threads / patch / 17130

patchgitweb: recognize six digit abbreviated SHA1

Subject: [PATCH] gitweb: recognize six digit abbreviated SHA1

## tl;dr

5 messages between Jan 13, 2009 and Jan 14, 2009. Diffs are folded; open one to read it.

replies: 4people: 3as markdown or json

Anders Melchiorsen· Jan 13, 2009, 00:04 UTC · lore

This lets gitweb hightlight abbreviated hashes as produced by git rev-parse --short.

Signed-off-by: Anders Melchiorsen <mail@cup.kalibalik.dk>
---

It seems like seven digit hashes are in vogue now. So, did I miss some reason for keeping it at eight in this spot?

Show changes to gitweb/gitweb.perl +1 −1

diff --git a/gitweb/gitweb.perl b/gitweb/gitweb.perl
index 0ac84d1..1a7d448 100755
--- a/gitweb/gitweb.perl
+++ b/gitweb/gitweb.perl
@@ -1337,7 +1337,7 @@ sub format_log_line_html {
 	my $line = shift;
 
 	$line = esc_html($line, -nbsp=>1);
-	if ($line =~ m/([0-9a-fA-F]{8,40})/) {
+	if ($line =~ m/([0-9a-fA-F]{7,40})/) {
 		my $hash_text = $1;
 		my $link =
 			$cgi->a({-href => href(action=>"object", hash=>$hash_text),

-- 
1.6.0.2.514.g23abd3

Anders Melchiorsen· Jan 13, 2009, 00:08 UTC · re: Anders Melchiorsen · lore

Re: [PATCH] gitweb: recognize six digit abbreviated SHA1

Anders Melchiorsen <mail@cup.kalibalik.dk> writes:

> +	if ($line =~ m/([0-9a-fA-F]{7,40})/) {

I could not make up my mind between the seven digits from "git rev-parse --short" and the six digits currently used by gitk.

So I put one option in the patch, and the other one in the subject.

Cheers, Anders.

Sam Vilain· Jan 13, 2009, 02:00 UTC · re: Anders Melchiorsen · lore

Re: [PATCH] gitweb: recognize six digit abbreviated SHA1

Anders Melchiorsen wrote:

Show 11 quoted lines

> Anders Melchiorsen <mail@cup.kalibalik.dk> writes:
>
>   
>> +	if ($line =~ m/([0-9a-fA-F]{7,40})/) {
>>     
>
> I could not make up my mind between the seven digits from "git
> rev-parse --short" and the six digits currently used by gitk.
>
> So I put one option in the patch, and the other one in the subject.
>

I think if you're going to go so short as 6 digits, it's probably worth making sure that the really short SHA1s check commits only. eg, if you've got a commit 'fa023473' and a tree 'fa023421', then 'fa0234' should match the commit and not the tree. But I don't think there's a plumbing way to do a query like that at the moment.

Even git.git isn't unique over 6 digits, even restricting to commits. Not since b0a3de ;-).

Sam

Junio C Hamano· Jan 13, 2009, 08:26 UTC · re: Sam Vilain · lore

Re: [PATCH] gitweb: recognize six digit abbreviated SHA1

Sam Vilain <sam@vilain.net> writes:

Show 5 quoted lines

> I think if you're going to go so short as 6 digits, it's probably worth
> making sure that the really short SHA1s check commits only. eg, if
> you've got a commit 'fa023473' and a tree 'fa023421', then 'fa0234'
> should match the commit and not the tree. But I don't think there's a
> plumbing way to do a query like that at the moment.

When people give an abbreviated object name, 99% of the time they mean commits (and "index deadbeef..acebead" in a patch is a good place to pick blob object names from, which would be what the 99% of the remaining 1% would name), so making sure it is a commit or a blob would be a very sensible thing to do.

Unfortunately, you fundamentally cannot do this without taking a performance hit of actually opening the object.

Sam Vilain· Jan 14, 2009, 00:33 UTC · re: Junio C Hamano · lore

Re: [PATCH] gitweb: recognize six digit abbreviated SHA1

Junio C Hamano wrote:

Show 9 quoted lines

> When people give an abbreviated object name, 99% of the time they mean
> commits (and "index deadbeef..acebead" in a patch is a good place to pick
> blob object names from, which would be what the 99% of the remaining 1%
> would name), so making sure it is a commit or a blob would be a very
> sensible thing to do.
>
> Unfortunately, you fundamentally cannot do this without taking a
> performance hit of actually opening the object.
>

Yeah. I envisioned this just for those cases where the tool would otherwise return 'ambiguous argument'; if there was a hint as to the object type, open *all* the matching objects and return the one that matched the type only.

Sam.