The impact of diabetes on life expectancy is an extremely difficult question to answer, because although there have been many studies they are hard to interpret because they are trying to hit a moving target. The traditional view of the medical profession is quite gloomy. One of the main current endocrinology textbooks says:
"Patients with type 1 diabetes have premature mortality compared with the general population. It is said that on the day that a patient is diagnosed with diabetes their life expecancy is reduced by around one-third."
Holt, RIG & Hanley, NA (2007). Essential Endocrinology and Diabetes. 5th Edition. Blackwell
However (typically of textbooks), this doesn't say where - or more importantly when - this information has come from. Although this is a recent book, the first edition was published in 1983 - and I don't know in which of its five editions this statement originated. The technology for managing diabetes was much more primitive in 1983 than it is today - so it may well be unduly pessimistic (especially since such figures would have to be derived from data collected many years
before this). Also, even if this figure is up to date, it doesn't make any distinction between people with tight control and not. It is horrendously difficult to design studies to put figures on such things, because to get meaningful results you would have to follow a group of patients who followed the same treatment regime over many years. In reality this never happens - people change treatment as they find ones better suited to themselves, and new treatments are developed - which are adopted by different people at different rates.
What it is possible to say is that diabetes (T1 or T2) is not an absolutely inevitable decline, and some people live to an extreme and relatively healthy old age despite the disease. There is always a certain probability of potentially fatal complications occurring, and these people have beaten the odds. As has been mentioned by others, it is currently looking very likely that the absence of C-Peptide in T1 diabetics makes it much more likely that side effects occur, but I have never seen any reliable figures put on this for a large population. What is certainly the case, though, is that the probability of developing complications or dying prematurely is directly correlated to BG control. It is all about probability - and the trick is to give yourself the best possible odds. If they have a really tight BG control then T1s are still much more likely to beat them than if their BG is less well controlled.
There are also many new treatments in development. Even ignoring all of the potential "cures", treatments like mixing insulin with synthetic C-Peptide might well have a major impact upon health and survival of T1s in the very near future. Asking by how much diabetes is likely to curtail your life really isn't a very useful question. Survival is always a lottery (after all, you could be hit by a bus tomorrow). The important thing is to do everything that you can to increase your odds so that you will be alive and healthy enough to benefit from new treatments (and there are many on the way). Too many people view diabetes as an "inevitable decline". Although T2s are luckier here, because so long as they can arrest the decline before their beta cells are completely fried, their fate is very much in their own hands. However, even for T1s, they can still do a lot to dramatically increase their odds - and ultimately that is all that survival is about.