git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v3] commit.c: use skip_prefix() instead of starts_with()
@ 2014-03-04  8:42 Tanay Abhra
  2014-03-04 19:16 ` Max Horn
  2014-03-04 19:33 ` Junio C Hamano
  0 siblings, 2 replies; 4+ messages in thread
From: Tanay Abhra @ 2014-03-04  8:42 UTC (permalink / raw)
  To: git; +Cc: mhagger, gitster, max, Tanay Abhra

In record_author_date() & parse_gpg_output() ,using skip_prefix() instead of
starts_with() is a more suitable abstraction.

Helped-by: Max Horn <max@quendi.de> 
Helped-by: Junio C Hamano <gitster@pobox.com>
Helped-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Tanay Abhra <tanayabh@gmail.com>
---
Patch V3 Variable naming improved, removed assignments inside conditionals.
	Thanks to Junio C Hamano and Max Horn.
	
Patch V2 Corrected email formatting ,reapplied the implementation according to suggestions.
	Thanks to Michael Haggerty.

This is in respect to GSoC microproject #10.

In record_author_date(), extra and useless calls to strlen due to using starts_with()
were removed by using skip_prefix(). Extra variable "skip" was used as "buf" is used in 
for loop update check.

Other usages of starts_with() in the same file can be found with,

$ grep -n starts_with commit.c

1116:		else if (starts_with(line, gpg_sig_header) &&
1196:		if (starts_with(buf, sigcheck_gpg_status[i].check + 1)) {

The starts_with() in line 1116 was left as it is, as strlen values were pre generated as 
global variables.
The starts_with() in line 1196 was replaced as it abstracts way the skip_prefix part by
directly using the function.
Also skip_prefix() is inline when compared to starts_with().

 commit.c | 17 +++++++++--------
 1 file changed, 9 insertions(+), 8 deletions(-)

diff --git a/commit.c b/commit.c
index 6bf4fe0..6c92acb 100644
--- a/commit.c
+++ b/commit.c
@@ -548,7 +548,7 @@ define_commit_slab(author_date_slab, unsigned long);
 static void record_author_date(struct author_date_slab *author_date,
 			       struct commit *commit)
 {
-	const char *buf, *line_end;
+	const char *buf, *line_end, *ident_line;
 	char *buffer = NULL;
 	struct ident_split ident;
 	char *date_end;
@@ -566,14 +566,16 @@ static void record_author_date(struct author_date_slab *author_date,
 	     buf;
 	     buf = line_end + 1) {
 		line_end = strchrnul(buf, '\n');
-		if (!starts_with(buf, "author ")) {
+		ident_line = skip_prefix(buf, "author ");
+		if (!ident_line) {
 			if (!line_end[0] || line_end[1] == '\n')
 				return; /* end of header */
 			continue;
 		}
+		buf = ident_line;
 		if (split_ident_line(&ident,
-				     buf + strlen("author "),
-				     line_end - (buf + strlen("author "))) ||
+				     buf,
+				     line_end - buf) ||
 		    !ident.date_begin || !ident.date_end)
 			goto fail_exit; /* malformed "author" line */
 		break;
@@ -1193,10 +1195,9 @@ static void parse_gpg_output(struct signature_check *sigc)
 	for (i = 0; i < ARRAY_SIZE(sigcheck_gpg_status); i++) {
 		const char *found, *next;
 
-		if (starts_with(buf, sigcheck_gpg_status[i].check + 1)) {
-			/* At the very beginning of the buffer */
-			found = buf + strlen(sigcheck_gpg_status[i].check + 1);
-		} else {
+		found = skip_prefix(buf, sigcheck_gpg_status[i].check + 1);
+		/* At the very beginning of the buffer */
+		if(!found) {
 			found = strstr(buf, sigcheck_gpg_status[i].check);
 			if (!found)
 				continue;
-- 
1.9.0

^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH v3] commit.c: use skip_prefix() instead of starts_with()
  2014-03-04  8:42 [PATCH v3] commit.c: use skip_prefix() instead of starts_with() Tanay Abhra
@ 2014-03-04 19:16 ` Max Horn
  2014-03-04 20:38   ` Junio C Hamano
  2014-03-04 19:33 ` Junio C Hamano
  1 sibling, 1 reply; 4+ messages in thread
From: Max Horn @ 2014-03-04 19:16 UTC (permalink / raw)
  To: Tanay Abhra; +Cc: git, mhagger, gitster

[-- Attachment #1: Type: text/plain, Size: 1948 bytes --]


On 04.03.2014, at 09:42, Tanay Abhra <tanayabh@gmail.com> wrote:

[...]

> commit.c | 17 +++++++++--------
> 1 file changed, 9 insertions(+), 8 deletions(-)
> 
> diff --git a/commit.c b/commit.c
> index 6bf4fe0..6c92acb 100644
> --- a/commit.c
> +++ b/commit.c

[...]

> @@ -566,14 +566,16 @@ static void record_author_date(struct author_date_slab *author_date,
> 	     buf;
> 	     buf = line_end + 1) {
> 		line_end = strchrnul(buf, '\n');
> -		if (!starts_with(buf, "author ")) {
> +		ident_line = skip_prefix(buf, "author ");
> +		if (!ident_line) {
> 			if (!line_end[0] || line_end[1] == '\n')
> 				return; /* end of header */
> 			continue;
> 		}
> +		buf = ident_line;
> 		if (split_ident_line(&ident,
> -				     buf + strlen("author "),
> -				     line_end - (buf + strlen("author "))) ||
> +				     buf,
> +				     line_end - buf) ||
> 		    !ident.date_begin || !ident.date_end)
> 			goto fail_exit; /* malformed "author" line */
> 		break;

Why not get rid of that assignment to "buf", and use ident_line instead of buf below? That seems like it would be more readable, wouldn't it?


> @@ -1193,10 +1195,9 @@ static void parse_gpg_output(struct signature_check *sigc)
> 	for (i = 0; i < ARRAY_SIZE(sigcheck_gpg_status); i++) {
> 		const char *found, *next;
> 
> -		if (starts_with(buf, sigcheck_gpg_status[i].check + 1)) {
> -			/* At the very beginning of the buffer */
> -			found = buf + strlen(sigcheck_gpg_status[i].check + 1);
> -		} else {
> +		found = skip_prefix(buf, sigcheck_gpg_status[i].check + 1);
> +		/* At the very beginning of the buffer */

Do we really need that comment, and in that spot? The code seemed clear enough to me without it. But if you think keeping is better, perhaps move it to *before* the skip_prefix, and add a trailing "?"

> +		if(!found) {
> 			found = strstr(buf, sigcheck_gpg_status[i].check);
> 			if (!found)
> 				continue;



[-- Attachment #2: Message signed with OpenPGP using GPGMail --]
[-- Type: application/pgp-signature, Size: 235 bytes --]

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH v3] commit.c: use skip_prefix() instead of starts_with()
  2014-03-04  8:42 [PATCH v3] commit.c: use skip_prefix() instead of starts_with() Tanay Abhra
  2014-03-04 19:16 ` Max Horn
@ 2014-03-04 19:33 ` Junio C Hamano
  1 sibling, 0 replies; 4+ messages in thread
From: Junio C Hamano @ 2014-03-04 19:33 UTC (permalink / raw)
  To: Tanay Abhra; +Cc: git, mhagger, max

Tanay Abhra <tanayabh@gmail.com> writes:

> In record_author_date() & parse_gpg_output() ,using skip_prefix() instead of
> starts_with() is a more suitable abstraction.

Thanks.  Will queue with a reworded message to clarify what exactly
"A more suitable" means.

Here is what I tentatively came up with.

-- >8 --
From: Tanay Abhra <tanayabh@gmail.com>
Date: Tue, 4 Mar 2014 00:42:20 -0800
Subject: [PATCH] commit.c: use skip_prefix() instead of starts_with()

In record_author_date() & parse_gpg_output(), the callers of
starts_with() not just want to know if the string starts with the
prefix, but also can benefit from knowing the string that follows
the prefix.

By using skip_prefix(), we can do both at the same time.

Helped-by: Max Horn <max@quendi.de>
Helped-by: Junio C Hamano <gitster@pobox.com>
Helped-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Tanay Abhra <tanayabh@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
---
 commit.c | 17 +++++++++--------
 1 file changed, 9 insertions(+), 8 deletions(-)

diff --git a/commit.c b/commit.c
index 6bf4fe0..6c92acb 100644
--- a/commit.c
+++ b/commit.c
@@ -548,7 +548,7 @@ define_commit_slab(author_date_slab, unsigned long);
 static void record_author_date(struct author_date_slab *author_date,
 			       struct commit *commit)
 {
-	const char *buf, *line_end;
+	const char *buf, *line_end, *ident_line;
 	char *buffer = NULL;
 	struct ident_split ident;
 	char *date_end;
@@ -566,14 +566,16 @@ static void record_author_date(struct author_date_slab *author_date,
 	     buf;
 	     buf = line_end + 1) {
 		line_end = strchrnul(buf, '\n');
-		if (!starts_with(buf, "author ")) {
+		ident_line = skip_prefix(buf, "author ");
+		if (!ident_line) {
 			if (!line_end[0] || line_end[1] == '\n')
 				return; /* end of header */
 			continue;
 		}
+		buf = ident_line;
 		if (split_ident_line(&ident,
-				     buf + strlen("author "),
-				     line_end - (buf + strlen("author "))) ||
+				     buf,
+				     line_end - buf) ||
 		    !ident.date_begin || !ident.date_end)
 			goto fail_exit; /* malformed "author" line */
 		break;
@@ -1193,10 +1195,9 @@ static void parse_gpg_output(struct signature_check *sigc)
 	for (i = 0; i < ARRAY_SIZE(sigcheck_gpg_status); i++) {
 		const char *found, *next;
 
-		if (starts_with(buf, sigcheck_gpg_status[i].check + 1)) {
-			/* At the very beginning of the buffer */
-			found = buf + strlen(sigcheck_gpg_status[i].check + 1);
-		} else {
+		found = skip_prefix(buf, sigcheck_gpg_status[i].check + 1);
+		/* At the very beginning of the buffer */
+		if(!found) {
 			found = strstr(buf, sigcheck_gpg_status[i].check);
 			if (!found)
 				continue;
-- 
1.9.0-186-gd464cb7

^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH v3] commit.c: use skip_prefix() instead of starts_with()
  2014-03-04 19:16 ` Max Horn
@ 2014-03-04 20:38   ` Junio C Hamano
  0 siblings, 0 replies; 4+ messages in thread
From: Junio C Hamano @ 2014-03-04 20:38 UTC (permalink / raw)
  To: Max Horn; +Cc: Tanay Abhra, git, mhagger

Max Horn <max@quendi.de> writes:

>> +		buf = ident_line;
>> 		if (split_ident_line(&ident,
>> -				     buf + strlen("author "),
>> -				     line_end - (buf + strlen("author "))) ||
>> +				     buf,
>> +				     line_end - buf) ||
>> 		    !ident.date_begin || !ident.date_end)
>> 			goto fail_exit; /* malformed "author" line */
>> 		break;
>
> Why not get rid of that assignment to "buf", and use ident_line
> instead of buf below? That seems like it would be more readable,
> wouldn't it?

Yes, and also now the argument list is much shorter, you could
probably do it on two lines instead of three:

                if (split_ident_line(&ident,
                                     ident_line, line_end - ident_line) ||
                    ...


>> @@ -1193,10 +1195,9 @@ static void parse_gpg_output(struct signature_check *sigc)
>> 	for (i = 0; i < ARRAY_SIZE(sigcheck_gpg_status); i++) {
>> 		const char *found, *next;
>> 
>> -		if (starts_with(buf, sigcheck_gpg_status[i].check + 1)) {
>> -			/* At the very beginning of the buffer */
>> -			found = buf + strlen(sigcheck_gpg_status[i].check + 1);
>> -		} else {
>> +		found = skip_prefix(buf, sigcheck_gpg_status[i].check + 1);
>> +		/* At the very beginning of the buffer */
>
> Do we really need that comment, and in that spot? The code seemed
> clear enough to me without it. But if you think keeping is better,
> perhaps move it to *before* the skip_prefix, and add a trailing
> "?"

Both good suggestions (I tend to prefer the removal).

Thanks.

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2014-03-04 20:38 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-03-04  8:42 [PATCH v3] commit.c: use skip_prefix() instead of starts_with() Tanay Abhra
2014-03-04 19:16 ` Max Horn
2014-03-04 20:38   ` Junio C Hamano
2014-03-04 19:33 ` Junio C Hamano

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).