Fixed compile error on latest GCC (14.2.1)

[NickyBoy89]

This PR contains a fix for a compile error that I noticed while compiling Cobra on my Linux machine:

cc -I. -Wall -O2 -pedantic -Werror -Wshadow -std=c99 -DYY_NO_INPUT -Wformat-truncation=1   -c -o cobra_prim.o cobra_prim.c
cobra_prim.c: In function ‘dogrep’:
cobra_prim.c:158:75: error: ‘%s’ directive output may be truncated writing up to 2030 bytes into a region of size between 0 and 2030 [-Werror=format-truncation=]
  158 |                         snprintf(cmd, sizeof(cmd), "grep -q -n -e  \"%s\" %s",
      |                                                                           ^~
cobra_prim.c:158:25: note: ‘snprintf’ output between 19 and 4079 bytes into a destination of size 2048
  158 |                         snprintf(cmd, sizeof(cmd), "grep -q -n -e  \"%s\" %s",
      |                         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  159 |                                 s, f->s);
      |                                 ~~~~~~~~
cc1: all warnings being treated as errors
make: *** [<builtin>: cobra_prim.o] Error 1

It appears that in the dogrep function, the compiler cannot adequately reason about the bounds check that we make to make sure that the resulting command that will be run through grep will not overflow the buffer that was created for it.

This PR separates those steps out into separate variables, and includes a minor refactor to re-copy the grep command into the buffer, instead of modifying the 5th and 6th indexes of the string to remove a flag before it is run again. This I thought made the intent of the code more readable, at the cost of an additional snprintf.

System Details

• OS: Arch Linux, Kernel 6.11.6-arch1-1
• gcc --version

  gcc (GCC) 14.2.1 20240910

[nimble-code]

it's interesting that the modified version doesn't trigger the same warning, since there's no more information there.
I believe this same warning came up in an earlier version but went away when the test just before the snprintf was added.
which version of gcc did you use?

[NickyBoy89]

Hmm, it's interesting that the problem seems intermittent, and seems to be very version-dependent. I remember solving a similar type of issue in #65, just in a different place, so something as simple as separating out the variable name fixes it.

I'm using GCC gcc (GCC) 14.2.1 20240910 (also in the original PR text)

Perhaps it would be worth looking at multiple callsites for snprintf at once next time?