Looks like it’s linked to pdb file being locked, which sometimes happen with more than one instance of Visual trying to access the same pdb file.
=> running with -j1 will most probably solve the issue (or better, try to understand synchronisation point in your makefile, which can happen to be tricky)

well, I tried to compile it again and got this error:

well, in screenshot, I only see a warning, it even looks like the linking happened correctly : are you sure you got a real issue in your compilation ?

Florent

well, the compilation stopped at this point
it was the second time I tried to compile it for MSVC2010
I guess the problem is with the ‘arthurplugin’, but I dont know how to solve it… =/

Have you tried it without jom?

// i have built 4.7.2 two times and it didn’t failed, and previous versions many times without any fails

// for me default options on configure are enough, and then i use nmake sub-scr (builds the framework) then nmake sub-tools (builds Creator and rest of the tools) and then nmake clean (reduces the folder size for ~4.5 GB to ~1.5)
//of course every command is launched from the VS cmd ;)

LE:
I don’t know what jom does, but on my system nmake uses both cores, and on top of that the processor is not the determinant factor in building large frameworks, but the HDD.

it the error is upper, you can try launching the compilation with > log.txt 2>err.txt

=> this willl produce 2 files, one containing stdout (= standard output) while the other one will only contain error and warnings.

Might be worth taking this into account as well:
http://msdn.microsoft.com/fr-fr/library/bdscwf1c(v=vs.80).aspx

And raising as recommended by warning your limit to 500 (like in CXXFLAGS)