[jira] Commented: (CLI-151) HelpFormatter wraps incorrectly on every line beyond the first

["J. Lewis Muir (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CLI-151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12590028#action_12590028 ] 

J. Lewis Muir commented on CLI-151:
-----------------------------------

I have observed this same problem.

I'm not sure the suggested fix to {{findWrapPos}} is correct.  When I tried it, the {{testFindWrapPos}} test case failed with

{noformat}
[junit] wrap position 2 expected:<-1> but was:<16>
[junit] junit.framework.AssertionFailedError:
[junit]     wrap position 2 expected:<-1> but was:<16>
[junit]   at org.apache.commons.cli.HelpFormatterTest.testFindWrapPos(
[junit]     HelpFormatterTest.java:62)
{noformat}

I think the problem is actually in {{HelpFormatter.renderWrappedText}} ({{HelpFormatter.java}} line 812).  It calls {{findWrapPos}} with a start position of {{nextLineTabStop}}, but that's not right; it should be a start position of 0.

The line wrapping works for me after making this correction.  I will attach a patch containing this fix as well as an addition to the test cases to validate it.

> HelpFormatter wraps incorrectly on every line beyond the first
> --------------------------------------------------------------
>
>                 Key: CLI-151
>                 URL: https://issues.apache.org/jira/browse/CLI-151
>             Project: Commons CLI
>          Issue Type: Bug
>          Components: CLI-1.x
>            Reporter: Dan Armbrust
>             Fix For: 1.2
>
>
> The method findWrapPos(...) in the HelpFormatter is a couple of bugs in the way that it deals with the "startPos" variable.  This causes it to format every line beyond the first line by "startPos" to many characters, beyond the specified width.  
> To see this, create an option with a long description, and then use the help formatter to print it.  The first line will be the correct length.  The 2nd, 3rd, etc lines will all be too long.
> I don't have a patch (sorry) - but here is a corrected version of the method.
> I fixed it in two places - both were using "width + startPos" when they should have been using width.
> {code}
>  protected int findWrapPos(String text, int width, int startPos)
>     {
>         int pos = -1;
>         // the line ends before the max wrap pos or a new line char found
>         if (((pos = text.indexOf('\n', startPos)) != -1 && pos <= width)
>             || ((pos = text.indexOf('\t', startPos)) != -1 && pos <= width))
>         {
>             return pos+1;
>         }
>         else if ((width) >= text.length())
>         {
>             return -1;
>         }
>         // look for the last whitespace character before startPos+width
>         pos = width;
>         char c;
>         while ((pos >= startPos) && ((c = text.charAt(pos)) != ' ')
>                && (c != '\n') && (c != '\r'))
>         {
>             --pos;
>         }
>         // if we found it - just return
>         if (pos > startPos)
>         {
>             return pos;
>         }
>         
>         // must look for the first whitespace chearacter after startPos 
>         // + width
>         pos = startPos + width;
>         while ((pos <= text.length()) && ((c = text.charAt(pos)) != ' ')
>                && (c != '\n') && (c != '\r'))
>         {
>             ++pos;
>         }
>         return (pos == text.length())        ? (-1) : pos;
>     }
> {code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.