(overriding the personal "Reply-to:" -- others might value replies.)
On 2014-04-01, at 20:26, Jon Perryman wrote: > This means a PE'd PTF will hold up ASM maintenance until the fixing PTF > becomes available. It also means we can't install a PTF to fix a specific > error. Does the ASM group ship APAR's to allow specific problems to be > resolved without applying all PTF's? > > The C compiler list 4 PTF's as the levelset Do you know if they follow the > same methodology as ASM? > There might be a more relevant list for that question, but I don't know. > Your group is lucky because they can simply tell us to bring the product > current because that's what will happen anyways. This isn't how most products > create PTF's. > I recognize a differece in character between HLASM and OS nucleus. I'll still continue from the ISP PoV. Suppose the vendor supplies ten PTFs with no declared interdependncies and presumed noninteractions. How many different configurations can customers create by mix-and-match? Over a thousand. How many of these configurations can customers reasonably expect the vendor to have submitted to regression test? A good guess is eleven. The remainder are untested. -- gil
