i have a project with lots of included code (data and macros)
format is elf64 using import64.inc

in one of the includes is

PLY_BUFFSIZE equ 1024*1024*100
ply_buffer: rb PLY_BUFFSIZE

a 100MB piece of memory

if i list all the includes in one order the binary will be over 100MB (as expected)
but if i include in another order the binary will be 22kB
and it will work

data from includes goes in "segment readable writeable" at the start of the binary

in curious why the smaller version works

The "R" in "RB" stands for "reserved". If it appears at the end of the data section it will be initialized at runtime, however, when there is defined data after it, this forces it to be initialized at compile time.

nice explained, thx

hmm
i have 15 includes
6'th is ply.inc where that 100MB reserve is defined
model.inc is last and is where a macro defined in ply.inc is used
thou the reserved memory declaration is separate from the macro

so ply.inc is 6'th and there are plenty of db, dd, dq and such in includes after it but fasm still puts it last and makes a 22k binary
but if i put model.inc as anything but last it will be 100MB, even thou the data is separate from the macro

i thought includes are flat too

I guess it is depending on sections.
If following includes are only code this would maybe not affect the data section.

almost all of them are

-------------
data
-------------
macros
-------------

so when i include them in a readable/writeable section i get the data there and macros where ever i use them