate those resources in a test project managed elsewhere (RIDE, InteliJ, etc). However, the Load and Save buttons only assume the existence of 3 types of sections: Settings, Variables and Test Cases. When loading and saving a file, any other section will be discarded. This brings the danger of eliminating information added elsewhere (e.g. a Keywords section in the same file).

The [Arguments] section of a new keyword is including variables declared inside the keyword.
This might be relevant with manually edited steps, not steps added automatically with FireRobot (considering the currently supported Sel2lib keywords on the context menu).
Spotted on v1.0.7.

It's pointing to https://github.com/rtomac/robotframework-selenium2library
and not to https://github.com/MarkusBernhardt/robotframework-selenium2library-java

Select and add path of additional RF libraries.

Add button to select resource file and add it automatically to the *Settings" section.

This doesn't seem to be a specifc issue of FireRobot.
Something like .//svg also doesn't work in Firepath
The solution seems to be the usage of something like .//*[local-name() = 'svg']

Currently (v1.0.8) we are only creating test steps for text boxes with tag name "input" or "textarea". This in the context of the "Fill Form As Is" option.
Expand the concept for elements with contenteditable="true".

Only fragment text when needed. Do not produce so many test steps.

Element.textContent will replace &nbsp with \u0020 instead of \u00A0.
This is an old bug on Firefox described here: https://bugzilla.mozilla.org/show_bug.cgi?id=359303.
A test step of type "Page Should Contain Link " will always fail if we have &nbsp inside the link text.
The problem was "solved" in text verifications (Page Should Contain) by splitting the verification into several steps.
The same strategy won't work with "Page Should Contain Link ".

Not sure how to solve this without a fix by Mozilla.

In float mode, you can use FR in a Firefox browser instance different form where it was started, but the toolbar icon will be grey, giving the impression that FR is off.

If the option to create variables is ON but the option to use them is OFF they will be used. Not a likely scenario, but wrong nevertheless.

Make it usable in 768px height.

This is valid for the preferences, help and report.

Steps to reproduce:

1. Open FR in float mode.
2. Open new Fiefox browser instance but leave FR window open
3. Click the "Select mode" button while having the new FF instance as the latest focused window.
   R: You get the dashed border, you can select stuff but the context menu does not open.

js error: contextMenu.openPopupAtScreen is not a function select.jsm:250

There seems to be some incompatibility with Firefox 32 (officially released yesterday) and the java selenium library used to run the tests.
This is a known issue reported in https://code.google.com/p/selenium/issues/detail?id=7642.
It's supposed to be fixed on release 2.43 (don't know when it will be available).

After the test suite is run, the test report won't open automatically in Linux and OS X.
This is probably a limitation of nsIProcess and the runAsync() method.
The observer returns immediately as if the test run is finished. At that point there is no report.

1. Highlight/select part of the generated test script.
2. Click somewhere to create a new keyword based on the selected steps.
3. Add the new keyword name.
4. Select the .txt/.robot resource file to add the new keyword.
5. Click OK.
6. The new keyword is added to the end of the resource file.
7. The resource file is automatically added to the Settings section on FireRobot (if not yet included).
8. The test steps, on FireRobot, are replaced by the keyword (if option is enabled).

form text field -> from test field

When closing and opening FireRobot, while maintaining the same browser instance, the reference to the last working file is not removed.
This leads to potencial file overwrite when clicking on "Save".

setup1: firefox 19/win7x64
setup2: firefox 32/macosx 10.9

problem:
Navigate to e.g. www.sme.sk; www.mmo-champion.com; (literally any webpage)
I started FireRobot
Pressed Select Elements # whole page is outlined
I can't select any element on this page # from time to time, popup with information that tells me that 'we are unable to create xpath to element' appears

What Am I missing?

If we select a big amount of text with thousands of sub-elements we will get a "Script is taking a lot of time..." message form the browser.

This should be enabled to save test run reports.
However try to keep window minimalistic.

This problem is solved when making XPath based on the element text, by using the concat() function.
It's not solved when we we make XPath based on the the title, placeholder or alt attributes.
This leads to invalid xpath when the char is used on these attributes value.

The FR editor does not allow inserting the tab char (\t). The tab key is used to change the focused element. So, to align manually edited test steps we will have to use spaces.
The \t representation is not the same in all text editors. The result is that when editing the test outside FR the steps may be misaligned, due to a mix of \t (added automatically) and spaces.
Replace \t by 4 spaces.

Issue:

The FireRobot add-on can be activated, however selecting the elements on the page does not populate the 'HTML' or 'Test Cases' windows. Right clicking on an element does not bring up the context window.

Notable is once the add-on is closed, all items on that TAB remain inaccessible changing to another TAB on the same browser window and right-clicking on an element now brings up the FireRobot context window. Selecting one of the context items presents the following message:

Warning!
There are no selected elements.

I've installed v1.0.5 on the following systems:
OSX 10.9.4, FF 31.1.0 (ESR)
Ubuntu 14.04 FF 30.0
WIndows 7(64) FF 31.1.0 (ESR).

I tested with the following website, http://www.wufoo.com/gallery/templates/registrations/workshop-registration/, to eliminate the possibility that the website is not fully loading the page.

In the Browser Console, I seeing the following messages when I attempt to select elements and right-click on them:

reflow: 0.18ms
selectedElements is undefined select.jsm:301
reflow: 0.2ms function _calcMouseTargetRect, tabbrowser.xml line 4993
reflow: 0.21ms
browserWindow is null select.jsm:249
selectedElements is undefined select.jsm:312

We can save them without the extension. But in that case we won't be able to open them in FireRobot.
In Windows, the .txt suffix is appended automatically.

If we are enabling a locator, it means we want to use it. It makes no sense to put on the bottom of the preference list.

At the moment it supports only .txt and .text but since .robot is the new official robotframework file extension, the option should be made available.

Most experts favor CSS over XPath, however, this is still mater for debate (http://elementalselenium.com/tips/34-xpath-vs-css-revisited-2).
The option should be made available.

Default dark grey color is shown as white in Ubuntu