On Fri, May 18, 2007 at 06:33:42PM +0700, Darmawan Salihun wrote: > Because this version of flashrom will be quite different in terms > of code due to inherent differences between Linux/other Unix and > Windows, I think it's better to put it as a "branch" of the > flashrom directory.
Please don't make a subdirectory of flashrom. I would instead suggest creating a new directory in util/ e.g. named winflashrom or whatever you prefer. > I'm still thinking about the way to incorporate the flash chip > and motherboard support in as seamless way as possible so that > every newer chip/motherboard support will be merged easily. I can come up with two ways: 1. Use a database separate from the programs (as suggest by Stefan) 2. Create a library used by both programs. I prefer 2 a lot; * No database file to keep track of, the binary ALWAYS works * For winflashrom the database needs to be sent to the kernel driver somehow, if not right now then I think at least in the future. That means extra work, when it could just be in code already. The benefit of 1 is that the database can be updated at any time. But I think at least the Windows version should have more intelligence in the kernel driver than in the application so upgrading will be a largeish operation anyway. > In the mean time I need the information on how to commit the code > once I have a tested version of the Winflashrom. What you do is send a signed off patch to the mailing list. Use svn diff to produce it. Then people on the list review and if it looks good it gets committed. :) > Anyway, I wonder how the qa (Quality Assurance / auto build, etc.) > system will handle a windows version source code :-? especially the > header files. What is your current development environment? I would like it a lot if you could write the software so that it is possible to build with MinGW, then it could be cross-compiled easily and binaries could be generated automatically. //Peter -- linuxbios mailing list [email protected] http://www.linuxbios.org/mailman/listinfo/linuxbios
