tcc -run: support custom stdin with -rstdin FILE

When tcc -run is used with input file from stdin, then run_main's
stdin (which is tcc's stdin) is already EOF after we've read it.
If the program operates exclusively on stdin, then it cannot be used.

This commit adds an option -rstdin FILE which sets run_main's stdin.

Can also be used in such case to set run_main's stdin to the tty,
using -rstdin /dev/tty (posix) or -rstdin con (windows).

libtcc.c

@@ -1547,6 +1547,7 @@ enum {
     TCC_OPTION_rdynamic,
     TCC_OPTION_pthread,
     TCC_OPTION_run,
+    TCC_OPTION_rstdin,
     TCC_OPTION_w,
     TCC_OPTION_E,
     TCC_OPTION_M,
@@ -1571,6 +1572,30 @@ enum {
 #define TCC_OPTION_HAS_ARG 0x0001
 #define TCC_OPTION_NOSEP   0x0002 /* cannot have space before option and arg */
 
+/*
+ * in tcc_options, if opt-string A is a prefix of opt-string B,
+ * it's un-ambiguous if and only if option A is without TCC_OPTION_HAS_ARG.
+ * otherwise (A with HAS_ARG), if, for instance, A is FOO and B is FOOBAR,
+ * then "-FOOBAR" is either A with arg BAR, or B (-FOOBARX too, if B HAS_ARG).
+ *
+ * tcc_parse_args searches tcc_options in order, so if ambiguous:
+ * - if the shorter (A) is earlier: the longer (B) is completely unreachable.
+ * - else B wins, and A can't be used with adjacent arg if it also matches B.
+ *
+ * there are few clashes currently, and the longer is always earlier/reachable.
+ * when it's ambiguous, shorter-concat-arg is not useful currently.
+ * the sh(1) script 'optclash' can identifiy clashes (tcc root dir, try "-h").
+ * at the time of writing, running './optclash' prints this:
+
+    -Wl,... (1642) overrides -W... (1644)
+    -Wp,... (1643) overrides -W... (1644)
+    -dumpmachine (1630) overrides -d... (1632)
+    -dumpversion (1631) overrides -d... (1632)
+    -dynamiclib (1623) overrides -d... (1632)
+    -flat_namespace (1624) overrides -f... (1650)
+    -mfloat-abi... (1647) overrides -m... (1649)
+
+ */
 static const TCCOption tcc_options[] = {
     { "h", TCC_OPTION_HELP, 0 },
     { "-help", TCC_OPTION_HELP, 0 },
@@ -1613,6 +1638,7 @@ static const TCCOption tcc_options[] = {
     { "o", TCC_OPTION_o, TCC_OPTION_HAS_ARG },
     { "pthread", TCC_OPTION_pthread, 0},
     { "run", TCC_OPTION_run, TCC_OPTION_HAS_ARG | TCC_OPTION_NOSEP },
+    { "rstdin", TCC_OPTION_rstdin, TCC_OPTION_HAS_ARG },
     { "rdynamic", TCC_OPTION_rdynamic, 0 },
     { "r", TCC_OPTION_r, 0 },
     { "Wl,", TCC_OPTION_Wl, TCC_OPTION_HAS_ARG | TCC_OPTION_NOSEP },
@@ -1808,6 +1834,27 @@ static void args_parser_add_file(TCCState *s, const char* filename, int filetype
         ++s->nb_libraries;
 }
 
+/*  parsing is between getopt(3) and getopt_long(3), and permuting-like:
+ *  - an option is 1 or more chars.
+ *  - at most 1 option per arg in argv.
+ *  - an option in argv is "-OPT[...]" (few are --OPT, if OPT is "-...").
+ *  - optarg is next arg, or adjacent non-empty (no '='. -std=.. is arg "=..").
+ *  - supports also adjacent-only optarg (typically optional).
+ *  - supports mixed options and operands ("--" is ignored, except with -run).
+ *  - -OPT[...] can be ambiguous, which is resolved using tcc_options's order.
+ *    (see tcc_options for details)
+ *
+ *  specifically, per arg of argv, in order:
+ *  - if arg begins with '@' and is not exactly "@": process as @listfile.
+ *  - elif arg is exactly "-" or doesn't begin with '-': process as input file.
+ *    - if -run... is already set: also stop, arg... become argv of run_main.
+ *  - elif arg is "--":
+ *    - if -run... is already set: stop, arg... become argv of run_main.
+ *    - else ignore it.
+ *  - else ("-STRING") try to apply it as option, maybe with next (opt)arg.
+ *
+ *  after all args, if -run... but no "stop": run_main gets our argv (tcc ...)
+ */
 /* using * to argc/argv to let "tcc -ar" benefit from @listfile expansion */
 PUB_FUNC int tcc_parse_args(TCCState *s, int *pargc, char ***pargv)
 {
@@ -1996,6 +2043,12 @@ PUB_FUNC int tcc_parse_args(TCCState *s, int *pargc, char ***pargv)
             goto set_output_type;
 #else
             return tcc_error_noabort("-run is not available in a cross compiler");
+#endif
+#ifdef TCC_IS_NATIVE
+        case TCC_OPTION_rstdin:
+            /* custom stdin for run_main */
+            s->run_stdin = optarg;
+            break;
 #endif
         case TCC_OPTION_v:
             do ++s->verbose; while (*optarg++ == 'v');

optclash

@@ -0,0 +1,38 @@
+#!/bin/sh
+
+export LC_ALL=C
+
+defname=libtcc.c
+
+# $1 is the line number, $2... is the actual source line
+extract_opts() { awk '/tcc_options\[\]/ {x=1}; x; x && $2~/^}/ {exit}'; }
+
+case $1 in -h|--help)
+    echo "Usage: $0 [INFILE]"
+    echo "Detect tcc_options[] clashes in $defname-like INFILE."
+    echo "If INFILE is missing, use $defname at the same dir as this script."
+    echo "Clashes are reported as longer-overrides, or longer-unreachable."
+    exit
+esac
+
+f=${1-$(dirname "$0")/$defname}
+[ -r "$f" ] || { >&2 echo "$0: can't read -- $f"; exit 1; }
+
+nl -b a <"$f" | extract_opts | tr \" ' ' | awk '$2=="{"' | sort -b -k 3 |
+# "<line-num> { <unquoted-opt> <rest-of-line>"  sorted-up by opt
+# unavoidable O(N^2). the sort simplifies the awk code - only test prior opts.
+awk '
+    {
+        n=$1; opt=$3; h=/HAS_ARG/
+        for (pn in prevs) {  # pn: line-num
+            po = prevs[pn]   # po: opt-with-has-arg
+            if (index(opt,po) == 1) {
+                clash=1
+                printf("-%s%s (%d) %s -%s... (%s)\n", opt,h?"...":"",n,
+                       n>pn? "is not reachable! by":"overrides", po,pn)
+            }
+        }
+    }
+    h   {prevs[n] = opt}
+    END {if (clash) exit 1; print "no clashes"}
+'

tcc.c

@@ -35,7 +35,7 @@ static const char help[] =
     "General options:\n"
     "  -c           compile only - generate an object file\n"
     "  -o outfile   set output filename\n"
-    "  -run         run compiled source\n"
+    "  -run         run compiled source [with custom stdin: -rstdin FILE]\n"
     "  -fflag       set or reset (with 'no-' prefix) 'flag' (see tcc -hh)\n"
     "  -Wwarning    set or reset (with 'no-' prefix) 'warning' (see tcc -hh)\n"
     "  -w           disable all warnings\n"

tcc.h

@@ -976,6 +976,7 @@ struct TCCState {
     const char *run_main; /* entry for tcc_run() */
     void *run_ptr; /* runtime_memory */
     unsigned run_size; /* size of runtime_memory  */
+    const char *run_stdin; /* custom stdin file for run_main */
 #ifdef _WIN64
     void *run_function_table; /* unwind data */
 #endif

tccrun.c

@@ -233,6 +233,17 @@ LIBTCCAPI int tcc_run(TCCState *s1, int argc, char **argv)
     prog_main = (void*)get_sym_addr(s1, s1->run_main, 1, 1);
     if ((addr_t)-1 == (addr_t)prog_main)
         return -1;
+
+    /* custom stdin for run_main, mainly if stdin is/was an input file.
+     * fileno(stdin) should remain 0, as posix mandates to use the smallest
+     * free fd, which is 0 after the initial fclose in freopen. windows too.
+     * to set stdin to the tty, use /dev/tty (posix) or con (windows).
+     */
+    if (s1->run_stdin && !freopen(s1->run_stdin, "r", stdin)) {
+        tcc_error_noabort("failed to reopen stdin from '%s'", s1->run_stdin);
+        return -1;
+    }
+
     errno = 0; /* clean errno value */
     fflush(stdout);
     fflush(stderr);