A gas line can be pneumatically or hydrostatically tested per B31.3. A hydrotest is safer during a catastrophic failure and hence is preferred, especially if pressure and/or stored volume in the line are comparatively high.
For future reference, you'll get a better targeted audience for piping and pipeline issues at forum378.
As moltenmetal pointed out, there is a substantial difference in the consequence of a failure when comparing liquid and air (or other gas) testing. Keep in mind that the reason we test something is because we expect that on occasion we will experience a failure. Otherwise there is no reason to test.
One reference for determining the explosive potential and safe distances for your proposed pneumatic test is ASME PCC-2 Repair of Pressure Equipment and Piping, specifically Article 5.1. I believe this article has been changed for the 2013 2014 edition which, depending on who one is to believe, is expected to be published sometime between "very shortly" and December.
There are exceptions such as large size flare headers which only being tested pneumatically; because if it would be intended to be hydrotested on the relevant pipe rack, it's needed having huge foundation and structure can tolerate the load of such full of water headers...
Also, under B31.3 each weld needs to be 'inspected' - with a hydro you feel for water weeping out. With a pneumatic, you have to 'soap' every inch of every weld. For [aboveground] gas piping I still hydro, then blow out with compressed air. Then pressurize with dry air or nitrogen for a few minutes, then exhaust that gas. Repeat until dry enough for your startup. Also, in B31.3 it is expected that the installer and/or the engineer-of-record will put in high-point vents and low-point drains. For future Lock-out and blow down, these vents will be very valuable. Same for thawing out frozen stuff, there is a way to put steam in and get condensate out.
