You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@shindig.apache.org by Dan Bentley <db...@google.com> on 2008/03/25 08:23:07 UTC
Running (slowly) Review Board
Hey shindig-dev,
I mentioned trying out review-board. I now have an instance set up
against shindig's respository. I've put in one revision that you can look
at before submitting anything.
http://metzby.dyndns.info:8000/r/1/
Caveats:
1) This is running on my iMac desktop. As such, DO NOT TRUST THIS COMPUTER
WITH CONFIDENTIAL INFORMATION OR PASSWORDS YOU WANT ME NOT TO KNOW. I'm not
going to be malicious, but why should you trust me?
2) It is (currently) using django's built-in server, which is slow. I'm
investigating setting it up behind Apache so it can run port 80 nicely.
3) It is running on my desktop; it may die. Ping me on email and I will
kick the server if necessary.
4) It is using sqlite instead of mysql, which is slow.
5) It is not using memcached, which is slow.
Check out http://metzby.dyndns.info:8000/r/1/diff/#index_header to see what
the online-review looks like.
Please let me know if you find this useful, useless,
would-be-useful-if-<suggestion>, or if you have a server where I could
productionize this.
Thanks for comments, and apologies if this has melted down by the time you
try to use it
-Dan
Re: Running (slowly) Review Board
Posted by Santiago Gala <sa...@gmail.com>.
El mar, 25-03-2008 a las 01:12 -0700, Kevin Brown escribió:
> That's pretty close to what I had in mind (although the interface is a
> little weird).
>
I was not able to put guice.txt there, but I managed in my git backed
instance. Not easily, as the tool (reviewboard) was unable to apply a
working patch both with dan's subversion backend and with my git
backend, until I turned the patch into a git one (here).
In any case, now it works here: http://marlow.memojo.com/rb/r/6/
(it is my laptop, so it depends if I went off with it or I'm connected)
and I attach the fairly similar output of
git diff --stat=110,90 -p -M -C -C --color --color-words
master^..test-patch
I guess anybody with a terminal with decent ansi color support can see
it. In particular, I find very interesting how well git -M and -C deal
with code movements. See for instance GadgetRenderer.java:
GadgetRenderer.java | 411 +---------
{GadgetRenderer.java => GadgetRenderingTask.java} | 47 +-
It has been mostly deleted, except parts have been copied to the
RenderingTast. It works really well in all examples I tried except en
cases where a template is repeatedly copied (.po in translations, for
instance), where the copy information is distracting.
> On Tue, Mar 25, 2008 at 12:23 AM, Dan Bentley <db...@google.com> wrote:
>
> > Hey shindig-dev,
> > I mentioned trying out review-board. I now have an instance set up
> > against shindig's respository. I've put in one revision that you can look
> > at before submitting anything.
> >
> > http://metzby.dyndns.info:8000/r/1/
> >
> > Caveats:
> > 1) This is running on my iMac desktop. As such, DO NOT TRUST THIS
> > COMPUTER
> > WITH CONFIDENTIAL INFORMATION OR PASSWORDS YOU WANT ME NOT TO KNOW. I'm
> > not
> > going to be malicious, but why should you trust me?
> > 2) It is (currently) using django's built-in server, which is slow. I'm
> > investigating setting it up behind Apache so it can run port 80 nicely.
> > 3) It is running on my desktop; it may die. Ping me on email and I will
> > kick the server if necessary.
> > 4) It is using sqlite instead of mysql, which is slow.
> > 5) It is not using memcached, which is slow.
> >
> >
> > Check out http://metzby.dyndns.info:8000/r/1/diff/#index_header to see
> > what
> > the online-review looks like.
> >
> > Please let me know if you find this useful, useless,
> > would-be-useful-if-<suggestion>, or if you have a server where I could
> > productionize this.
> >
> > Thanks for comments, and apologies if this has melted down by the time you
> > try to use it
> > -Dan
> >
>
>
>
--
Santiago Gala
http://memojo.com/~sgala/blog/
Re: Running (slowly) Review Board
Posted by Kevin Brown <et...@google.com>.
That's pretty close to what I had in mind (although the interface is a
little weird).
On Tue, Mar 25, 2008 at 12:23 AM, Dan Bentley <db...@google.com> wrote:
> Hey shindig-dev,
> I mentioned trying out review-board. I now have an instance set up
> against shindig's respository. I've put in one revision that you can look
> at before submitting anything.
>
> http://metzby.dyndns.info:8000/r/1/
>
> Caveats:
> 1) This is running on my iMac desktop. As such, DO NOT TRUST THIS
> COMPUTER
> WITH CONFIDENTIAL INFORMATION OR PASSWORDS YOU WANT ME NOT TO KNOW. I'm
> not
> going to be malicious, but why should you trust me?
> 2) It is (currently) using django's built-in server, which is slow. I'm
> investigating setting it up behind Apache so it can run port 80 nicely.
> 3) It is running on my desktop; it may die. Ping me on email and I will
> kick the server if necessary.
> 4) It is using sqlite instead of mysql, which is slow.
> 5) It is not using memcached, which is slow.
>
>
> Check out http://metzby.dyndns.info:8000/r/1/diff/#index_header to see
> what
> the online-review looks like.
>
> Please let me know if you find this useful, useless,
> would-be-useful-if-<suggestion>, or if you have a server where I could
> productionize this.
>
> Thanks for comments, and apologies if this has melted down by the time you
> try to use it
> -Dan
>
--
~Kevin
Re: Running (slowly) Review Board
Posted by Santiago Gala <sa...@gmail.com>.
El mié, 26-03-2008 a las 16:54 -0700, Kevin Brown escribió:
> So, I can't for the life of me figure out how to upload a patch to this
> thing. What format is it expecting?
>
I got a couple in.
I think each backend expects the corresponding patch format (silly but
true). Evidence:
- the git backend can't take your guice patch but takes a "git diff"
after I "path -p0 <..." it onto a git checkout.
- the subversion repository can't take this one, but takes a svn
generated one
Also, when the subversion backend is used, a "." must be specified in
the path. And it will do at least one round trip to the server per file
*two* times in the process (at least). So you need to be very patient
when uploading patches there.
My laptop contains it as:
http://marlow.memojo.com/rb/r/6/ (third version of the patch; second is
an error, first is the first one not too well imported) with some silly
comments. But it will be offline substantial portions of the day,
starting in a few minutes.
Also I managed to upload the two guice patches to the git repo and push
them to github:
http://github.com/sgala/apache-incubator-shindig/commits/guice-test-2
http://github.com/sgala/apache-incubator-shindig/commits/guice-test-1
but their UI has no facility for diff'ing two branches. I can view
locally the difference between both patches, which is handy.
Just FYI I attach the output of "git diff --stat -p -M -C -C master
guice-test-2 >/tmp/guice-t2.patch" and "git diff --stat -p -M -C -C
guice-test-1 guice-test-2 >/tmp/guice-t2-t1.patch" i.e. the git'ized
second version of the patch and the difference between the first version
of the guice patch and the second one, on SHINDIG-152. The previous one
I attached had the problem that applying the subversion patch on top of
git didn't automatically "rm" the deleted files, which introduced some
artifacts. I only noticed when comparing both patches.
I found the diff between patches useful to know what really changed
between both.
Regards
Santiago
> On Tue, Mar 25, 2008 at 12:23 AM, Dan Bentley <db...@google.com> wrote:
>
> > Hey shindig-dev,
> > I mentioned trying out review-board. I now have an instance set up
> > against shindig's respository. I've put in one revision that you can look
> > at before submitting anything.
> >
> > http://metzby.dyndns.info:8000/r/1/
> >
> > Caveats:
> > 1) This is running on my iMac desktop. As such, DO NOT TRUST THIS
> > COMPUTER
> > WITH CONFIDENTIAL INFORMATION OR PASSWORDS YOU WANT ME NOT TO KNOW. I'm
> > not
> > going to be malicious, but why should you trust me?
> > 2) It is (currently) using django's built-in server, which is slow. I'm
> > investigating setting it up behind Apache so it can run port 80 nicely.
> > 3) It is running on my desktop; it may die. Ping me on email and I will
> > kick the server if necessary.
> > 4) It is using sqlite instead of mysql, which is slow.
> > 5) It is not using memcached, which is slow.
> >
> >
> > Check out http://metzby.dyndns.info:8000/r/1/diff/#index_header to see
> > what
> > the online-review looks like.
> >
> > Please let me know if you find this useful, useless,
> > would-be-useful-if-<suggestion>, or if you have a server where I could
> > productionize this.
> >
> > Thanks for comments, and apologies if this has melted down by the time you
> > try to use it
> > -Dan
> >
>
>
>
--
Santiago Gala
http://memojo.com/~sgala/blog/
Re: Running (slowly) Review Board
Posted by Kevin Brown <et...@google.com>.
So, I can't for the life of me figure out how to upload a patch to this
thing. What format is it expecting?
On Tue, Mar 25, 2008 at 12:23 AM, Dan Bentley <db...@google.com> wrote:
> Hey shindig-dev,
> I mentioned trying out review-board. I now have an instance set up
> against shindig's respository. I've put in one revision that you can look
> at before submitting anything.
>
> http://metzby.dyndns.info:8000/r/1/
>
> Caveats:
> 1) This is running on my iMac desktop. As such, DO NOT TRUST THIS
> COMPUTER
> WITH CONFIDENTIAL INFORMATION OR PASSWORDS YOU WANT ME NOT TO KNOW. I'm
> not
> going to be malicious, but why should you trust me?
> 2) It is (currently) using django's built-in server, which is slow. I'm
> investigating setting it up behind Apache so it can run port 80 nicely.
> 3) It is running on my desktop; it may die. Ping me on email and I will
> kick the server if necessary.
> 4) It is using sqlite instead of mysql, which is slow.
> 5) It is not using memcached, which is slow.
>
>
> Check out http://metzby.dyndns.info:8000/r/1/diff/#index_header to see
> what
> the online-review looks like.
>
> Please let me know if you find this useful, useless,
> would-be-useful-if-<suggestion>, or if you have a server where I could
> productionize this.
>
> Thanks for comments, and apologies if this has melted down by the time you
> try to use it
> -Dan
>
--
~Kevin