RFR: JDK-8029512 : webrev.ksh: fix parsing of parent repo paths from hgrc file

[Mike Duigou]

On Jan 7 2014, at 03:46 , Dave Pointon <dpointo8 at linux.vnet.ibm.com> wrote:

> Hi Mike ,
> 
> On Mon, 2014-01-06 at 19:45 -0800, Mike Duigou wrote:
>> I have updated my previous webrev to make the regular expressions simpler and more portable based on feedback. I have tested on Mac, Solaris, Linux and Cygwin.
>> 
>> http://cr.openjdk.java.net/~mduigou/JDK-8029512/1/webrev/
>> 
>> I did also note that the hgrc format allows for definitions of the form:
>> 
>> foo =
>>  bar
>> 
>> where any line beginning with white space is considered a continuation line. This is not easy to support with sed and I have not attempted to solve it in this patch. (I did document the limitation).
>> 
>> To be honest the best approach may have been to write the script in Python and use the ConfigParser package but I suspect that's what Mercurial does so this would re-expose us to the original bug which apparently necessitated manual parsing.
>> 
>> Cheers,
>> 
>> Mike
>> 
> <snip>
> 
> It looks a little over-complicated than is needed, but that aside, there
> appear to be 2 identical typos - on lines 1949 and 1953 - whereby the
> 'OR' operator ('||') appears instead of the pipe operator ('|').

Thanks for the correction. My tests didn't require the filter stage so I failed to notice that this typo had crept in.

> 
> I would suggest the use of perl for JDK 9 since all whitespace
> characters are matched &/or captured by the '\s' regex operator.

The hope is that webrev will be retired within the next year. The current efforts are intended only as a stopgap until the "glorious future" arrives.

> Best rgds ,
> 
> -- 
> Dave Pointon FIAP MBCS
> 
> Now I saw, tho' too late, the folly of beginning a work before we count
> the cost and before we we judge rightly of our strength to go thro'
> with it - Robinson Crusoe
>