<html><head></head><body><div>On Thu, 2015-08-06 at 13:47 -0400, DJ-Pfulio wrote:</div><blockquote type="cite">In a 1-man shop, editing on prod is fine.<br>
<br>
In any larger org, it should lead to termination. I've seen hundreds of<br>
"tiny changes" bring down production systems because they bypassed<br>
dev-test, pre-prod, and deployment processes.<br>
<br>
<br>
Yes - I've done it a few times too. Now I know better.</blockquote><div><br></div><div>About the only thing I will ever agree to is a hot-change that is perfectly understood and is so isolated that domino effect is not possible. If I am unable to prove either (which is 99% of all situations), I won't hot-patch at all.</div><div><br></div><div>I most recently had to patch a single ASM instruction in an ARM program I deployed to change the initial value of a variable. (Why, you might ask? Old project, and reference hardware and toolchain were archived a long time back, and the change needed to apparently be made rapidly.) Takes "forever" to do relative to edit C module and re-make, unless the toolchain isn't there. But relative to the time of rebuilding the toolchain or digging it out of archives, it is pretty speedy...</div><div><br></div><div><span class="Apple-tab-span" style="white-space:pre"> </span>— Mike</div></body></html>