[
https://jira.codehaus.org/browse/SCM-735?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=349767#comment-349767
]
Todd Currie commented on SCM-735:
---------------------------------
This patch should be adjusted to take into account SCM-745. The regular
expression should not make use of the greedy wild card match, otherwise the
author name can become corrupted depending on the text of the change
description.
>From the attached patch:
@@ -36,7 +36,7 @@
public class PerforceFilelogConsumer {
private static final String PERFORCE_TIMESTAMP_PATTERN = "yyyy/MM/dd";
- private static final String LINE_PATTERN = "#(\\d+).*on (.*) by (.*)@";
+ private static final String LINE_PATTERN = "#(\\d+) change (\\d+) .* on
(.*) by (.*)@";
Should be:
@@ -36,7 +36,7 @@
public class PerforceFilelogConsumer {
private static final String PERFORCE_TIMESTAMP_PATTERN = "yyyy/MM/dd";
- private static final String LINE_PATTERN = "#(\\d+).*on (.*) by (.*)@";
+ private static final String LINE_PATTERN = "#(\\d+) change (\\d+) .*? on
(.*?) by (.*?)@";
> Blame result on branched file is incorrect
> ------------------------------------------
>
> Key: SCM-735
> URL: https://jira.codehaus.org/browse/SCM-735
> Project: Maven SCM
> Issue Type: Bug
> Components: maven-scm-provider-perforce
> Affects Versions: 1.8.1
> Reporter: Gregory SSI-YAN-KAI
> Attachments: scm.patch
>
>
> When a file is branched, its revision id restarts from 1.
> So, 2 different lines can have the same revision id even though they have
> been submitted in 2 different changelists.
> Since the perforce blame command is based on the revision id of each line,
> the result might be incorrect.
> The solution I'm proposing in the attached patch is to use changelist number
> to identify the author of a line.
--
This message was sent by Atlassian JIRA
(v6.1.6#6162)
