OK:
FORMAT PE GUI 6.0 on 'nul'
MajorSubSystemVersion 0x0006
MinorSubSystemVersion 0x0000

OK:
FORMAT PE GUI 6.01 on 'nul'
MajorSubSystemVersion 0x0006
MinorSubSystemVersion 0x0001

BUG:
FORMAT PE GUI 6.1 on 'nul'
MajorSubSystemVersion 0x0006
MinorSubSystemVersion 0x000A ??!


ImageBase x64:
example: vs set to 0x0000000140000000 / 0x0000000180000000 exe/dll

FORMAT PE64 GUI 6.0 DLL at 0x0000000100000000 on 'nul'

????
zzz.asm [100]:
mov qword[rax+8],xxxxxz
error: value out of range.



Resource:
directory RT_RCDATA,files

resource files,1,LANG_NEUTRAL,_test
resdata _test
file "res\test.dat"
endres

resource files,2,LANG_NEUTRAL,_test2
resdata _test2
file "res\test2.dat"
endres

????
error: symbol already defined.

already defined? this is same directory...

1. Version 6.10 is 0x6 + 0xa. Not a bug.

2. "mov qword[rax+8],1 shl 32" cannot be encoded in x86-64. Not a bug, it is a processor instruction set limitation. Use "mov reg,1 shl 32" first. Then "mov qword[rax+8],reg".

3. The resources must be defined as a single set. The exe format only supports one set of resources. Not a bug, it is just a misunderstanding of how the macros work.

1 anyway user set 6.1 and expect 6.1 not 6.10 or 6.989898 xD

2 thx, forgot that.
mov rdx,xxxxxz
mov [rax+8],rdx

3 please example how to put 1 2 3.. files in the same RT_RCDATA dir

>_<
1. That's questionable. There are multiple ways to define it:
a) 6.1 is the same as 6.01
b) 6.1 is the same as 6.10
The former way is by representing the version with two integers separated by a dot. The latter way is by representing the version with a fractional number. The documentation says it's "the minimum version of system
that the executable is targeted to (specified in form of floating-point value)". So there's no justification for the user to expect it the other way around.

3. Is the manual unclear about it? See also the fasm.pdf coming together with the fasm package.

1 fractional number lol... ms link not sets 6.10 when /version:6.1 and ulink too = bug
3 hmm NOTHING

The syntax of fasm does not necessarily follow the rules used by other tools, that's what l_inc meant when he pointed out that this behavior is correct according to how it is documented in fasm's manual, even if may be non-standard (but there are many details of fasm's syntax that are very non-standard).

The interesting point here is what was the justification of the choice of floating-point interpretation of this setting. At the time when I made initial PE output implementation, the executables that had a version 3.10 there were still the most common ones. This version (3 and 0Ah) was referring to the version of Windows that was commonly known as Windows 3.1 (but when you looked at the exact version it was something like 3.10.xxx). This led me to believe that at the time the version of Windows was in fact interpreted like a floating-point value, and thus I chose such syntax for fasm (so you could write "3.1" instead of "3.10"), with the minor version field being filled with a number of hundredths. Nowadays this remains simply as a fasm's own backward-compatibility, and perhaps the manual should openly recommend to always use the two-digit form for any non-zero minor version.

when i first time "fail" on that options - block run for winxp, but think block only win2k (that was a "fun release")
hmm win3.1.. lucky that i haven't seen her


finaly...
directory RT_RCDATA,files

resource files,1,LANG_NEUTRAL,file1,2,LANG_NEUTRAL,file2
resdata file1
file "test1.dat"
endres

resdata file2
file "test2.dat"
endres