certain code causes onMount() to have hidden slow execution

[Adrian-Sanchez]

I first noticed it with a cfhttp() in the onMount(). But i have to say, this took me a lot of time to figure out because when I found out the thing you use to troubleshoot, also is the problem..SMH, it was brutal.

Symptom: Certain code will cause cbwire to take 2-3 seconds longer. What's more interesting is that it's not always. You can have an offending piece of code run seamingly fine but wait about 40 seconds with no requests to the server, then run the full lifecycle and it will hang. Wait another 40 seconds, it will do it again. But back to back full lifecycle requests will run promptly with no hanging.

I'm sure Lucee is partially at fault. I'm not entirely sure if it's a cbwire issue. BoxLang doesn't have this issue.

Test case:

`function onMount() {
var startTime = getTickCount();
//writeDump(var="onMount called",format="html");//triggers the issue
//writeOutput("onMount called
");//doesn't trigger the issue
writeOutput(serializeJson(callstackget()) & "
");//triggers the issue
var executionTime = getTickCount() - startTime;
data.executiontimes.push("onMount executed in " & executionTime & "ms");
//runCfhttp()//triggers the issue
}

I've also pushed a test case to a repo
https://github.com/Adrian-Sanchez/cfhttp-bug-cbwire.git

[mrigsby]

@Adrian-Sanchez Are you testing on Windows/Mac/Linux and what versions of Lucee have you tested it on?

If memory serves, I ran into something similar with CBDebugger in the past but I can't remember the workaround/fix. I'll have to take a look at your test case repo and see if I can recreate the issue.

[Adrian-Sanchez]

I’m on windows The test case I uploaded is using lucee 6 It should be relatively easy to reproduce.

…

On Wed, Dec 10, 2025 at 3:35 PM Michael Rigsby ***@***.***> wrote: *mrigsby* left a comment (coldbox-modules/cbwire#254) <#254 (comment)> @Adrian-Sanchez <https://github.com/Adrian-Sanchez> Are you testing on Windows/Mac/Linux and what versions of Lucee have you tested it on? If memory serves, I ran into something similar with CBDebugger in the past but I can't remember the workaround/fix. I'll have to take a look at your test case repo and see if I can recreate the issue. — Reply to this email directly, view it on GitHub <#254 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGJSJ5ZWGVSO7PQHNA5FTU34BCU2RAVCNFSM6AAAAACOVO52B6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTMMZZGM3TMNRWGA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[grantcopley]

@Adrian-Sanchez I suspect this is a Lucee issue. There is nothing particularly special about the onMount() method that would cause it to run slower versus any other CFML method. What exact version of Lucee 6 are you on?

[mrigsby]

@Adrian-Sanchez & @grantcopley

This does seem to be the same issue I ran into awhile ago. Take a look at https://community.ortussolutions.com/t/cbwire-20-second-random-loading-delays/9721/14

I am able to replicate the delay on my windows dev environment however it does not occur on my Mac. I fired up FusionReactor and it does show the delay is related to java.io.File.exists & java.io.winntfilesystem.getbooleanattributes

The fact that the delay occurs when using the "Run through wire" button for the CFHTTP call tells me it isn't related specifically to the onMount() lifecycle event.

Also of note is that I did add an exclusion in my anti-virus and it still happens.

If you google WinNTFileSystem.getBooleanAttributes performance you can see it seems to be an issue for many in other java applications

FusionReactor Profiler Screenshot:

[Adrian-Sanchez]

@mrigsby What's weird is that it takes about 40s for it to trigger. It's not every time so when i was coding, cbwire was just zooming and i'd get sidetracked, come back to the page, navigate to a new page and it would take longer than usual. Side Note: I did not have Fusion-Reactor installed on that machine so i'm not going to say how long it took me to figure out it was coming from that onMount() function; Log messages everywhere 🤦‍♂️

But if you stay on the same page, it doesn't happen at all.

@grantcopley I didn't specify for the test case, it's running Lucee 6.2.3+35. But i did see it in Lucee 5 as well

[Adrian-Sanchez]

@mrigsby

"Run through wire" button

I never got it to trigger the delay on this button...Interesting.

[mrigsby]

I think I only got it to happen using the button after a longer wait and commenting out setInterval function

Either way, you can test that it isn't specifically onMount() if you:

1. comment out writeOutput( serializeJson( callstackget() ) & "<br>" );
2. add <cfdump var="#callstackget()#"> to the test.cfm wire template

You will still get the delay issue.

I think this is because CBWire handles the initial render differently than the subsequent updates. it seems something in the initial wire() function is triggering this bug.

[Adrian-Sanchez]

@mrigsby

2. add <cfdump var="#callstackget()#"> to the test.cfm wire template

That's such a good catch. That would have drove me insane to see it come back but have an empty onMount().

The good news for me is LangBox is runnable for my app that I was working on.... and production is going to be in a containerized (not windows).