Richard Corfield

Software Engineer / Technical Architect working in the UK. on and around the Salesforce platform. Degree and interest in control and information engineering

[Richard Corfield @RichardSoftware]

Unit test passed with a single depth subquery with 500 records.

[Richard Corfield @RichardSoftware]

A question for anyone using queryWithBinds() on Salesforce.

Database.query() has to be used in cursor mode if a subquery has more than 200 records, otherwise you are left with a dangling cursor exception reading the subquery after statement completion.

...

[Richard Corfield @RichardSoftware]

Database.queryWithBinds() does not support cursor mode.

Can this handle subqueries larger than 200?

[Richard Corfield @RichardSoftware]

Today I did 50 climbs to raise money for Yorkshire Cancer Research and Diabetes UK.

Total climb between 500m and 750m (as big as one of our local mountains). Time about 6 hours.

https://t.co/XtR8NkEwhL

[Richard Corfield @RichardSoftware]

@FishOfPrey @PhilW_SF Got to love your comment "Bad developer bad!"

I have in capital letters in mine "THIS MUST NEVER EVER BE USED IN PRODUCTION CODE!"

I found I could demo basic races without needing the delay given 10 threads.

[Richard Corfield @RichardSoftware]

Ouch! I'm writing demos to present on locking in Salesforce and thought I'd use a "publish immediately" platform event for locking.

Firing platform events breaks locks!!!!

Here the demo with logging inline, and again with logging delayed until end of execution.

[Richard Corfield @RichardSoftware]

@FishOfPrey @PhilW_SF Thanks a lot.
I didn't know if I could publish code I'd developed as part of my work, so glad you made some.
I was wondering whether to raise a case, but this way you've got direct knowledge of the problem.

[Richard Corfield @RichardSoftware]

@FishOfPrey @PhilW_SF Here's the Bean Counter.

You _really_ don't want to know what's in Util.sleep() ;-)

[Richard Corfield @RichardSoftware]

@FishOfPrey @PhilW_SF Link to video demonstration (too long for Twitter)
https://t.co/DqEqKBcdes

[Richard Corfield @RichardSoftware]

@FishOfPrey @PhilW_SF My comment hints at another possible issue. Things were stopping dead in their tracks when I exceeded a field width on a text field in the platform event definition.

[Richard Corfield @RichardSoftware]

@FishOfPrey @PhilW_SF Here's the code as it finally stood. This is not production code! (It was quick demos of locking patterns for a seminar on locking). The original called publish() immediately rather than storing things up.

[Richard Corfield @RichardSoftware]

@kranzforce @PhilW_SF That is apparently what is happening.

[Richard Corfield @RichardSoftware]

@PhilW_SF Drat - it was a typo when I was writing the log code. I thought I'd corrected it. Maybe that was an old screenshot. I'm using FOR UPDATE. Dunno where I got the ACCESS idea from.

[Richard Corfield @RichardSoftware]

@PhilW_SF My workaround has been to batch up events and fire them at the end of the transaction. I then reconstruct the order on the client from the timestamps.

[Richard Corfield @RichardSoftware]

@PhilW_SF So I use Platform Events in "Fire Immediately" mode. The allows me to stream debugging information to the front end.

What I'm seeing is that firing the event causes the FOR ACCESS lock to be dropped, but not a DML UPDATE lock.

[Richard Corfield @RichardSoftware]

The client is reconstructing event order from timestamps.

[Richard Corfield @RichardSoftware]

I may need to include timestamps in my logs and have the client reconstruct the message order.