Am Freitag 31 Oktober 2008 00.23:32 schrieb Wolfgang Denk:
Dear Niklaus Giger,
In message 200810231231.15846.niklaus.giger@member.fsf.org you wrote:
This time I took a closer look at img2brec.sh. My conclusions are img2brec.sh was done for development under Windows (probably cygwin), as it uses a command filesize which is not available as such under Unix. Therefore I replaced with a functional equivalent (wc --bytes) which should also work under cygwin (not tested). I modified the script to git also rid of extra \M (LF) caracters in the output. Now the outputfile on my Debian system after calling tools/img2brec.sh u-boot.bin [u-boot.brec] looks fine.
Hm.... but the \r\n sequences are probably intentional because theyu are needed on Win* systems?
I think they were needed in previous versions of Cygwin. Also often programs transforrm automatically \n into \l\n sequences. Also if the destination program does not handle correctly Unix-linefeed one would have to add something like unix2dos into the shell.
Also cygwin has a variety of options whether mount points automatically transform textfiles from/to Unix to MS/DOS-format. Therefore I cannot perform tests a Win* systems.
Do you now of any user of this batch who could answer these question?
Best regards,
Wolfgang Denk
Best regards
Niklaus Giger