Appium cannot find element - appium

I am trying to find the header of an iOS app in a page-agnostic way. This header is visible on all pages and always found with the following class chain:
header_class_chain = '**/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther/XCUIElementTypeOther'
This class chain runs a global search for all elements of type XCUIElementTypeOther that are 14 levels deep. I cannot search by accessibility_id because those tags do not exist, nor do I expect them to be added anytime soon.
In the Appium Inspector, when I search on the above class chain using self.driver.find_element((MobileBy.IOS_CLASS_CHAIN, header_class_chain)), I always find the header element I want, whether I am on a page directly accessed from the in-app menu or on some child page. Thus, I assumed I could use that class chain to find the header regardless of page content.
However, when I run a test script under automation, searching on that class chain finds the header ONLY on pages that are directly accessible via the menu. I expect this class chain to function under automation the same way that it functions manually, that is, it finds the header on any displayed page.
Two interesting observations that may or may not help. First, the header's element ID as reported by Appium does not change while the app is in a given workflow. That is, if I were to progress through the purchase workflow, the header's ID in Appium remains constant until I either finish the workflow or I navigate to another page via the menu. I have observed this behavior both manually through the Appium Inspector and in logs from automated test runs.
Second, while searching for the header on a child page, I saw the following lines in the Appium logs:
[WD Proxy] Got response with status 200: "{\n \"value\" : false,\n \"sessionId\" : \"4C72B56A-C6C6-4D9A-850E-3C95EE014E29\",\n \"status\" : 0\n}"
[WD Proxy] Replacing sessionId 4C72B56A-C6C6-4D9A-850E-3C95EE014E29 with 3a91c7cb-d660-48b4-8b2d-144dbaecfaae
When I started my test, the logs reported [info] [35m[BaseDriver][39m Session created with session id: 3a91c7cb-d660-48b4-8b2d-144dbaecfaae.
Shortly after, another sessionId was introduced to my test run: [debug] [35m[WD Proxy][39m Got response with status 200: {"value":{"sessionId":"4C72B56A-C6C6-4D9A-850E-3C95EE014E29","capabilities":{"device":"iphone","browserName":"APPNAME","sdkVersion":"12.1","CFBundleIdentifier":"com.COMPANY.APPNAME"}},"sessionId":"4C72B56A-C6C6-4D9A-850E-3C95EE014E29","status":0} (app name redacted, as app is not yet publicly available.)
I am not familiar enough with Appium to rule out the possibility that these 2 sessionIds is what's causing my search to fail. If I were to guess, I would say the base driver running the app carries the sessionId starting with 3a91 and the WebDriverAgent is assigned the 4C72 sessionID and that the WebDriverAgent somehow interacts with the base driver. Again, this is just an educated guess and is likely tangential to my real problem: Appium not finding a header element where I clearly expect it to.
Why is this call succeeding manually, but failing under automation?
How could I change my code so that I can find the header, regardless of what page I'm on?

Turns out I was searching on the wrong element. On pages directly accessible from the menu, the header is 14 levels of XCUIElementTypeOther deep. On child pages, the header is 15 levels deep. I must modify my search to find a header that is visible at either 14 or 15 levels deep.

Related

Vaadin23: Expected router outlet to be a valid DOM Node (but got null) and HomeView rendered several times in production mode

when I run our vaadin 23 app in production mode, I get following error in the browser console, while the corresponding site is rendered twice:
Uncaught (in promise) TypeError: [Vaadin.Router] Expected router outlet to be a valid DOM Node (but got null)
First it's rendered with a corrupt textfield (UI written in Vaadin); the second page looks fine. When I try to debug (IntelliJ), it gets rendered correctly, so I added log messages, where I learned, that the HomeView gets initialized even three times, whereas in dev mode it is initialized once. I find it hard to figure out why, since that is run in a thread (I'm far from knowing Vaadin well).
We have two apps, a backoffice and a webshop. The non-responsive backoffice does not show this issue, only the webshop. The webshop also uses two lit web components (but even when I comment them out, I have the same error). The rest is all kept in Java.
Does anybody have an idea how to solve this, or in what direction to search and debug?
Thanks a lot! Sura
This is likely caused by a known bug with the eagerServerLoad flag. As a workaround, try disabling this flag to prevent the issue.
Add vaadin.eagerServerLoad=false to application.properties to disable the flag, assuming that your application uses Spring Boot. You could find alternative ways of setting the property see the Configuration Properties article.

Appium works very slow on screen having dynamic mobile elements

I am working on automating an Android app. We have a screen that has a table wherein the cell values keep on changing in seconds. I have observed that it takes more than a minute for AndroidDriver to execute a single action.
I have read on some other forums about this and got to know that this is how the UiAutomator2 works and it has nothing to do with Appium. UiAutomator2 waits for elements to come to a static state and then performs the actions.
Since the dynamic elements on the screen are unavoidable, is there any workaround for this to make Appium scripts execute with good speed? Let me know what you all think of this. I will really appreciate your comments on this. Thanks.
Yes, that is right. It has nothing to do with Appium itself, but with UiAutomator. UiAutomator takes accessibility snapshots only when the application is in idle mode and the framework itself is not optimized for cases when there are constant changes in the page view. As a workaround, try to play around with waitForIdleTimeout such as setting it to zero or try with setting the disableWindowAnimation to true in the desired capabilities.
You could also try with running the following commands:
adb shell settings put global window_animation_scale 0
adb shell settings put global transition_animation_scale 0
adb shell settings put global animator_duration_scale 0
As a last resort, you could ask for the developers to provide you with a special build that has the animations removed in the source code.
I had the same issue, my specific problem was that the mobile app had a timer in the UI, that was changing its value every second.
Looks like the driver was waiting for the DOM to become idle, so it was SUPER slow because of that.
This simple command fixed the issue for me:
# Python code
driver.update_settings({"waitForIdleTimeout": 0})
Additional references:
https://appium.io/docs/en/commands/session/settings/update-settings/
https://github.com/appium/appium/issues/14155

Difference in manual execution of running app in background and runAppInBackground() using appium

While executing runAppInBackground() for Android application through Appium the app gets restarted but when executed the same manually couldn't be able to reproduce the same. I Would like to deep dive into implementation of a runAppInBackground() method to reproduce the same issue in a manual way.
You need to look behind the code of runAppInBackground
From java client side (your test code) perspective, it is a single call to Appium server:
POST "/session/:sessionId/appium/app/background"
If you continue looking into where its implemented on server side, you finish with appium-android-driver function.
In short what it does:
Get current activity and package
Press physical Home button
Wait for time you provided as argument (seconds)
Bring up back in focus based on different conditions; from the code you can understand what activity is being started
Basically its a sequence of adb shell commands, that you can run from terminal.
My guess is that step 4 you did manually may differ from what Appium is doing: different activities/arguments for activity been called

IBM i (AS/400) command works locally but not remotely

I have a Windows service written in VB.NET 2.0 which connects to an IBM AS/400 server. Queries work fine, but when I try to do something like deleting a spool file, I get errors. For example:
CPYSPLF FILE(PO630A) TOFILE(MPLCDATPAR/PO630APF) JOB(083064/ARUSER/POASYNCMON) SPLNBR(80) MBROPT(*REPLACE)
Running this command with ExecuteNonQuery yields:
CPF3342 - Job not found 083064/ARUSER/POASYNCMON
However, if I run that same command locally in AS/400, it works just fine. We already checked permissions. What else could be causing the command to fail this way? How can I get more information about the error, or go about troubleshooting this?
EDIT: This problem (and a lot of other ones) appeared when we migrated our server (where the .NET service runs) from Windows Server 2003 to Windows Server 2008.
How can I get more information about the error, or go about troubleshooting this?
The first thing is to verify that the IBM AS/400 server [what OS Version Release and Modification level, Technical Refresh (TR) level (if instead IBM i), Cumulative PTF level were all omitted.?.?] used for the connection is the same server used to perform the command-line invocation; i.e. on the server where the command-line invocation will be made to verify the command is functional, find the active server job in which the CPF3342 is still visible in the log.
The second thing to do is to get the spooled joblog showing the full details of the CPF3342 [and possibly any preceding message(s) that might be related]. If for example the message is not actually that message or is not sent by the expected program QSPCPYF, then immediately the direction of investigation probably would change. What is shown is apparently what is presented at the client, not what came from the server joblog; the USEnglish formatting I believe is "Job &5/&4/&3 not found." for which the formatting "CPF3342 - Job not found &5/&4/&3" is suspect.
To ensure the most appropriate comparison to the request made from the client:
• the local user that is signed-on to perform the same request should be the same user as the Current User of the active job found to be servicing the client request
• the local user should establish the same System Library List as the active job found to be servicing the client request
If such an incident recurs or even if the same incident persists, then verify the once again the re-create is still possible using the same interface [i.e. the condition\failure persists] and again verify the command-line request is successful [i.e. the circumvention is confirmed, that the same request is possible to be performed at the command-line]; and according to my earlier comment, first ensuring the same server by finding the active job that logs the CPF3342. Immediately afterward:
• Collect a job trace for the Copy Spooled File (CPYSPLF) request; for the failing case, review for any exception\interrupt conditions [with or without a message as accompanying trace data] that precede the program flow for the issuance of the msg CPF3342.
• Review the audit log for any T-AF or anything odd\unexpected at very close to the time of the failing request; expansive auditing should have been established since before the connection to the server.
• Contrast those data collections of the failing case with the same data taken from the successful processing.
Although the symptom [as lightly described, without the full joblog] the possibility of command-exits seems remote, the trace would reveal if the command in either scenario were intercepted by the Command Exit points; these can be reviewed separately [rather than looking in the trace] for any Exit Program, using the Work With Registration Information (WRKREGINF) to review any QIBM_QCA* entries in the repository for what exit programs might impact the CL Command request. But IIRC the trace-data shows which command was invoked, so the trace would also reveal if the unqualified command requests resolved to different *CMD objects.

JBoss not responding to valid URLs

I encountered strange behavior of an JBoss AS 7 on linux server. When I deploy war with my application, the server doesn't respond for valid HTTP requests for this app. When I try to GET a valid URL, I can see in logs, that backend functions (e.g. DAO methods) are called, debug logs shows that subsequent tags in my JSF are rendered, I even can see a message Rendering View index.xml, but the response never reaches the client.
When I use a non-existent URL (e.g. index.asd) response is 404 and when I use a wrong page name (e.g. inswxasd.xhtml) the response is 500. Thus it fails only with valid requests.
I tried to connect both remotely (using firefox) and locally (with wget) and I reproduced the problem. The strange thing is that I deploy the war that I already used and then it worked.
EDIT 1:
I've just noticed that when I send a request to, the server process takes 200% of CPU. Additionally it happens for only one application with stack: Hibernate, Spring, JSF 2.0, Primefaces.
EDIT 2:
Here is pastie with jstack output (jstack -l -F <PID>). All threads are blocked.
Additionally I noticed (using top and dumping stack of the JBoss process) that the problem is most probably caused by a thread called http--0.0.0.0-8080-1. Any ideas?
Thanks for help. I reverted the database to previous state, and it started to work. Apparently there was a problem in the data that caused this infinite loop. Now we need to find it, but as it is reproducible, I'll manage.

Resources