indigo_ccd_qhy2 crashes the server if launced with the -i option

[Paolo97Gll]

Following the documentation, I launched the indigo_ccd_qhy2 driver using the -i option.

indigo/indigo_drivers/ccd_qhy2/README.md

Lines 34 to 41 in 1b3d02b

	### Underlaying SDK is not stable
	Due to instability in the vendor provided SDK problems on all platforms should be expected. Therefore it is
	advised to use this driver with -i option like this:
	indigo_server -i indigo_ccd_qhy2
	This will execute the driver in a separate process and in case of a driver crash the server will not be affected.
	This will come at the cost of somewhat reduced performance.

I normally use the following command:

indigo_server --enable-info --disable-bonjour indigo_ccd_asi indigo_ccd_atik -i indigo_ccd_qhy2 indigo_wheel_sx indigo_agent_alpaca

For the test I used the following command.

indigo_server --enable-trace -i indigo_ccd_qhy2  > indigo.log 2>&1

And then the server crashed when I tried to take a 0.1 second shot (it happens also with 0.001, 0.01 and 1). The behaviour is the same that happens with the first ("long") command. Here the log file: indigo.log.

Without the -i option, everything works well and no crash happen. The camera is a QHY5 III 178 mono.

[polakovic]

We'll check it, but for now try to avoid -i switch. For such a fast camera it may slow down everything by many orders of magnitude (because of very inefficient INDI legacy protocol enforced by -i switch).

@rumengb, it looks like a memory overrun in base64 decoding.

[Paolo97Gll]

Ok, thanks!

[rumengb]

@peter Polakovic ***@***.***> I have checked the problem... it does not crash here but there is another strange issue. Image retrieval hangs until timeout occurs... 08:16:33.070885 indigo_server: Worker thread started socket = 30 08:16:33.070973 indigo_server: 4 clients 08:16:33.112826 indigo_server: GET /blob/0x2fa39f10.fits -> OK (16240320 bytes) 08:16:34.114445 indigo_server: 3 clients 08:16:34.117261 indigo_server: Worker thread finished 08:17:04.917428 indigo_server: indigo_write(24): Broken pipe 08:17:04.928936 indigo_server: Worker thread started socket = 25 I was able to reproduce it with asi too...

…

On Sat, Feb 12, 2022 at 9:47 PM Paolo Galli ***@***.***> wrote: Ok, thanks! — Reply to this email directly, view it on GitHub <#441 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AE5EZBPQDOGLEAHSMI5GAWTU222ODANCNFSM5OHGLROQ> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you were mentioned.Message ID: ***@***.***>