I've found an very interesting article who state that the new legislation (directive 2005/32/EC) will not ban the magnetic ballasts, but only inefficient ballasts(it's not important if electronic or magnetic). You can read the section 8 (especially subsection 8.2 and 8.3).
By the way you find also many others interesting things in the others sections and in general on the website.

Ok I see now that Directive 2009/125/EC supersedes Directive 2005/32/EC, so the article is a bit outdated. But fortunately nothing is changed for the electrical and electronic products.
From here:

Directive 2009/125/EC supersedes Directive
2005/32/ECwith regard to the eco-design
of energy-related products. As a result, other
products such as windows, taps, shower heads
and insulation materials have also been included
in the scope of the new eco-design directive.
However, nothing has changed in terms of
eco-design requirements for manufacturers of
electrical and electronic products (in this case
luminaires and luminaire accessories).

So are B1 magnetic ballasts to stay ? If yes thats excellent as they are the best magnetic ballasts out there, in this case it will cause a step up in magnetic ballast quality instead of them disappearing

So are B1 magnetic ballasts to stay ? If yes thats excellent as they are the best magnetic ballasts out there, in this case it will cause a step up in magnetic ballast quality instead of them disappearing

I haven't looked at the document in detail, but from what I see on page 5 of the last link, ballast classes B2, B1 and A3 (low-performance electronic) will be banned on 2017-04-13, and it would be the end of magnetic fluoro ballasts.

Fortunately there doesn't seem to be any problems from magnetic HID ballasts. From 2017 a 400 W HID ballast (for example, the requirements vary depending on lamp power) would have to be at least 90 % efficient, which wouldn't be a problem at all (the losses today are typically like 25-30 W from what I have read in catalogues, but at room temperature so in reality a bit higher).

Not all magnetic ballasts will be banned!
The great thing about the new directive is that the EEI (Energy Efficiency Index) is no more based on the system power, but on the ballast efficiency; so lamp power divided by system power. There are now three different power values of lamps: a nominal power, which is, so to say, only the name of the respective lamp, a rated power for mains frequency operation and a rated power for HF operation.  Thereby an impartial treatment of both magnetic and electronic ballasts is now granted.

The definitive directive that will start in 2017 say that all ballast efficiency must be at least:
If PLamp  5 W EBbFL = 0.71
If 5 W < PLampe < 100 W EBbFL = PLamp/(2*sqrt(PLamp/36)+38/36*PLamp +1)
If PLamp  100 W EBbFL = 0.91

A 0.95 factor must be used to takes account of the characteristics of a lamp operated by magnetic ballast.

I put here some examples for the most common fluorescent size:

T8 18W lamp. Lamp power of 18W for magnetic and 16W for electronic ballast.
With the old 2000/55/EU directive:
Maximum input power for a T8 18W lamp: B1= 24W, A2= 19W
From these values I calculate the ballast losses permitted for the magnetic: B1= 24-18= 6W, A2= 19-18= 1W.
For electronic ballast: B1= 24-16= 8W, A2= 19-16= 3W.

With the new 2009/125/EC directive:
Maximum losses for magnetic ballast:
EBbFL = 18/(2*sqrt(18/36)+38/36*18 +1)= 0.84
Plosses= 18*0.95/0.84 - 18*0.95 = 3.25W

Maximum losses for electronic ballast:
EBbFL = 16/(2*sqrt(16/36)+38/36*16 +1)= 0.83
Plosses= 16/0.83 - 16 = 3.22W

T8 36W lamp. Lamp power of 36W for magnetic and 32W for electronic ballast.
With the old 2000/55/EU directive:
Maximum input power for a T8 36W lamp: B1= 41W, A2= 36W
From these values I calculate the ballast losses permitted for the magnetic: B1= 41-36= 5W, A2= 36-36= 0W(which is impossible to do!).
For electronic ballast: B1= 41-32= 9W, A2= 36-32= 4W.

With the new 2009/125/EC directive:
Maximum losses for magnetic ballast:
EBbFL = 36/(2*sqrt(36/36)+38/36*36 +1)= 0.88
Plosses= 36*0.95/0.88 - 36*0.95 = 4.66W
Maximum losses for electronic ballast:
EBbFL = 32/(2*sqrt(32/36)+38/36*32 +1)= 0.87
Plosses= 32/0.87 - 32 = 4.66W

T8 58W lamp. Lamp power of 58W for magnetic and 50W for electronic ballast.
With the old 2000/55/EU directive:
Maximum input power for a T8 58W lamp: B1= 64W, A2= 55W
From these values I calculate the ballast losses permitted for the magnetic: B1= 64-58= 6W, A2= 55-58= -3W(which is impossible to do!).
For electronic ballast: B1= 64-50= 14W, A2= 55-50= 5W.

With the new 2009/125/EC directive:
Maximum losses for magnetic ballast:
EBbFL = 58/(2*sqrt(58/36)+38/36*58 +1)= 0.895
Plosses= 58*0.95/0.895 - 58*.95 = 6.46W
Maximum losses for electronic ballast:
EBbFL = 50/(2*sqrt(50/36)+38/36*50 +1)= 0.890
Plosses= 50/0.890 - 50 = 6.13W

To summarize regarding the B1 magnetic ballast refering 2000/55/EU:
The T8 18W will not pass the new requirement (old B1 Maxlosses=6W, New A2 Maxlosses=3.25W).
The T8 36W will not(but nearly) pass the new requirement(old B1 Maxlosses=5W, New A2 Maxlosses=4.66W).
The T8 58W will pass the new requirement (old B1 Maxlosses=6W, New A2 Maxlosses=6.46W).

The use of two T8 18W in tandem with a 36W ballast will also pass the new requirement.

In general the magnetic ballast for lamps with a low arc voltage (like the T8 18W) will have more problem to pass the new requirement.

Edit 0n 29.01.2013:
Corrected the formula for the magnetic ballast with the 0.95 factor used to takes account of the characteristics of a lamp operated by magnetic ballast.
More info can be read here:
http://www.tridonic.com/ch/download/CELMA_Ecodesign_1st_edition_Dec2009_full.pdf

That's good news! I had previously assumed that all magnetic fluoro ballast would be banned in 2017. I suppose the availability of magnetic ballasts in 2017 and beyond will be more limited by lack of demand than by regulation: I rarely see magnetically ballasted tubes (or recently, even T8 tubes) in new installations. I think that ballasts from China would be available for a long time anyway, but I hope that quality ballast manufacturers will not discontinue the production of magnetic ballasts.

PLs, T8s, magnetic ballasts and even MH lamps, are already in fading, as almost all new indoor installations are only of one of three lightsources: T5 (Usually 54W HO [Really saves energy: replaces the lower wattage 36W T8 and causes overlighting and increased electricity consumption]), CFL (Mostly in closed fixtures that ensures that they willn't last for more than a single month) and LED.

Here in Switzerland and in Nord Italy we can still quite easily find magnetically ballasted fittings. For example last week I've bought in Italy a Sylvania sylproof superia single 58W lamp with a nice Helvar magnetic ballast (EEI B2).

Update:
VS make a new range of A2 magnetic ballast for T8 and T5 lamps which complies with the new EU directive and will not be banned in 2017.

Your discovery is both enlightening and worrying. It shows how dangerous misreading/misunderstanding a document can be, and how easily people simply followed the misunderstanding rather than challenging or examining it further. I have also read an article published in Germany, backed by numerous experiments and lab tests showing magnetic ballasts to be both efficient and reliable, and that the wording of the EU directive in question stops short of defining a technology as such to be phased out, only the least efficient of all products. Therefore magnetic ballasts, which are still being made more efficient, will remain. There is only one thing that will take them off the market - manufacturers who no longer wish to make them.

I think that the common misunderstanding is also caused by the old directive which classified as "A" class(A3, A2 or A1) only electronic ballasts and with the new directive all B class will be banned after 2017(but now also magnetic ballasts can be A2 if efficient enough).
Yes I've already read a similar article, it's  this one?
The problem is that manufacturers will make them only if the market volume will remain high enough to be profitable. Fortunately VS seem to believe it.