ChimeraX hangs if the blastprotein module can't connect

[goddard@…]

During a meeting yesterday (26 August) our infrastructure was down. When attempting to contact ChimeraX Web Services to run BLAST, ChimeraX froze.

[Zach Pearson]

I tested this by disconnecting from the internet and experienced a traceback instead of a hang, so this may be obsolete now that the interface is in Qt (or may have been a one-time thing).

After replacing the traceback with an error message, some extra text was printed to the log alongside my error message.

[Zach Pearson]

Extra text

[Tom Goddard]

I tried a simple test with current ChimeraX daily build using Blast Protein gui with wifi turned off (no internet). ChimeraX gave reasonable error message.

Was not able to figure out how to change the web services url to see what error I would get with network connectivity but no server responding. Tried changing prereqs/cxservices and looking through webservices and blastprotein bundles, but failed to get it to use a new URL for the web service. But Zach says he tested this and got a reasonable error message.

Possibly this error required a zombie web server that only gave a partial response. That may be a rare case and if we cannot easily reproduce it then fine to close this ticket and wait until it is reported again.

[Zach Pearson]

Can't reproduce, but testing did uncover the need to replace a nasty traceback with a clean error message, which we did.