In our scenegraph app, after about an hour of HLS live stream playback, a regular lookup on the global node crashes the app. This code is in a task that runs an http request at periodic intervals.
BRIGHTSCRIPT: ERROR: roSGNode.<lookup>: Rendezvous aborted for Node: pkg:/components/cpc/metrics/tasks/cpc_http.brs(19)
BrightScript Micro Debugger.
Enter any BrightScript statement, debug commands, or HELP.
Suspending threads...
Thread selected: 5* ...pc/metrics/tasks/cpc_http.brs(19)
Current Function:
011: http.SetPort(port)
012: timeout = Int(m.global.http_timeout_seconds) * 1000
http.setCertificatesFile("common:/certs/ca-bundle.crt")
013: http.InitClientCertificates()
014: http.enablehostverification(false)
015: http.enablepeerverification(false)
016: http.SetUrl(url)
017: timeout = 30000
018: if m.global.http_timeout_seconds <> invalid
019:* timeout = Int(m.global.http_timeout_seconds) * 1000
020: end if
021: if http.AsyncGetToString() Then
022: msg = wait(timeout, port)
023: if (type(msg) = "roUrlEvent")
Type Mismatch. (runtime error &h18) in pkg:/components/cpc/metrics/tasks/cpc_http.brs(19)
019: timeout = Int(m.global.http_timeout_seconds) * 1000
Backtrace:
#0 Function request() As Dynamic
file/line: pkg:/components/cpc/metrics/tasks/cpc_http.brs(19)
Local Variables:
global Interface:ifGlobal
m roAssociativeArray refcnt=2 count:2
url roString refcnt=1 val:"http://..."
http bsc:roUrlTransfer refcnt=1
port bsc:roMessagePort refcnt=2
timeout Integer val:30000 (&h7530)
msg <uninitialized>
If I run some logging in this debug session, i get strange results. The global node is there, but accessing this particular value (which is set when the app starts up) gives me weirdness:
Brightscript Debugger> ? m.global
<Component: roSGNode> =
{
change: <Component: roAssociativeArray>
focusable: false
focusedChild: <Component: roInvalid>
id: ""
_rsgdisplaysize: <Component: roAssociativeArray>
analytics: <Component: roSGNode>
appconfig: <Component: roAssociativeArray>
brandconfig: <Component: roAssociativeArray>
cpc_log_enabled: true
cpc_log_level: "ALL"
deeplinkservice: <Component: roSGNode>
encryptionkey: "86CCB539521E6554D623F1K12A4CF29H"
http_timeout_seconds: 30
initializationvectorkey: "95FBC64A9F7837FE"
migrate: <Component: roSGNode>
moduleproperties: <Component: roAssociativeArray>
retrycount: 0
router: <Component: roSGNode>
userinfo: <Component: roAssociativeArray>
watchlist: <Component: roSGNode>
}
Brightscript Debugger> ? m.global.http_timeout_seconds
BRIGHTSCRIPT: ERROR: roSGNode.<lookup>: Rendezvous aborted for Node
invalid
To be clear, this lookup has worked without issue dozens of times previous to this crash.
Any ideas what is going on here? Memory issue? this is on a Roku Stick:
- - - - - - - - - - - - - - - -
Model: 3500X
Display Name: Roku Stick
Firmware: 098.00E04128A
Friendly Name: OldStick
Display Size: 1280x 720
UI Resolution: <Component: roAssociativeArray> =
{
height: 720
name: "HD"
width: 1280
}
Video Mode: 1080p
-------------------------------------
