Re: [PATCH v3][GSOC] fsck: use bitwise-or assignment operator to set flag

2 messages, 2 authors, 2016-06-15 · open the first message on its own page

Re: [PATCH v3][GSOC] fsck: use bitwise-or assignment operator to set flag

From: Junio C Hamano <hidden>
Date: 2016-06-15 23:00:23

Hiroyuki Sano [off-list ref] writes:
fsck_tree() has two different ways to set a flag,
which are the followings:

  1. Using a if-statement that guards assignment.

  2. Using a bitwise-or assignment operator.

Currently, many with the former way,
and one with the latter way.

In this patch, unify them to the latter way,
because it makes the code shorter and easier to read,
and it is brief and to the point.
Two issues:

 * "In this patch," is redundant.

 * "it is brief and to the point" are equally applicable to both
   styles, so that is not a *reason* to choose one over the other.

If a condition were *not* brief and to the point, then a rewrite to
the latter style will make the resulting code worse:

	if (a very complex condition
            that potentially have to consume a
            lot of brain-cycles to understand) {
		has_that_condition = 1;
	}

is a lot easier to extend than

	has_that_condition = (a very complex condition
                              that potentially have to consume a
                              lot of brain-cycles to understand);

because it is a lot more likely that we would need to later extend
such a complex condition is more likely than a simple singleton
condition, and we could end up with

	if (a very complex condition
            that potentially have to consume a
            lot of brain-cycles to understand) {
		futher computation to check if
                the condition really holds
                will be added here later
                if (does that condition really hold true?)
			has_that_condition = 1;
	}


which may be harder to express in the latter form.

In other words, "it is brief and to the point" merely _allows_ these
statements to be expressed in the latter form; it does not say
anything about which is better between the former and the latter.
quoted hunk
Signed-off-by: Hiroyuki Sano <redacted>
---
 fsck.c | 18 ++++++------------
 1 file changed, 6 insertions(+), 12 deletions(-)
diff --git a/fsck.c b/fsck.c
index b3022ad..abed62b 100644
--- a/fsck.c
+++ b/fsck.c
@@ -165,18 +165,12 @@ static int fsck_tree(struct tree *item, int strict, fsck_error error_func)
 
 		sha1 = tree_entry_extract(&desc, &name, &mode);
 
-		if (is_null_sha1(sha1))
-			has_null_sha1 = 1;
-		if (strchr(name, '/'))
-			has_full_path = 1;
-		if (!*name)
-			has_empty_name = 1;
-		if (!strcmp(name, "."))
-			has_dot = 1;
-		if (!strcmp(name, ".."))
-			has_dotdot = 1;
-		if (!strcmp(name, ".git"))
-			has_dotgit = 1;
+		has_null_sha1 |= is_null_sha1(sha1);
+		has_full_path |= !!strchr(name, '/');
+		has_empty_name |= !*name;
+		has_dot |= !strcmp(name, ".");
+		has_dotdot |= !strcmp(name, "..");
+		has_dotgit |= !strcmp(name, ".git");
 		has_zero_pad |= *(char *)desc.buffer == '0';
 		update_tree_entry(&desc);

Re: [PATCH v3][GSOC] fsck: use bitwise-or assignment operator to set flag

From: Hiroyuki <hidden>
Date: 2016-06-15 23:00:24

Junio and Eric,

Thank you for the reviews and helpful advice.
I should have read more past commit messages before send patch.


Regards,


2014-03-21 3:20 GMT+09:00 Junio C Hamano [off-list ref]:
Hiroyuki Sano [off-list ref] writes:
quoted
fsck_tree() has two different ways to set a flag,
which are the followings:

  1. Using a if-statement that guards assignment.

  2. Using a bitwise-or assignment operator.

Currently, many with the former way,
and one with the latter way.

In this patch, unify them to the latter way,
because it makes the code shorter and easier to read,
and it is brief and to the point.
Two issues:

 * "In this patch," is redundant.

 * "it is brief and to the point" are equally applicable to both
   styles, so that is not a *reason* to choose one over the other.

If a condition were *not* brief and to the point, then a rewrite to
the latter style will make the resulting code worse:

        if (a very complex condition
            that potentially have to consume a
            lot of brain-cycles to understand) {
                has_that_condition = 1;
        }

is a lot easier to extend than

        has_that_condition = (a very complex condition
                              that potentially have to consume a
                              lot of brain-cycles to understand);

because it is a lot more likely that we would need to later extend
such a complex condition is more likely than a simple singleton
condition, and we could end up with

        if (a very complex condition
            that potentially have to consume a
            lot of brain-cycles to understand) {
                futher computation to check if
                the condition really holds
                will be added here later
                if (does that condition really hold true?)
                        has_that_condition = 1;
        }


which may be harder to express in the latter form.

In other words, "it is brief and to the point" merely _allows_ these
statements to be expressed in the latter form; it does not say
anything about which is better between the former and the latter.
quoted
Signed-off-by: Hiroyuki Sano <redacted>
---
 fsck.c | 18 ++++++------------
 1 file changed, 6 insertions(+), 12 deletions(-)
diff --git a/fsck.c b/fsck.c
index b3022ad..abed62b 100644
--- a/fsck.c
+++ b/fsck.c
@@ -165,18 +165,12 @@ static int fsck_tree(struct tree *item, int strict, fsck_error error_func)

              sha1 = tree_entry_extract(&desc, &name, &mode);

-             if (is_null_sha1(sha1))
-                     has_null_sha1 = 1;
-             if (strchr(name, '/'))
-                     has_full_path = 1;
-             if (!*name)
-                     has_empty_name = 1;
-             if (!strcmp(name, "."))
-                     has_dot = 1;
-             if (!strcmp(name, ".."))
-                     has_dotdot = 1;
-             if (!strcmp(name, ".git"))
-                     has_dotgit = 1;
+             has_null_sha1 |= is_null_sha1(sha1);
+             has_full_path |= !!strchr(name, '/');
+             has_empty_name |= !*name;
+             has_dot |= !strcmp(name, ".");
+             has_dotdot |= !strcmp(name, "..");
+             has_dotgit |= !strcmp(name, ".git");
              has_zero_pad |= *(char *)desc.buffer == '0';
              update_tree_entry(&desc);


-- 
Hiroyuki Sano
sh19910711@gmail.com
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help