webrev failure?

Martijn Verburg martijnverburg at gmail.com
Tue Aug 14 17:12:41 UTC 2012


Hi Chris,

> Martin,
>
> Unless you have changes that span multiple repositories, then it can be
> quite straight forward to run webrev ( without '-f ) on just the individual
> repository that you are changing. It will run with simply 'hg out'.

That did seem to work, thanks.

> If you have changes in multiple repo's then you have other choices:
>  1) run webrev on each individually, and include all the links in the
>     review mail
>  2) create a 'file.list' ( a file containing your changed files, one
>     file name per line ), and pass it as the final arg to webrev
>  3) Run with a really really old hg ( I use 0.9.5 ) and an installed
>     forest extension. Now, 'webrev -f' should work just fine ;-)

We could go with 1) or 2) - I'll work that into our process.

Cheers,
Martijn

> On 11/08/12 17:04, Martijn Verburg wrote:
>>
>> Hi all,
>>
>> Apologies if these are the wrong mailing lists, but I'm assuming that
>> the webrev tool falls under this domain somewhat.
>>
>> In order to get patches into the OpenJDK as cleanly as possible, we're
>> looking to utilise webrev (since it's the std and all).
>>
>> Running "ksh ./make/scripts/webrev.ksh -f" over the latest jdk8/tl
>> source (no patches, but a full build was completed using build-infra)
>> and I got a host of errors:
>>
>> --------
>>
>>     SCM detected: mercurial
>> hg: unknown command 'foutgoing'
>> abort: cannot follow file not in parent revision: "Mercurial"
>> abort: cannot follow file not in parent revision: "basic"
>> abort: cannot follow file not in parent revision: "add"
>> abort: cannot follow file not in parent revision: "annotate"
>> ..
>> ..
>>
>> --------
>>
>> A separate, yet related set of errors are at
>> http://pastebin.com/q0tF1A4m for sake of brevity in this mail.
>>
>> I assume I'm running this tool incorrectly, I expected a result of
>> something like "You've changed nothing, nothing to se here move along
>> please" :-)
>>
>> Cheers,
>> Martijn
>>
>



More information about the build-dev mailing list