A .32-20 Update

Joe, you got it.  In measuring cylinder gap the "pass" dimension is the thickness of blade feeler gage you can insert between the barrel extension and cylinder, and pull the trigger DA with empty brass in the chambers through a full rotation without resistance.  The "hold" dimension is the thickness of gage which binds the cylinder repeating the same test when you try to pull it through DA.

To measure end shake or end play, repeat the "hold" dimension test with an empty cylinder, no brass.  We use fired brass as an expedient replacement for Go headspace gages.  End shake is the difference between the "hold" dimension measured with and without "rear gage" in place.

I'm not sure what current S&W specs are...When I was at Ruger minimum "pass" was 0.003" before proofing.  After proof and function test it is normal for the cylinder gap to increase up to 0.001 or so as the parts settle into place.  Statistical average for new revolvers is close to the Mean Assembly Tolerance of 0.004 pass/0.006 hold, with no more than 0.002" end play.  Max. cylinder gap for new guns going into shipping after proof and targeting was 0.008" pass, 0.009" hold, but this would be very unusual unless the LE customer specified a larger gap as insurance against cylinder binding with lead +P service ammo, typical of St. Louis, Kansas City, NYPD, Boston, Baltimore, DCMPD.  Agencies using jacketed ammo, US Secret Service, Customs and Border Patrol, US Marshalls, Postal Inspectors, ONI, USAFOSI, etc. would specify gap towards min. as their revolvers would never see lead ammo. 

Max gap for customer service repairs in the 1980s was 0.009 pass, 0.010 hold.  I have seen recent S&Ws which were this large and have measured dozens of wartime S&W Victory Model .38s which came out of Factory Through Repair in UK, Canada and Australia which were as large as 0.012." 

I can tell you from personal experience that if cylinder gap is over 0.012" that a S&W Victory revolver firing UK Mk2z .380 Rimmed service ammo with 178-grain FMJ bullet WILL lodge a bullet in the barrel before you complete a 100-round function test.  I have seen Webley MkVI .455 revolvers with cylinder gaps as large as 0.018" which shot perfectly OK with service loads or lubricated lead bullets, but my recommendation is that revolvers having cylinder gaps over 0.012" probably should not be fired until repaired by a competent gunsmith.

If the cylinder gap was found excessive the factory fix at Ruger was to refit a new cylinder. 

"Long" cylinders were used in customer service and were selected by sorting to be at max. tolerance +0.005" from Mean Assembly.  The substitution would bring most loose guns into spec after the end shake was taken out by stretching the crane arbor.

Occasionally when an older S&W is returned to the factory for repair of a loose gap, I have been told they may return it as "Unrepairable OFG".  This means "Open Front Gage", which is a factory term for excessive cylinder gap.  While the usual gunsmith fix is to set the barrel back a thread, take out the end shake and reface the barrel to correct the loose gap, the factory customer service depts. seldom do this any more, and will return a gun as "unrepairable" if it cannot be fixed by replacing parts which are in stock.  They don't want to get into a custom shop, repair operation.

Dave Reiss please chime in if I have left anything important out. 

My experience is more to Rugers than S&Ws and there are important differences.