"if you repair or remove the defect afterwards by some grinding or other means - it does not negate the fact that it had been a "fail" upon earlier inspection earlier"

If it had been a fail upon earlier (initial) inspection, for me, the qualification test at that point is over and there would not be an opportunity to go back and bring it into compliance.  Once the welder tells me it's finished, if it doesn't meet visual criteria requirements, the welder fails.  My only input is prior to the test when I tell the welder that if it doesn't meet the visual requirements of D1.1, I have to reject it.  I only do this because some newly hired and experienced welders either come from other backgrounds and are not familiar with D1.1 visual requirements, or they come from a structural shop where they were not fully held to meeting visual quality criteria.  They were basically told to just weld it and get it out the door.  When this is the case, I take a minute to explain to them that the weld must pass visual before I can do the bend test, and then I go over the visual criteria.    
 
I don't have a problem with a welder using a grinder to remove a string of rejectable porosity as long is it's done before the welder tells me he's finished.

I find it hard to justify from any of the applicable code in D1.1 to fail a welder in the middle of the test because he ground out any amount of discontinuities providing the WPS, job specs, and customer description of what he wants me to accomplish do not restrict the use of grinders for testing. 

That is indeed "stretching it a bit". 

Test is conducted outside or with shop doors open, gust of wind comes up and blows gas shield away, boom, porosity in the weld and you are going to disqualify them for having the observance to remove it and then proceed with the test.  That is exactly what I want to see, responsibility in doing their own QC without me saying anything and making a correct weld. 

He Is In Control, Have a Great Day,  Brent