Re: How to reuse the the part of codes about file system in grub ?

[Andrei Borzenkov]

27.08.2015 18:27, Arch Stack пишет:
> I find that grub-fstest can help me to read files in different file systems
> on Windows. But I have some questions:
> 1.When I compile GRUB on Cygwin, it reports " Error: junk at end of line,
> first unrecognized character is `-' " when compile efiemu32.o. I must use
> '--disable-efiemu' when configure.
> When I compile GRUB with mingw-w64-x86_64-gcc on MSYS2, the error was
> 'Unsupported TCHAR size' when compile grub-core/osdep/windows/init.c.
> How to solve it?
> 2.I find that I can use 'make grub-fstest.exe' to compile grub-fstest, but
> it need libgnu.a. How to compile libgnu.a separately?

Theoretically it should be handled by Makefile dependencies but I have 
found in several cases that it did not properly work unless you start 
from the top target. If you are really interested in fixing that you'll 
need to debug and send patch.

> 3. If I want to use grub-fstest on Cygwin, I must change the codes in
> grub_file_open function, because /dev/sdaX was translated to
> \\.\GLOBALROOT\Device\Harddisk0\PartitionX and file_name[0] == '\\'. Can I
> compile grub-fstest without changing the codes?

Not sure I understand the question. Device names are obviously platform 
specific. On Windows raw devices are indeed named like above. Could you 
please give example of what you are trying to do?