idle_*.cmd scripts are no longer called

[hb020]

Perform all steps below and tick them with [x]

• Check the related part of the Documentation
• Update openHASP to the latest version
• Reproduce the issue and describe all steps

Describe the bug

with today's git HEAD (8d5ae79), the following has arisen:

The idle_*.cmd scripts are no longer called. Very likely due to this commit and #601

Right now, I have
MQTT PUB: idle => long in the logs, but no MSGR: Loading L:/idle_long.cmd

To Reproduce

• get the latest commit (8d5ae79)
• create a number of pages (one having number 11 in my example)
• have the following in idle_long.cmd

json ['page 11']

• when sending the command "page 11", the page shows.
• go to another page
• upon idle_long, the script will not be called and the page will not be called.
• when reverting to commit 0ebfe7d, it works again.

Expected behavior

upon idle_long, the idle_long.cmd script should be called. (same for and idle_short, and idle_off)

Screenshots or video

none

[oklona]

Hello. To have this work, I just added the following lines to hasp_dispatch.cpp, under line 1359 (the end of function dispatch_idle_state):

if(state == HASP_SLEEP_OFF) { dispatch_run_script(NULL, "L:/idle_off.cmd", state ); } else if(state == HASP_SLEEP_SHORT) { dispatch_run_script(NULL, "L:/idle_short.cmd", state ); } else if(state == HASP_SLEEP_LONG) { dispatch_run_script(NULL, "L:/idle_long.cmd",state ); } else { LOG_WARNING(TAG_MSGR, F("Invalid idle value %d"), state); return; }

I don't know the purpose of the "source" parameter to dispatch_run_script, and it looks like it is not actually used for anything, so I just passed the uint8_t "state".

[yeyeto2788]

Hey @oklona aren't the filenames incorrect? Based on the docs those are supposed to have - instead of _.

Even if the names are different, did that fully work?

[oklona]

Yeah, I just copied it from the "dispatch_idle" function below in the source file, which I don't see being called from anywhere. I didn't research much, but it seems to me that "dispatch_idle" was previously used, but was replaced by "dispatch_idle_state". Therefore I assumed I had simply misunderstood the docs. So I did change my file names to use underscore instead of dash. -But apart from that, it worked 100%. Obviously, changing the underscore to dash in the source code would do the same, but during debug I see that underscore is used also for mqtt_on.cmd, mqtt_off.cmd and such. So to me it seems like at some point developers have decided to use underscore.

[yeyeto2788]

Oh yeah!!! It works 😄. Thanks for that, I guess it would be great to create a PR to the repo.

[hb020]

It could be wise to check with the reason behind the revert commit mentioned above. It moved the script calls to dispatch_idle(), and right now the only places that function gets called is during init phases. It might just be a typo, maybe there is a mix up between dispatch_idle() and dispatch_idle_state(), but anyway, there is reason behind that revert commit. I was hoping for @fvanroie to chime in, he did the commit.

[oklona]

Agreed. This looks like it has been forgotten somewhere (I cannot find anywhere in the code actually calling dispatch_idle_state() currently). So, since I am very new to this, I only chime in with what could possibly fix this, and leave up to others to decide whether my solution is what they want :-)

[hb020]

Created pull request

[hb020]

PR accepted, tested, closing the issue.