feat: implement streq and printErr

user/redirwriter.c

@@ -0,0 +1,43 @@
+#include "ulib.h"
+
+/*
+ * Live ring-3 proof for shell.c's new `cmd > file` / `cmd >> file` file
+ * redirection (jobctlApplyRedirects/jobctlSpawnForegroundRedirected,
+ * jobctl.h): from this program's own point of view it just writes its
+ * stdout exactly like pipesrc.c already does for the pipe case -- it has
+ * no idea whether fd 1 is the real console, a pipe, or (here) a real file
+ * on tmpfs, the same "redirection is invisible to the child" property a
+ * real shell's dup2()-before-exec gives every Unix program. fd 2 is
+ * deliberately left alone (only fd 0/fd 1 are ever redirected, matching
+ * jobctlSpawnPiped's own pipeline convention), so the diagnostic line
+ * below reaches the real console even while fd 1's own payload goes to
+ * the file instead -- writing it through fd 1 would corrupt the exact
+ * bytes redirreader.c later compares.
+ *
+ * kmain.c's redirDriverThread runs this program TWICE against the same
+ * path -- once through `>` (truncate/create) and once through `>>`
+ * (append) -- so the resulting file holds two back-to-back copies of this
+ * same payload with no separator, proving `>>` genuinely continues from
+ * the end of what `>` already wrote rather than overwriting it.
+ */
+#define redirectPayload "redirect-payload-ok"
+
+static void printErr(const char* text)
+{
+    userWrite(2, text, stringLength(text));
+}
+
+int main(void)
+{
+    printErr("redirwriter: writing payload to stdout (redirected)\n");
+
+    unsigned long length = stringLength(redirectPayload);
+    long written = userWrite(1, redirectPayload, length);
+    if (written != (long)length) {
+        printErr("redirwriter: FAIL short/failed write\n");
+        return 1;
+    }
+
+    printErr("redirwriter: wrote payload\n");
+    return 0;
+}