Skip to contentSkip to conversation

cmd: avoid decimal conversion

PATCH New Michal Simek 4 participants 10 messages Sep 11, 2019 – Oct 12, 2019

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 cmd: avoid decimal conversion

From: T Karthik Reddy <[email protected]>

This patch uses auto instead of decimal in simple_strtoul().

---

 cmd/test.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)
cmd/test.c +12 -12
index fa0c349f0827..258bfd880653 100644
@@ -113,28 +113,28 @@ static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
 			expr = strcmp(ap[0], ap[2]) > 0;
 			break;
 		case OP_INT_EQ:
-			expr = simple_strtol(ap[0], NULL, 10) ==
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) ==
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_NEQ:
-			expr = simple_strtol(ap[0], NULL, 10) !=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) !=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_LT:
-			expr = simple_strtol(ap[0], NULL, 10) <
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_LE:
-			expr = simple_strtol(ap[0], NULL, 10) <=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_GT:
-			expr = simple_strtol(ap[0], NULL, 10) >
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_GE:
-			expr = simple_strtol(ap[0], NULL, 10) >=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_FILE_EXISTS:
 			expr = file_exists(ap[1], ap[2], ap[3], FS_TYPE_ANY);
Signature · 1 line
2.17.1
#1 lore raw

PATCH 1/1 cmd: avoid decimal conversion

Quoted 46 lines · Michal Simek

From: T Karthik Reddy <[email protected]>

This patch uses auto instead of decimal in simple_strtoul().

---

 cmd/test.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)
cmd/test.c +9 -10
index fa0c349f0827..258bfd880653 100644
@@ -113,28 +113,28 @@ static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
 			expr = strcmp(ap[0], ap[2]) > 0;
 			break;
 		case OP_INT_EQ:
-			expr = simple_strtol(ap[0], NULL, 10) ==
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) ==
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_NEQ:
-			expr = simple_strtol(ap[0], NULL, 10) !=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) !=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_LT:
-			expr = simple_strtol(ap[0], NULL, 10) <
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_LE:
-			expr = simple_strtol(ap[0], NULL, 10) <=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_GT:
-			expr = simple_strtol(ap[0], NULL, 10) >
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >

cmd/test.c static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) · line 140

Quoted 10 lines
cmd/test.c +3 -2
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_GE:
-			expr = simple_strtol(ap[0], NULL, 10) >=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_FILE_EXISTS:
 			expr = file_exists(ap[1], ap[2], ap[3], FS_TYPE_ANY);

I'm going to NAK this, but could be argued around to changing my mind. While it's true that in general command inputs are hex and not decimal, this has been decimal since introduction in 2009. So changing it now is breaking ABI and other peoples test scripts, so I don't think we can do this, sorry.

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20190913/da42964f/attachment.sig>
#2 lore raw

PATCH 1/1 cmd: avoid decimal conversion

Quoted 57 lines · Tom Rini
Quoted 54 lines · Michal Simek

From: T Karthik Reddy <[email protected]>

This patch uses auto instead of decimal in simple_strtoul().

---

 cmd/test.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)
cmd/test.c +12 -12
index fa0c349f0827..258bfd880653 100644
@@ -113,28 +113,28 @@ static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
                      expr = strcmp(ap[0], ap[2]) > 0;
                      break;
              case OP_INT_EQ:
-                     expr = simple_strtol(ap[0], NULL, 10) ==
-                                     simple_strtol(ap[2], NULL, 10);
+                     expr = simple_strtol(ap[0], NULL, 0) ==
+                                     simple_strtol(ap[2], NULL, 0);
                      break;
              case OP_INT_NEQ:
-                     expr = simple_strtol(ap[0], NULL, 10) !=
-                                     simple_strtol(ap[2], NULL, 10);
+                     expr = simple_strtol(ap[0], NULL, 0) !=
+                                     simple_strtol(ap[2], NULL, 0);
                      break;
              case OP_INT_LT:
-                     expr = simple_strtol(ap[0], NULL, 10) <
-                                     simple_strtol(ap[2], NULL, 10);
+                     expr = simple_strtol(ap[0], NULL, 0) <
+                                     simple_strtol(ap[2], NULL, 0);
                      break;
              case OP_INT_LE:
-                     expr = simple_strtol(ap[0], NULL, 10) <=
-                                     simple_strtol(ap[2], NULL, 10);
+                     expr = simple_strtol(ap[0], NULL, 0) <=
+                                     simple_strtol(ap[2], NULL, 0);
                      break;
              case OP_INT_GT:
-                     expr = simple_strtol(ap[0], NULL, 10) >
-                                     simple_strtol(ap[2], NULL, 10);
+                     expr = simple_strtol(ap[0], NULL, 0) >
+                                     simple_strtol(ap[2], NULL, 0);
                      break;
              case OP_INT_GE:
-                     expr = simple_strtol(ap[0], NULL, 10) >=
-                                     simple_strtol(ap[2], NULL, 10);
+                     expr = simple_strtol(ap[0], NULL, 0) >=
+                                     simple_strtol(ap[2], NULL, 0);
                      break;
Quoted 8 lines

cmd/test.c static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) · line 140

Quoted 2 lines
cmd/test.c +0 -0
              case OP_FILE_EXISTS:
                      expr = file_exists(ap[1], ap[2], ap[3], FS_TYPE_ANY);

I'm going to NAK this, but could be argued around to changing my mind. While it's true that in general command inputs are hex and not decimal, this has been decimal since introduction in 2009. So changing it now is breaking ABI and other peoples test scripts, so I don't think we can do this, sorry.

Well, I think that this is automatic. There should be a back compatibility with the ABI if base is 10

Michael

Quoted 7 lines

--
Tom
_______________________________________________
U-Boot mailing list
U-Boot at lists.denx.de
https://lists.denx.de/listinfo/u-boot

--

| Michael Nazzareno Trimarchi                     Amarula Solutions BV |
| COO  -  Founder                                      Cruquiuskade 47 |
| +31(0)851119172                                 Amsterdam 1018 AM NL |
|                  [`as] http://www.amarulasolutions.com               |
#3 lore raw

PATCH 1/1 cmd: avoid decimal conversion

In reply to Tom Rini
Quoted 56 lines · Tom Rini
Quoted 54 lines · Michal Simek

From: T Karthik Reddy <[email protected]>

This patch uses auto instead of decimal in simple_strtoul().

---

 cmd/test.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)
cmd/test.c +12 -12
index fa0c349f0827..258bfd880653 100644
@@ -113,28 +113,28 @@ static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
 			expr = strcmp(ap[0], ap[2]) > 0;
 			break;
 		case OP_INT_EQ:
-			expr = simple_strtol(ap[0], NULL, 10) ==
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) ==
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_NEQ:
-			expr = simple_strtol(ap[0], NULL, 10) !=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) !=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_LT:
-			expr = simple_strtol(ap[0], NULL, 10) <
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_LE:
-			expr = simple_strtol(ap[0], NULL, 10) <=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_GT:
-			expr = simple_strtol(ap[0], NULL, 10) >
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_GE:
-			expr = simple_strtol(ap[0], NULL, 10) >=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >=
+					simple_strtol(ap[2], NULL, 0);
 			break;
Quoted 8 lines

cmd/test.c static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) · line 140

Quoted 2 lines
cmd/test.c +0 -0
 		case OP_FILE_EXISTS:
 			expr = file_exists(ap[1], ap[2], ap[3], FS_TYPE_ANY);

I'm going to NAK this, but could be argued around to changing my mind. While it's true that in general command inputs are hex and not decimal, this has been decimal since introduction in 2009. So changing it now is breaking ABI and other peoples test scripts, so I don't think we can do this, sorry.

I also think that this is not breaking any ABI. test_hush_if_test.py is around for a while to capture issues in this space and I can't see any single failure in connection to this change.

If this accepted then we can add more tests like this

    ('test 0x2000000 -gt 0x2000001', False),
    ('test 0x2000000 -gt 0x2000000', False),
    ('test 0x2000000 -gt 0x1ffffff', True),
    ('test 2000000 -gt 0x1ffffff', False),
    ('test 0x2000000 -gt 1ffffff', True),

    ('test 0x2000000 -lt 1ffffff', False),
    ('test 0x2000000 -eq 2000000', False),
    ('test 0x2000000 -ne 2000000', True),

where some of them are failing without this patch ... test_hush_if_test[test 0x2000000 -gt 0x1ffffff-True]

... test_hush_if_test[test 2000000 -gt 0x1ffffff-False]

... test_hush_if_test[test 0x2000000 -gt 1ffffff-True]

... test_hush_if_test[test 0x2000000 -lt 1ffffff-False]

Thanks,
Michal

#4 lore raw

PATCH 1/1 cmd: avoid decimal conversion

Hi Tom,

Quoted 81 lines · Michal Simek
Quoted 56 lines · Tom Rini
Quoted 54 lines · Michal Simek

From: T Karthik Reddy <[email protected]>

This patch uses auto instead of decimal in simple_strtoul().

---

 cmd/test.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)
cmd/test.c +12 -12
index fa0c349f0827..258bfd880653 100644
@@ -113,28 +113,28 @@ static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
 			expr = strcmp(ap[0], ap[2]) > 0;
 			break;
 		case OP_INT_EQ:
-			expr = simple_strtol(ap[0], NULL, 10) ==
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) ==
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_NEQ:
-			expr = simple_strtol(ap[0], NULL, 10) !=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) !=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_LT:
-			expr = simple_strtol(ap[0], NULL, 10) <
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_LE:
-			expr = simple_strtol(ap[0], NULL, 10) <=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_GT:
-			expr = simple_strtol(ap[0], NULL, 10) >
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_GE:
-			expr = simple_strtol(ap[0], NULL, 10) >=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >=
+					simple_strtol(ap[2], NULL, 0);
 			break;
Quoted 8 lines

cmd/test.c static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) · line 140

Quoted 2 lines
cmd/test.c +0 -0
 		case OP_FILE_EXISTS:
 			expr = file_exists(ap[1], ap[2], ap[3], FS_TYPE_ANY);

I'm going to NAK this, but could be argued around to changing my mind. While it's true that in general command inputs are hex and not decimal, this has been decimal since introduction in 2009. So changing it now is breaking ABI and other peoples test scripts, so I don't think we can do this, sorry.

I also think that this is not breaking any ABI. test_hush_if_test.py is around for a while to capture issues in this space and I can't see any single failure in connection to this change.

If this accepted then we can add more tests like this

    ('test 0x2000000 -gt 0x2000001', False),
    ('test 0x2000000 -gt 0x2000000', False),
    ('test 0x2000000 -gt 0x1ffffff', True),
    ('test 2000000 -gt 0x1ffffff', False),
    ('test 0x2000000 -gt 1ffffff', True),

    ('test 0x2000000 -lt 1ffffff', False),
    ('test 0x2000000 -eq 2000000', False),
    ('test 0x2000000 -ne 2000000', True),
Quoted 9 lines

where some of them are failing without this patch ... test_hush_if_test[test 0x2000000 -gt 0x1ffffff-True]

... test_hush_if_test[test 2000000 -gt 0x1ffffff-False]

... test_hush_if_test[test 0x2000000 -gt 1ffffff-True]

... test_hush_if_test[test 0x2000000 -lt 1ffffff-False]

Any comment on this?

Thanks,
Michal

#5 lore raw

PATCH 1/1 cmd: avoid decimal conversion

Quoted 87 lines · Michal Simek

Hi Tom,

Quoted 84 lines · Michal Simek
Quoted 56 lines · Tom Rini
Quoted 54 lines · Michal Simek

From: T Karthik Reddy <[email protected]>

This patch uses auto instead of decimal in simple_strtoul().

---

 cmd/test.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)
cmd/test.c +12 -12
index fa0c349f0827..258bfd880653 100644
@@ -113,28 +113,28 @@ static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
 			expr = strcmp(ap[0], ap[2]) > 0;
 			break;
 		case OP_INT_EQ:
-			expr = simple_strtol(ap[0], NULL, 10) ==
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) ==
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_NEQ:
-			expr = simple_strtol(ap[0], NULL, 10) !=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) !=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_LT:
-			expr = simple_strtol(ap[0], NULL, 10) <
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_LE:
-			expr = simple_strtol(ap[0], NULL, 10) <=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <=
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_GT:
-			expr = simple_strtol(ap[0], NULL, 10) >
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >
+					simple_strtol(ap[2], NULL, 0);
 			break;
 		case OP_INT_GE:
-			expr = simple_strtol(ap[0], NULL, 10) >=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >=
+					simple_strtol(ap[2], NULL, 0);
 			break;
Quoted 8 lines

cmd/test.c static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) · line 140

Quoted 2 lines
cmd/test.c +0 -0
 		case OP_FILE_EXISTS:
 			expr = file_exists(ap[1], ap[2], ap[3], FS_TYPE_ANY);

I'm going to NAK this, but could be argued around to changing my mind. While it's true that in general command inputs are hex and not decimal, this has been decimal since introduction in 2009. So changing it now is breaking ABI and other peoples test scripts, so I don't think we can do this, sorry.

I also think that this is not breaking any ABI. test_hush_if_test.py is around for a while to capture issues in this space and I can't see any single failure in connection to this change.

If this accepted then we can add more tests like this

    ('test 0x2000000 -gt 0x2000001', False),
    ('test 0x2000000 -gt 0x2000000', False),
    ('test 0x2000000 -gt 0x1ffffff', True),
    ('test 2000000 -gt 0x1ffffff', False),
    ('test 0x2000000 -gt 1ffffff', True),

    ('test 0x2000000 -lt 1ffffff', False),
    ('test 0x2000000 -eq 2000000', False),
    ('test 0x2000000 -ne 2000000', True),

where some of them are failing without this patch ... test_hush_if_test[test 0x2000000 -gt 0x1ffffff-True]

Quoted 8 lines
Quoted 6 lines

... test_hush_if_test[test 2000000 -gt 0x1ffffff-False]

... test_hush_if_test[test 0x2000000 -gt 1ffffff-True]

... test_hush_if_test[test 0x2000000 -lt 1ffffff-False]

Any comment on this?

Sorry, yes, OK, we can take this then. I should have this in my first batch of non-python general changes I grab.

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20191009/81ec862b/attachment-0001.sig>
#6 lore raw

PATCH 1/1 cmd: avoid decimal conversion

Am 09.10.2019 um 18:26 schrieb Tom Rini:

Quoted 92 lines · Michal Simek
Quoted 91 lines

Hi Tom,

Quoted 88 lines · Michal Simek
Quoted 56 lines · Tom Rini
Quoted 54 lines · Michal Simek

From: T Karthik Reddy <[email protected]>

This patch uses auto instead of decimal in simple_strtoul().

---

  cmd/test.c | 24 ++++++++++++------------
  1 file changed, 12 insertions(+), 12 deletions(-)
cmd/test.c +12 -12
index fa0c349f0827..258bfd880653 100644
@@ -113,28 +113,28 @@ static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])
  			expr = strcmp(ap[0], ap[2]) > 0;
  			break;
  		case OP_INT_EQ:
-			expr = simple_strtol(ap[0], NULL, 10) ==
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) ==
+					simple_strtol(ap[2], NULL, 0);
  			break;
  		case OP_INT_NEQ:
-			expr = simple_strtol(ap[0], NULL, 10) !=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) !=
+					simple_strtol(ap[2], NULL, 0);
  			break;
  		case OP_INT_LT:
-			expr = simple_strtol(ap[0], NULL, 10) <
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <
+					simple_strtol(ap[2], NULL, 0);
  			break;
  		case OP_INT_LE:
-			expr = simple_strtol(ap[0], NULL, 10) <=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) <=
+					simple_strtol(ap[2], NULL, 0);
  			break;
  		case OP_INT_GT:
-			expr = simple_strtol(ap[0], NULL, 10) >
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >
+					simple_strtol(ap[2], NULL, 0);
  			break;
  		case OP_INT_GE:
-			expr = simple_strtol(ap[0], NULL, 10) >=
-					simple_strtol(ap[2], NULL, 10);
+			expr = simple_strtol(ap[0], NULL, 0) >=
+					simple_strtol(ap[2], NULL, 0);
  			break;
Quoted 8 lines

cmd/test.c static int do_test(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) · line 140

Quoted 2 lines
cmd/test.c +0 -0
  		case OP_FILE_EXISTS:
  			expr = file_exists(ap[1], ap[2], ap[3], FS_TYPE_ANY);

I'm going to NAK this, but could be argued around to changing my mind. While it's true that in general command inputs are hex and not decimal, this has been decimal since introduction in 2009. So changing it now is breaking ABI and other peoples test scripts, so I don't think we can do this, sorry.

I also think that this is not breaking any ABI. test_hush_if_test.py is around for a while to capture issues in this space and I can't see any single failure in connection to this change.

If this accepted then we can add more tests like this

     ('test 0x2000000 -gt 0x2000001', False),
     ('test 0x2000000 -gt 0x2000000', False),
     ('test 0x2000000 -gt 0x1ffffff', True),
     ('test 2000000 -gt 0x1ffffff', False),
     ('test 0x2000000 -gt 1ffffff', True),

     ('test 0x2000000 -lt 1ffffff', False),
     ('test 0x2000000 -eq 2000000', False),
     ('test 0x2000000 -ne 2000000', True),

where some of them are failing without this patch ... test_hush_if_test[test 0x2000000 -gt 0x1ffffff-True]

... test_hush_if_test[test 2000000 -gt 0x1ffffff-False]

... test_hush_if_test[test 0x2000000 -gt 1ffffff-True]

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

... test_hush_if_test[test 0x2000000 -lt 1ffffff-False]

Any comment on this?

Sorry, yes, OK, we can take this then. I should have this in my first batch of non-python general changes I grab.

But strtoul("010", NULL, 0) is 8 while strtoul("010", NULL, 10) is 10. Do we care for that change?

Regards,
Simon

#7 lore raw

PATCH 1/1 cmd: avoid decimal conversion

Quoted 100 lines · Simon Goldschmidt

Am 09.10.2019 um 18:26 schrieb Tom Rini:

Quoted 99 lines · Michal Simek
Quoted 98 lines

Hi Tom,

Quoted 94 lines · Michal Simek
Quoted 60 lines · Tom Rini
Quoted 49 lines · Michal Simek

From: T Karthik Reddy <[email protected]>

This patch uses auto instead of decimal in simple_strtoul().

---

  cmd/test.c | 24 ++++++++++++------------

  1 file changed, 12 insertions(+), 12 deletions(-)
cmd/test.c +0 -0
index fa0c349f0827..258bfd880653 100644
@@ -113,28 +113,28 @@ static int do_test(cmd_tbl_t *cmdtp, int
 flag, int argc, char * const argv[])
               expr = strcmp(ap[0], ap[2]) > 0;
               break;

          case OP_INT_EQ:
-            expr = simple_strtol(ap[0], NULL, 10) == -                    simple_strtol(ap[2], NULL, 10); +            expr = simple_strtol(ap[0], NULL, 0) == +                    simple_strtol(ap[2], NULL, 0); break;
          case OP_INT_NEQ:
-            expr = simple_strtol(ap[0], NULL, 10) != -                    simple_strtol(ap[2], NULL, 10); +            expr = simple_strtol(ap[0], NULL, 0) != +                    simple_strtol(ap[2], NULL, 0); break;
          case OP_INT_LT:
-            expr = simple_strtol(ap[0], NULL, 10) < -                    simple_strtol(ap[2], NULL, 10); +            expr = simple_strtol(ap[0], NULL, 0) < +                    simple_strtol(ap[2], NULL, 0); break;
          case OP_INT_LE:
-            expr = simple_strtol(ap[0], NULL, 10) <= -                    simple_strtol(ap[2], NULL, 10); +            expr = simple_strtol(ap[0], NULL, 0) <= +                    simple_strtol(ap[2], NULL, 0); break;
          case OP_INT_GT:
-            expr = simple_strtol(ap[0], NULL, 10) > -                    simple_strtol(ap[2], NULL, 10); +            expr = simple_strtol(ap[0], NULL, 0) > +                    simple_strtol(ap[2], NULL, 0); break;

Quoted 8 lines

          case OP_INT_GE:
-            expr = simple_strtol(ap[0], NULL, 10) >=
-                    simple_strtol(ap[2], NULL, 10);
+            expr = simple_strtol(ap[0], NULL, 0) >=
+                    simple_strtol(ap[2], NULL, 0); break;
          case OP_FILE_EXISTS:
              expr = file_exists(ap[1], ap[2], ap[3], FS_TYPE_ANY);

Quoted 8 lines

I'm going to NAK this, but could be argued around to changing my mind. While it's true that in general command inputs are hex and not decimal,
this has been decimal since introduction in 2009.  So changing it now is
breaking ABI and other peoples test scripts, so I don't think we can do
this, sorry.

I also think that this is not breaking any ABI. test_hush_if_test.py is around for a while to capture issues in this space and I can't see any single failure in connection to this change.

If this accepted then we can add more tests like this ('test 0x2000000 -gt 0x2000001', False), ('test 0x2000000 -gt 0x2000000', False), ('test 0x2000000 -gt 0x1ffffff', True), ('test 2000000 -gt 0x1ffffff', False), ('test 0x2000000 -gt 1ffffff', True),

     ('test 0x2000000 -lt 1ffffff', False),
     ('test 0x2000000 -eq 2000000', False),
     ('test 0x2000000 -ne 2000000', True),

where some of them are failing without this patch ... test_hush_if_test[test 0x2000000 -gt 0x1ffffff-True]

... test_hush_if_test[test 2000000 -gt 0x1ffffff-False]

... test_hush_if_test[test 0x2000000 -gt 1ffffff-True]

... test_hush_if_test[test 0x2000000 -lt 1ffffff-False]

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

Any comment on this?

Sorry, yes, OK, we can take this then.  I should have this in my first batch of non-python general changes I grab.

But strtoul("010", NULL, 0) is 8 while strtoul("010", NULL, 10) is 10. Do we care for that change?

Leading zeroes are problematic

Before the patch this pass
test 010 -eq 10 && echo yes

After the patch this fail.

I can't see any test in test/py/tests/test_hush_if_test.py to cover this case that's why I can't guess if this is used or not.

Thanks,
Michal

#8 lore raw

PATCH 1/1 cmd: avoid decimal conversion

Quoted 112 lines · Michal Simek
Quoted 100 lines · Simon Goldschmidt

Am 09.10.2019 um 18:26 schrieb Tom Rini:

Quoted 99 lines · Michal Simek
Quoted 98 lines

Hi Tom,

Quoted 94 lines · Michal Simek
Quoted 60 lines · Tom Rini
Quoted 56 lines · Michal Simek

From: T Karthik Reddy <[email protected]>

This patch uses auto instead of decimal in simple_strtoul().

---

  cmd/test.c | 24 ++++++++++++------------
  1 file changed, 12 insertions(+), 12 deletions(-)
cmd/test.c +12 -12
index fa0c349f0827..258bfd880653 100644
@@ -113,28 +113,28 @@ static int do_test(cmd_tbl_t *cmdtp, int
 flag, int argc, char * const argv[])
              expr = strcmp(ap[0], ap[2]) > 0;
              break;
          case OP_INT_EQ:
-            expr = simple_strtol(ap[0], NULL, 10) ==
-                    simple_strtol(ap[2], NULL, 10);
+            expr = simple_strtol(ap[0], NULL, 0) ==
+                    simple_strtol(ap[2], NULL, 0);
              break;
          case OP_INT_NEQ:
-            expr = simple_strtol(ap[0], NULL, 10) !=
-                    simple_strtol(ap[2], NULL, 10);
+            expr = simple_strtol(ap[0], NULL, 0) !=
+                    simple_strtol(ap[2], NULL, 0);
              break;
          case OP_INT_LT:
-            expr = simple_strtol(ap[0], NULL, 10) <
-                    simple_strtol(ap[2], NULL, 10);
+            expr = simple_strtol(ap[0], NULL, 0) <
+                    simple_strtol(ap[2], NULL, 0);
              break;
          case OP_INT_LE:
-            expr = simple_strtol(ap[0], NULL, 10) <=
-                    simple_strtol(ap[2], NULL, 10);
+            expr = simple_strtol(ap[0], NULL, 0) <=
+                    simple_strtol(ap[2], NULL, 0);
              break;
          case OP_INT_GT:
-            expr = simple_strtol(ap[0], NULL, 10) >
-                    simple_strtol(ap[2], NULL, 10);
+            expr = simple_strtol(ap[0], NULL, 0) >
+                    simple_strtol(ap[2], NULL, 0);
              break;
          case OP_INT_GE:
-            expr = simple_strtol(ap[0], NULL, 10) >=
-                    simple_strtol(ap[2], NULL, 10);
+            expr = simple_strtol(ap[0], NULL, 0) >=
+                    simple_strtol(ap[2], NULL, 0);
              break;
          case OP_FILE_EXISTS:
Quoted 1 line
              expr = file_exists(ap[1], ap[2], ap[3], FS_TYPE_ANY);
Quoted 8 lines

I'm going to NAK this, but could be argued around to changing my mind. While it's true that in general command inputs are hex and not decimal,
this has been decimal since introduction in 2009. So changing it now is
breaking ABI and other peoples test scripts, so I don't think we can do
this, sorry.

I also think that this is not breaking any ABI. test_hush_if_test.py is around for a while to capture issues in this space and I can't see any single failure in connection to this change.

If this accepted then we can add more tests like this

     ('test 0x2000000 -gt 0x2000001', False),
     ('test 0x2000000 -gt 0x2000000', False),
     ('test 0x2000000 -gt 0x1ffffff', True),
     ('test 2000000 -gt 0x1ffffff', False),
     ('test 0x2000000 -gt 1ffffff', True),

     ('test 0x2000000 -lt 1ffffff', False),
     ('test 0x2000000 -eq 2000000', False),
     ('test 0x2000000 -ne 2000000', True),

where some of them are failing without this patch ... test_hush_if_test[test 0x2000000 -gt 0x1ffffff-True]

... test_hush_if_test[test 2000000 -gt 0x1ffffff-False]

... test_hush_if_test[test 0x2000000 -gt 1ffffff-True]

... test_hush_if_test[test 0x2000000 -lt 1ffffff-False]

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

Any comment on this?

Sorry, yes, OK, we can take this then. I should have this in my first batch of non-python general changes I grab.

But strtoul("010", NULL, 0) is 8 while strtoul("010", NULL, 10) is 10. Do we care for that change?

Leading zeroes are problematic

Quoted 7 lines

Before the patch this pass
test 010 -eq 10 && echo yes

After the patch this fail.

I can't see any test in test/py/tests/test_hush_if_test.py to cover this case that's why I can't guess if this is used or not.

Hmm, it might not be a problem then...?

Regards,
Simon

#9 lore raw

PATCH 1/1 cmd: avoid decimal conversion

In reply to Michal Simek
Quoted 6 lines · Michal Simek

From: T Karthik Reddy <[email protected]>

This patch uses auto instead of decimal in simple_strtoul().

Applied to u-boot/master, thanks!

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20191012/553888ff/attachment.sig>