Sat Mar 13 09:21:57 2021 UTC ()
indent: add debug logging for actually writing to the output file
Together with the results of the tokenizer and the 4 buffers for token,
label, code and comment, the debug log now provides a good high-level
view on how the indentation happens and where to look for the many
remaining bugs.
(rillig)
diff -r1.47 -r1.48 src/usr.bin/indent/indent.c
diff -r1.8 -r1.9 src/usr.bin/indent/indent.h
diff -r1.35 -r1.36 src/usr.bin/indent/io.c
diff -r1.38 -r1.39 src/usr.bin/indent/lexi.c
--- src/usr.bin/indent/indent.c 2021/03/13 00:26:56 1.47
+++ src/usr.bin/indent/indent.c 2021/03/13 09:21:57 1.48
@@ -1,4 +1,4 @@
-/* $NetBSD: indent.c,v 1.47 2021/03/13 00:26:56 rillig Exp $ */
+/* $NetBSD: indent.c,v 1.48 2021/03/13 09:21:57 rillig Exp $ */
/*-
* SPDX-License-Identifier: BSD-4-Clause
@@ -46,7 +46,7 @@
#include <sys/cdefs.h>
#ifndef lint
#if defined(__NetBSD__)
-__RCSID("$NetBSD: indent.c,v 1.47 2021/03/13 00:26:56 rillig Exp $");
+__RCSID("$NetBSD: indent.c,v 1.48 2021/03/13 09:21:57 rillig Exp $");
#elif defined(__FreeBSD__)
__FBSDID("$FreeBSD: head/usr.bin/indent/indent.c 340138 2018-11-04 19:24:49Z oshogbo $");
#endif
@@ -1389,3 +1389,46 @@
ps.want_blank = false;
}
}
+
+#ifdef debug
+void
+debug_printf(const char *fmt, ...)
+{
+ FILE *f = output == stdout ? stderr : stdout;
+ va_list ap;
+
+ va_start(ap, fmt);
+ vfprintf(f, fmt, ap);
+ va_end(ap);
+}
+
+void
+debug_println(const char *fmt, ...)
+{
+ FILE *f = output == stdout ? stderr : stdout;
+ va_list ap;
+
+ va_start(ap, fmt);
+ vfprintf(f, fmt, ap);
+ va_end(ap);
+ fprintf(f, "\n");
+}
+
+void
+debug_vis_range(const char *prefix, const char *s, const char *e,
+ const char *suffix)
+{
+ debug_printf("%s", prefix);
+ for (const char *p = s; p < e; p++) {
+ if (isprint((unsigned char)*p) && *p != '\\' && *p != '"')
+ debug_printf("%c", *p);
+ else if (*p == '\n')
+ debug_printf("\\n");
+ else if (*p == '\t')
+ debug_printf("\\t");
+ else
+ debug_printf("\\x%02x", *p);
+ }
+ debug_printf("%s", suffix);
+}
+#endif
--- src/usr.bin/indent/indent.h 2021/03/13 00:26:56 1.8
+++ src/usr.bin/indent/indent.h 2021/03/13 09:21:57 1.9
@@ -1,4 +1,4 @@
-/* $NetBSD: indent.h,v 1.8 2021/03/13 00:26:56 rillig Exp $ */
+/* $NetBSD: indent.h,v 1.9 2021/03/13 09:21:57 rillig Exp $ */
/*-
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
@@ -30,7 +30,7 @@
#if 0
#if defined(__NetBSD__)
-__RCSID("$NetBSD: indent.h,v 1.8 2021/03/13 00:26:56 rillig Exp $");
+__RCSID("$NetBSD: indent.h,v 1.9 2021/03/13 09:21:57 rillig Exp $");
#elif defined(__FreeBSD__)
__FBSDID("$FreeBSD: head/usr.bin/indent/indent.h 336333 2018-07-16 05:46:50Z pstef $");
#endif
@@ -51,7 +51,14 @@
int count_spaces_until(int, const char *, const char *);
void init_constant_tt(void);
#ifdef debug
+void debug_vis_range(const char *, const char *, const char *, const char *);
+void debug_printf(const char *, ...) __printflike(1, 2);
+void debug_println(const char *, ...) __printflike(1, 2);
const char *token_type_name(token_type);
+#else
+#define debug_printf(fmt, ...) do { } while (false)
+#define debug_println(fmt, ...) do { } while (false)
+#define debug_vis_range(prefix, s, e, suffix) do { } while (false)
#endif
token_type lexi(struct parser_state *);
void diag(int, const char *, ...) __printflike(2, 3);
--- src/usr.bin/indent/io.c 2021/03/13 09:06:12 1.35
+++ src/usr.bin/indent/io.c 2021/03/13 09:21:57 1.36
@@ -1,4 +1,4 @@
-/* $NetBSD: io.c,v 1.35 2021/03/13 09:06:12 rillig Exp $ */
+/* $NetBSD: io.c,v 1.36 2021/03/13 09:21:57 rillig Exp $ */
/*-
* SPDX-License-Identifier: BSD-4-Clause
@@ -46,7 +46,7 @@
#include <sys/cdefs.h>
#ifndef lint
#if defined(__NetBSD__)
-__RCSID("$NetBSD: io.c,v 1.35 2021/03/13 09:06:12 rillig Exp $");
+__RCSID("$NetBSD: io.c,v 1.36 2021/03/13 09:21:57 rillig Exp $");
#elif defined(__FreeBSD__)
__FBSDID("$FreeBSD: head/usr.bin/indent/io.c 334927 2018-06-10 16:44:18Z pstef $");
#endif
@@ -68,12 +68,14 @@
output_char(char ch)
{
fputc(ch, output);
+ debug_vis_range("output_char '", &ch, &ch + 1, "'\n");
}
static void
output_range(const char *s, const char *e)
{
fwrite(s, 1, (size_t)(e - s), output);
+ debug_vis_range("output_range \"", s, e, "\"\n");
}
static inline void
@@ -93,14 +95,15 @@
if (n > 0)
ind -= ind % tabsize;
for (int i = 0; i < n; i++) {
- output_char('\t');
+ fputc('\t', output);
ind += tabsize;
}
}
for (; ind < new_ind; ind++)
- output_char(' ');
+ fputc(' ', output);
+ debug_println("output_indent %d", ind);
return ind;
}
@@ -196,9 +199,19 @@
{
int i;
- for (i = 0; i < ps.p_l_follow; i++)
- if (ps.paren_indents[i] >= 0)
- ps.paren_indents[i] = -(ps.paren_indents[i] + target_col);
+ for (i = 0; i < ps.p_l_follow; i++) {
+ if (ps.paren_indents[i] >= 0) {
+ int ind = ps.paren_indents[i];
+ /*
+ * XXX: this mix of 'indent' and 'column' smells like
+ * an off-by-one error.
+ */
+ ps.paren_indents[i] = -(ind + target_col);
+ debug_println(
+ "setting pi[%d] from %d to %d for column %d",
+ i, ind, ps.paren_indents[i], target_col);
+ }
+ }
}
cur_col = 1 + output_indent(cur_col - 1, target_col - 1);
output_range(s_code, e_code);
@@ -265,8 +278,11 @@
*(e_com = s_com = combuf + 1) = '\0';
ps.ind_level = ps.i_l_follow;
ps.paren_level = ps.p_l_follow;
- if (ps.paren_level > 0)
+ if (ps.paren_level > 0) {
+ /* TODO: explain what negative indentation means */
paren_indent = -ps.paren_indents[ps.paren_level - 1];
+ debug_println("paren_indent is now %d", paren_indent);
+ }
not_first_line = 1;
}
--- src/usr.bin/indent/lexi.c 2021/03/12 23:10:18 1.38
+++ src/usr.bin/indent/lexi.c 2021/03/13 09:21:57 1.39
@@ -1,4 +1,4 @@
-/* $NetBSD: lexi.c,v 1.38 2021/03/12 23:10:18 rillig Exp $ */
+/* $NetBSD: lexi.c,v 1.39 2021/03/13 09:21:57 rillig Exp $ */
/*-
* SPDX-License-Identifier: BSD-4-Clause
@@ -46,7 +46,7 @@
#include <sys/cdefs.h>
#ifndef lint
#if defined(__NetBSD__)
-__RCSID("$NetBSD: lexi.c,v 1.38 2021/03/12 23:10:18 rillig Exp $");
+__RCSID("$NetBSD: lexi.c,v 1.39 2021/03/13 09:21:57 rillig Exp $");
#elif defined(__FreeBSD__)
__FBSDID("$FreeBSD: head/usr.bin/indent/lexi.c 337862 2018-08-15 18:19:45Z pstef $");
#endif
@@ -261,32 +261,22 @@
static void
print_buf(const char *name, const char *s, const char *e)
{
- if (s == e)
- return;
-
- printf(" %s \"", name);
- for (const char *p = s; p < e; p++) {
- if (isprint((unsigned char)*p) && *p != '\\' && *p != '"')
- printf("%c", *p);
- else if (*p == '\n')
- printf("\\n");
- else if (*p == '\t')
- printf("\\t");
- else
- printf("\\x%02x", *p);
+ if (s < e) {
+ debug_printf(" %s ", name);
+ debug_vis_range("\"", s, e, "\"");
}
- printf("\"");
}
static token_type
lexi_end(token_type code)
{
- printf("in line %d, lexi returns '%s'", line_no, token_type_name(code));
+ debug_printf("in line %d, lexi returns '%s'",
+ line_no, token_type_name(code));
print_buf("token", s_token, e_token);
print_buf("label", s_lab, e_lab);
print_buf("code", s_code, e_code);
print_buf("comment", s_com, e_com);
- printf("\n");
+ debug_printf("\n");
return code;
}