Now I have many(about 200) c files like this:

// hello.c

#include <stdio.h>

main()
{
printf("hello, world\n");

exit(0);
}

and I want to add "#include <stdlib.h>" to these files

and I try many times but can not succed

can you give me some advice ?


I am sorry ,my "Wrap [CODE]" can not use

how did you try that?

I read the c file and try to write "#include <stdlib.h>" before the function "main()"

this could the the core of a cycle:

sed 's|^#include <stdio.h>$|#include <stdio.h>\n#include <stdlib.h>|' "$FILE.before" >"$FILE"

this is a very good suggestion.thanks a lot!

i am not very familiar with the tool sed, you help me a lot!


by the way ,do you konw why my "Wrap [CODE]" could not use?

Only type [code] before your code, and [/code] after it. It's easy, just practice diligently.

OK！ thanks！

---------- Post added 09-17-14 at 03:05 AM ----------

There's one thing about learning techniques for editing multiple files via sed or awk, but don't just blindly add stdlib.h to every file because you "think" you need it. I'm a proponent of not adding include files unless I employ library calls which require those include files. Otherwise one perpetuates this highly lengthy list of include files when they don't really know or care whether or not the list is right, if there are replacements of definitions, or whether or not much of the includes are required. Many's the time I've encountered a lengthy list of includes, some of them standard library ones, but in truth many of them locally defined ones for a particular coding project; where you end up having something defined in duplicate and therefore re-defined to a value you didn't expect. I've seen an include file which checked whether or not a value was defined, performed an undefine action and then forced that value to be what the coder wished. In the end it was "odd" but not destructive; however it made us wary that two or more coders were using the same designations; they were different, and one would win if a particular include file was used. Meanwhile if the other include file was solely used, that definition would still exist, but be a different declaration. As I say, it was not a real problem, but we saw potential there for it to be.