The plastic handle of my son's Spyder paintball gun had chipped around a captive nut and broken off leaving a small crater so that the nut had fallen out and the CO2 cartidge fitting could no longer be bolted up securely.
You can see in this first picture that I took a standard #8-32 SHCS, started it into a nut, and set it into the now fully exposed slot where the old nut had once sat. I did no prep work to the broken area other than wrapping electrical tape around the end of the handle to hold the urethane in and form the vertical side. I assumed that the rough surface of the fracture would contribute to more area and a better grip between the two mating surfaces that would be formed.
I obviously could have done a better job getting the electrical tape to lay straight, but I wasn't too concerned as the folded over portion was in a non-critcal area and I knew from experience that this urethane is easy to work with after it sets up. As a matter of fact, I couldn't have cared less about the final appearance of this project - it was to be a function-only fix.
I possibly could have used JB Weld, but I feel that this pourable urethane epoxy was a much better choice in this particular case for a number of reasons. Most notable is that it will flow down into every nook and crevice, which in this case means that it will automatically form a perfectly fitted "foot" that hooks under the unbroken portion of the handle, which will ultimately strengthen the repair's hold. I would be afraid that JB Weld (or any equivalent putty-type epoxy) would only form a nice mold of the cavity, but then just fall out as a plug releif - and this very thing has happened to me on more than one occasion.
In JB Weld's defense, I once used it to seal a small leak from a heat transfer fluid resevoir. The 15 gallon resevoir had gotten a bottom corner worn through by 30 years of being rubbed up against by hydraulic hoses, no less. Without shutting the machine down I stuck a JB Weld ball on the leak, being sure to get it wedged in the hole as well as on both the vertical and upside down surfaces. These surfaces were also at about 250 degrees F, mind you. This was in 1998 and as of mid-2006 that band-aid job was still holding the Ucon heat transfer fluid in the barrel coolant resevoir.