chesslive.de ... interesting, I didn't know this one before 
After 8.Nb5,Bxc5 9.Bf4,a6 10.Bxe5,axb5 11.Qxg7,Rf8 Glek & Leconte give 12.Bf6,Qb6 13.Bd3! sacrificing a pawn and leading to an advantage for White ...  On the other hand 8...h5 was not mentionnned in their article in SOS 7, and it seems interesting I would play 9.Nd6+,Bxd6 10.Qxg7 with a complicated fight.

Back in 10/2009 I was annoying by this line too. I started to analyse it and  I found an antidote. Then I asked Mr Moskalenko what he tought about it. He was kind to answer and at the end of our discussion, he says "Anyway, I don’t see any danger in the Gledhill attack, but of course, some surprising effect." 
According to tony37 :"6.dxc5 Nxe5 7.Qg3 Nbc6 scores very well for black" ... I check in Megadatabase 2013 and found only 2 games  ... so I wonder if it really scores well for black ? Tony37 could you tell us more about this one ? Thx.
After 6.dxc5 Nxe5 7.Qg3 Nbc6 the correct move is 8.Nb5 if White wants to play for an attack. Unfortunately, as I said before, I've found a way to stop the white attack. 
Let's go : a) - 8...Be7 was my first attempt to improve the variation for black, but in fact it's a waste of time : 9.Qxg7!,Ng6 10.Nf3,a6 and now not 11.Nc3??,Bf8 12.Bg5,Bxg7 13.Bxd8,Kxd8 -+ but 11.Bd3!,axb5 12.Bxg6,Rf8 13.Bxh7,Bf6 14.Qh6 and Black is near to be busted
             b) 8...a6! is the key move. Why ? Because if you let White plays, there are threats with Qg3+Bf4+Nb5. So : 9.Nd6+,Bxd6 10.Qxg7 (10.cxd6?,0-0 11.Bf4,Qxd6 and if 12.Nf3??,Nxf3+ 13.Qxf3,Qb4+ -+) 10...Ng6 11.cxd6,Qxd6 12.Nf3,Bd7 13.Bd3 (13.Be2,0-0-0 14.0-0,Rhg8 15.Qh6,Nce7 and I really like Black here) 13...0-0-0 intending 14...e5 with an attack for Black. But Moskalenko : "in variation 8…a6 you can play 12…e5!?, instead of …Bd7." is better anyway.
It's just some of the lines I've done 4 years ago now. Am I alone ? Give us your antidote of the Gledhill too !!

I still wonder what Black should play against the Gledhill with 6.dxc5.

I think it's a dangerous line to face as Black, and in the SOS article that is mentioned in this thread practically all variations end with at least a slight edge for White.

But there has to be a good antidote, otherwise this line would be played and discussed a lot more often.

The latest books that cover the Steinitz French don't give much advice. Antic & Maksimovic and Eingorn don't even mention 5.Qg4 and Watson doesn't cover 6.dxc5.

Does anyone know what Black's best continuation after 1.e4 e6 2.d4 d5 3.Nc3 Nf6 4.e5 Nfd7 5.Qg4 c5 6.dxc5 should be?

dom's line 6...Nc6 7.Nf3 h5 at least looks quite interesting and isn't mentioned by Glek & Leconte in SOS 7.

anti-classical or..."Gledill variation"

Here are an URL (french language) about it: http://64cases.blogspot.com/2007/08/la-variante-gledhill.html ... it's the blog of Mr and Mrs Leconte. Maria Lepeina Leconte is one women french champion. Both of them know well French defence (I can testify from my lost game two weeks ago in McCutcheon).

Coming back to the subject...blog paper from August 2007 and subject is SOS N°7 (authors: Igor Glek et Jean Olivier Leconte).

One analysis of the game Marshall-Marockzy,USA 1926 is here: http://www.kenilworthchessclub.org/games/lake_hopatcong/1926/java/marshall-maroc...

I am interested to know what is White plan after 5.Qg4 c5 6.dxc5!? Nc6
Black has plan to take on e5 with tempo.

A) 7.Nf3 h5 8. Qf4 (8.Qg3 h4 9.Qg4 Ndxe5 10.Nxe5 Nxe5 11.Qf4 Nc6 and White has an early e6-e5 pawn move to free his position) Bxc5=

B) 7.Qg3 Qa5 8.Bd2 Qxc5 

C) 7.f4 Nb4 (c2 pawn is very weak) 8.Qd1 Bxc5

D) 7.Be3 h5 and difficult to play with d4 pawn fork threat