Re: Active Directory Authentication with TLS
On Tue, Jan 17, 2012 at 21:09, Brad Hards wrote: > On Wednesday 18 January 2012 02:08:50 Thilo-Alexander Ginkel wrote: >> When disabling TLS, everything works like expected. > Are you just enabling / disabling TLS? Yes. >> Any ideas? Do I need to register the AD DC's CA certificate as trusted >> root somewhere? > Check you're using the right port for your AD configuration (perhaps 636 or > 3269 depending on whether you're trying to talk to the GC or not). According to tcpdump Review Board uses port 389 (ldap) to connect to the AD if TLS is enabled. I verified using another LDAP client that the AD server supports StartTLS on port 389 and ldaps on port 636. AFAICS, there is no way to change the port from the RB admin interface. Is there anything obvious that I am missing? Thanks, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Active Directory Authentication with TLS
Hi there, we are currently trying to switch Review Board from the Standard Registration mechanism to Active Directory Authentication. Unfortunately, this only seems to work as long as TLS is not enabled. The Domain Controller seems to support TLS and according to tcpdump I can see some data being exchanged that could be a TLS handshake. Unfortunately, the authentication fails after hitting a timeout while the log output is not too helpful: -- 8< -- 2012-01-17 15:44:23,662 - WARNING - Active Directory: Failed login for user xyz -- 8< -- When disabling TLS, everything works like expected. Any ideas? Do I need to register the AD DC's CA certificate as trusted root somewhere? Thanks, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Handling of files attached to a review request deleted at a later revision in Subversion
On Tue, Jan 10, 2012 at 11:22, Thilo-Alexander Ginkel wrote: > we are currently struggling with a minor issue with our Review Board > (v1.6.1) installation: If we reference a file from Subversion (as r1) > in a given review request that has been deleted in a later revision > (let's say, r2) along with the directory hosting it, Review Board is > unable to display the review request (after clearing memcached) as it > is unable to find the file in the Subversion repository. AFAICS this > is caused by the fact that Review Board attempts to locate the file > (using a PROPFIND) in a later revision r2+x, which fails. Just for completeness: This is the backtrace I am getting (50149 is the current HEAD revision): -- 8< --Traceback (most recent call last): File "/usr/lib64/python2.6/site-packages/reviewboard/diffviewer/views.py", line 151, in view_diff interdiffset, highlighting, True) File "/usr/lib64/python2.6/site-packages/reviewboard/diffviewer/diffutils.py", line 1071, in get_diff_files large_data=True) File "/usr/lib64/python2.6/site-packages/djblets/util/misc.py", line 156, in cache_memoize data = lookup_callable() File "/usr/lib64/python2.6/site-packages/reviewboard/diffviewer/diffutils.py", line 1070, in enable_syntax_highlighting)), File "/usr/lib64/python2.6/site-packages/reviewboard/diffviewer/diffutils.py", line 551, in get_chunks old = get_original_file(filediff) File "/usr/lib64/python2.6/site-packages/reviewboard/diffviewer/diffutils.py", line 364, in get_original_file large_data=True)[0] File "/usr/lib64/python2.6/site-packages/djblets/util/misc.py", line 156, in cache_memoize data = lookup_callable() File "/usr/lib64/python2.6/site-packages/reviewboard/diffviewer/diffutils.py", line 363, in data = cache_memoize(key, lambda: [fetch_file(file, revision)], File "/usr/lib64/python2.6/site-packages/reviewboard/diffviewer/diffutils.py", line 342, in fetch_file data = tool.get_file(file, revision) File "/usr/lib64/python2.6/site-packages/reviewboard/scmtools/svn.py", line 138, in get_file raise FileNotFoundError(path, revision, str(e)) FileNotFoundError: The file '/x/X.java' could not be found in the repository: '/subversion/x/!svn/bc/50149/x/X.java' path not found -- 8< -- Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Handling of files attached to a review request deleted at a later revision in Subversion
Hi there, we are currently struggling with a minor issue with our Review Board (v1.6.1) installation: If we reference a file from Subversion (as r1) in a given review request that has been deleted in a later revision (let's say, r2) along with the directory hosting it, Review Board is unable to display the review request (after clearing memcached) as it is unable to find the file in the Subversion repository. AFAICS this is caused by the fact that Review Board attempts to locate the file (using a PROPFIND) in a later revision r2+x, which fails. Is anyone else seeing this issue? Thanks, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Unable to upload screenshots
On Mon, Feb 14, 2011 at 23:01, Christian Hammond wrote: > I don't believe they're stored temporarily anywhere. If so, it's internal to > Django and I don't know it off-hand. > > The only thing I've ever seen permission-wise is when there's a directory > within media/uploaded that doesn't have the proper write permissions for the > server. Hm... Actually, after doing some more testing I think that permissions aren't an issue as GIF files can be uploaded properly. I also can't imaging why libpng should be loadable from the command line, but not from the Apache process... Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Unable to upload screenshots
On Tue, Jan 26, 2010 at 18:33, Thilo-Alexander Ginkel wrote:
> On Thu, Jan 7, 2010 at 08:34, David Resnick wrote:
>> Well, loading the image in PIL from the command line works.
>
> I am facing the same issue. Command line image loading works, but
> uploading through the RB UI fails:
>
> $ python
> Python 2.6 (r26:66714, Feb 21 2009, 02:16:04)
> [GCC 4.3.2 [gcc-4_3-branch revision 141291]] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
>>>> from PIL import Image
>>>> Image.open("screenshot.png")
>
I'd like to revive this thread as I just received another user report
about this issue. As this is still working from the command line, but
not through the Web UI, I was wondering whether this may be a
permission problem for the file upload. Which directory are the
uploaded screenshots supposed to be written to temporarily?
Thanks,
Thilo
--
Want to help the Review Board project? Donate today at
http://www.reviewboard.org/donate/
Happy user? Let us know at http://www.reviewboard.org/users/
-~--~~~~--~~--~--~---
To unsubscribe from this group, send email to
reviewboard+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/reviewboard?hl=en
Re: How to submit review for merging branch to trunk?
On Tue, Oct 19, 2010 at 18:55, wrote: > (Using Subversion...) > > I have some code checked in on a branch. Before merging back to trunk I > would like to create a review request for the merge. It's not clear at all > to me how I would do that with post-review (or even using the form on the > website). In post-review --help output I see mention of a --branch=BRANCH > option, but there's no example to suggest the form BRANCH is supposed to > take. In the past I've just updated my trunk sandbox from the branch and > used post-review there. Is that the correct approach? Yep, I'd merge to your working copy of trunk and submit the review from there before committing. --branch=BRANCH is just for documentation purposes and will show up in the review request. Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Review performance issue
On Fri, Oct 15, 2010 at 09:22, Mohan wrote: > Only RB is running on the server. > > when i did 'ps' it shows the following apache instances. > > [r...@rboard log]# ps -ef | grep httpd > apache 1279 22272 0 01:51 ? 00:00:03 /usr/sbin/httpd > apache 12843 22272 0 07:29 ? 00:00:00 /usr/sbin/httpd > [...] > apache 22453 22272 0 Oct14 ? 00:00:04 /usr/sbin/httpd > > just thinking on Prefork model. You're already running prefork... Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Review performance issue
On Thu, Oct 14, 2010 at 15:29, mohan kumar wrote: > Thanks for your response. > > apache (httpd) uses more cpu and memory: > > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND > 14071 apache 25 0 449m 116m 6124 R 89.0 2.1 0:34.34 httpd Hm, just had a secons look at your first memory statistics mail. I may be mistaken, but ~ 5.7 GB of buffers seems suspicious. Is there anything else on the machine in addition to RB? Apart from that, is there only a single httpd process? I have used apache in threaded mode in the past, but went back to the prefork model as that turned out to be much more reliable for me. If you use prefork, make sure that the number of available (idle) forks matches your load requirements. Forking mostly uses COW, so having a couple of extra idle workers is typically not as expensive as it may initially seem. Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: RB 1.0.9 -> 1.5 upgrade woes
Hi all, is there someone, who uses RB 1.5 with X.509 auth, Django 1.2.x (not sure whether that makes any difference) and mod_python? After thinking a while about the issue I assume that it is also present with the stock X.509 auth middleware. Regards, Thilo On 12 Oct 2010 17:12, "Thilo-Alexander Ginkel" wrote: > On Mon, Oct 11, 2010 at 19:02, Thilo-Alexander Ginkel wrote: >> ACK. I have set up a test installation based on production data and >> will play around with it a little over the next days to pinpoint what >> exactly went wrong. > > After some further investigation it seems that mod_python seems to > recycle request objects, which causes > X509AuthMiddleware.process_request to be invoked with a request, which > has a non-anonymous user. This isn't a big deal as long as all > requests come in with a X509_USERNAME_FIELD in the environment. > However, requests that do not use X.509 authentication will most > likely just inherit the user that is already attached to the request. > > I'd appreciate some feedback if that makes any sense and if this > pre-population of the request with an old user identity is intentional > or a defect in some other component. > > Regards, > Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Review performance issue
On Thu, Oct 14, 2010 at 07:53, Mohan wrote: > I have noticed that high memory usage on this machine. > > [r...@rboard ~]# free > total used free shared > buffers cached > Mem: 5760760 5726016 34744 0 220 > 6680 > -/+ buffers/cache: 5719116 41644 > Swap: 2096472 2096472 0 Can you identify (using ps or top) which processes consume so much memory? Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: starting review #
On Tue, Oct 12, 2010 at 23:00, Christian Hammond wrote: > Another possibility is that it's your repository. If that's being slow to > reach, things will time out in Review Board. During the creation of a new > review request, it tries to access files from there, and can time out if it > can't access it. Yet another possibility: The SMTP server. Mails are sent synchronously on review request creation, so if the SMTP server is slow, this will slow down review request creation. Solution: Operate a local SMTP server that relays to the corporate SMTP. Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: RB 1.0.9 -> 1.5 upgrade woes
On Mon, Oct 11, 2010 at 19:02, Thilo-Alexander Ginkel wrote: > ACK. I have set up a test installation based on production data and > will play around with it a little over the next days to pinpoint what > exactly went wrong. After some further investigation it seems that mod_python seems to recycle request objects, which causes X509AuthMiddleware.process_request to be invoked with a request, which has a non-anonymous user. This isn't a big deal as long as all requests come in with a X509_USERNAME_FIELD in the environment. However, requests that do not use X.509 authentication will most likely just inherit the user that is already attached to the request. I'd appreciate some feedback if that makes any sense and if this pre-population of the request with an old user identity is intentional or a defect in some other component. Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: RB 1.0.9 -> 1.5 upgrade woes
Hi Christian, On Sat, Oct 9, 2010 at 08:15, Christian Hammond wrote: > We don't have any permission enforcement for submitting review requests, so > I can't imagine what would be causing that problem. Nor can I imagine why > everything would be using your admin user. These are base things that > haven't really changed in 1.5, and we certainly haven't seen these issues > yet. > > Do you happen to have a diff of your changes to post-review? This one may be a little tricky. ;-) While I have a full history of my changes to post-review in SVN, the initial "official" post-review codebase is most likely not amongst them. So, if I figure out the upstream release I may be able to create a full diff. > And do you have any changes to Review Board itself? Yep. Custom SCMTool for our proprietary SCM + X.509+ActiveDirectory integration. So, yes, the authentication part is custom and I cannot completely rule out that I messed it up when porting it to RB 1.5. > What happens if you upload a diff through the web UI instead, or through > standard post-review? Didn't try this yet, as our SCMs (Perforce + Proprietary) require some post-processing of the diffs to be usable for RB. > My gut feeling is that this is something specific to your customizations, but > I'd like to know for sure. ACK. I have set up a test installation based on production data and will play around with it a little over the next days to pinpoint what exactly went wrong. Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Active Directory
On Mon, Oct 11, 2010 at 15:18, Al West wrote: > Okay so at least I know it works! I did try using LDAP settings but > it's only my second time using LDAP on Linux so I was getting the > parameters wrong and locked myself out a few times. How does one > reset the authentication settings if this happens. Though I don't > have anything uploaded to review board it would save me doing a > reinstall each time I lock myself out. Reinstalling is not strictly necessary as you can tweak the site configuration (and reset it to safe defaults) on the database. Apart from that you may want to look into tcpdump to figure out whether any communication is happening at all and where it starts failing (does it talk to the LDAP server, does the DNS lookup fail, etc.). Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: KeyError at /admin/db/scmtools/repository/add/
Hi Alex, On Sat, Oct 9, 2010 at 16:47, Alex wrote: > Exception Type: KeyError at /admin/db/scmtools/repository/add/ > Exception Value: 'tool' I guess your scmtools on the database is empty. I have seen this once on a test upgrade - I guess something went wrong during the evolution. There is a manage.py command available to populate the table again (don't remember it right now). Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
RB 1.0.9 -> 1.5 upgrade woes
Hello everyone, today I tried to upgrade our RB 1.0.9 installation to RB 1.5, but in the end had to restore the old state from backup. After upgrading RB and its dependencies as well as the site (using rb-site), Review Board initially seemed to continue working like a charm, i.e., existing reviews were still displaying correctly (after restarting memcached and increasing its permissible key size). However, when testing submission of new reviews, a couple of issues started surfacing: - Regular users could no longer submit review requests (permission / privilege problem) - Superusers submitting review requests could do so successfully, but all of these requests would show up as drafts under my user ID (2) I could explain the privilege issue (does 1.0.x perform privilege checks on review submission?), but the new magnetic personality of my user attracting any new review request is something that I have no good explanation for. ;-) Any ideas helping to sort out the root cause are much appreciated. Regards, Thilo P.S.: I should add that we are using a post-review client, which is heavily customized and thus still using the 1.0.x API. -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: automated script to mark a review closed?
Hi Jen, On Thu, Oct 7, 2010 at 22:11, Jennifer wrote: > Firstly, we have been using reviewboard for about a year now and it is > really great! Thanks to all the devs for your hard work. I am > looking forward to trying out the new release. > > We still have a problem where people will check in an approved fix, > but forget to close the review, which clutters up the dashboard. Has > anyone written a command line script that can automatically close a > review? Our developers are required to put the review number in the > checkin message, so we would like to add a hook to cvs (yes we are > *still* using cvs) to call a script that will close the review. I guess you can use this SVN hook as a starting point: http://github.com/reviewboard/reviewboard/blob/master/contrib/tools/svn-hook-set-submitted Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Review Board 1.5 Released!
On Fri, Oct 1, 2010 at 09:15, Christian Hammond wrote: > We just released Review Board 1.5! A lot went into this release, and I think > it turned out really well. Congratulations and thanks a lot for the continued effort of all contributors! Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: RB integration with Redmine bug tracker
On Thursday 02 September 2010 21:09:59 Christian Hammond wrote: > I want to first make sure we're both clear on what you're trying to do and > what you can do. Review Board can't post to bug trackers today. It's a way > to link to existing bug trackers, so that if someone references an existing > filed bug in a review request, we can provide a handy link to it. In this > case, to link, you'd probably want something like > http://xx.xx.x.xxx:xxx/projects/test_project/issues/%s > (where %s is the placeholder for the bug number). The bug URL for Redmine should read: http(s)://:/issues/%s [Assuming you installed Redmine in the root directory of the respective (virtual) host.] Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: User Account Management through REST API
On Wednesday 30 June 2010 20:34:31 Christian Hammond wrote: > You actually can add a user in LDAP who is not in Review Board as a > reviewer. They won't appear in the autocomplete, but if you type the > name it will work. Review Board checks the auth backend for any > unrecognized names that are entered and pulls them down then. Is this a new feature in 1.5? We are currently still running 1.0.5.1, which does not seem to support this. Would be another incentive to upgrade. ;-) Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: User Account Management through REST API
On Wed, Jun 30, 2010 at 11:19, Christian Hammond wrote: > I know what other people have done is written custom auth backends to talk > to a central database for users, so instead of creating/updating users in > Review Board based on redmine's data, you'd just have Review Board use > redmine for authentication in some way. > > I'm not familiar enough with redmine, but if it could use something like > LDAP for authentication, that'd be even easier. Redmine does support LDAP as a backend and adding a custom authentication backend to Redmine is also trivial (just implement an AuthSource - takes less than 40 lines of code for simple DB-based authentication). What I could consider somewhat problematic with these custom auth backends in both Review Board and Redmine is the "pull" approach that is implemented by them: The user account data is only pulled from the backend when an account is used for the first time, so you cannot add somebody as a reviewer before he has signed on at least once. Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Marking a review as currently being reviewed
On Wed, May 26, 2010 at 21:47, David Trowbridge wrote: > There aren't any better ways to do it, since this is kind of contrary > to the use cases we had when building RB. I personally like it when > more than one person looks at a review--more eyes means more bugs are > caught early on. I can second this. From my experience with a medium-sized RB installation it is the reviews involving more than a single reviewer, which end up finding most bugs. Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: RBTools UnicodeDecodeError of fresh VS2K8 project from fresh Win7 x64 + Py26 + RBTools0.2RC1
On Thursday 25 February 2010 01:55:42 Pv wrote: > I am pretty sure this aborts in RBTools itself before it ever gets to > the server. > Again, a manual upload of the diff file to the server works fine. Could http://reviews.reviewboard.org/r/1298/ have caused this? Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: ReviewBoard Error while posting review for GIT
On Tuesday 26 January 2010 18:55:20 Kunjal wrote: > Here is my config looks like: > > [core] > repositoryformatversion = 0 > filemode = true > bare = false > logallrefupdates = true > [remote "origin"] > url = > git://mobcom-git.sj.broadcom.com/git_repos/repo_mydroid/vendor.git fetch = > +refs/heads/*:refs/remotes/origin/* > [branch "donut"] > remote = origin > merge = refs/heads/donut > [reviewboard] > url = http://mob-rb-test.sj.broadcom.com What happens when you manually specify the RB server URL on the post-review command line? Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Unable to upload screenshots
On Thu, Jan 7, 2010 at 08:34, David Resnick wrote:
> Well, loading the image in PIL from the command line works.
I am facing the same issue. Command line image loading works, but
uploading through the RB UI fails:
$ python
Python 2.6 (r26:66714, Feb 21 2009, 02:16:04)
[GCC 4.3.2 [gcc-4_3-branch revision 141291]] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> from PIL import Image
>>> Image.open("screenshot.png")
>>>
Have you been able to figure out what was wrong in your case?
Unfortunately, I cannot add any further details as Review Board only
intermittently writes to its log file (the last time on Jan 16th -
nothing has been changed since then).
Regards,
Thilo
--
Want to help the Review Board project? Donate today at
http://www.reviewboard.org/donate/
Happy user? Let us know at http://www.reviewboard.org/users/
-~--~~~~--~~--~--~---
To unsubscribe from this group, send email to
reviewboard+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/reviewboard?hl=en
Re: What is the status of ClearCase integration on Windows?
On Friday 15 January 2010 23:20:32 Sassan wrote: > I am also thinking it might be a good idea to add a repository > independent base functionality to the post-review script where it is > handed the root directory of two source trees, before and after the > change and it will then just compare the files and post a review. > > This way anyone with any source repository can just create the before > and after soure trees outside RB and pass the roots of the source > trees to the post-review script for posting. This won't work as Review Board needs to be able to access the respective SCM repository from the server-side to apply the posted diff to the base revision. Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Enable search in RB 1.0.5.
Hi,
IIRC the root cause for this issue is that your Lucene version is too new.
Using the latest 2.x version should do the trick.
Regards,
Thilo
On 10 Jan 2010 03:56, "Kunjal" wrote:
Hello,
We want to enable search in RB. We are having issues.
I've finally built java6,ANT,jcc,and pylucerne and I can see that I
can enable the search function with /opt/lampp/htdocs/reviewboard/
search-index where index files will reside:
I get the following errors when doing a full index:
#[r...@mob-rb-test reviewboard]# rb-site manage /opt/lampp/htdocs/
reviewboard index -- --full
Traceback (most recent call last):
File "/opt/lampp/bin/rb-site", line 8, in
load_entry_point('ReviewBoard==1.0.5.1', 'console_scripts', 'rb-
site')()
File "/projects/mob_tools/xampp/1.6.4-brcm-v2/lib/python2.5/site-
packages/ReviewBoard-1.0.5.1-py2.5.egg/reviewboard/cmdline/rbsite.py",
line 1478, in main
command.run()
File "/projects/mob_tools/xampp/1.6.4-brcm-v2/lib/python2.5/site-
packages/ReviewBoard-1.0.5.1-py2.5.egg/reviewboard/cmdline/rbsite.py",
line 1412, in run
site.run_manage_command(args[0], args[1:])
File "/projects/mob_tools/xampp/1.6.4-brcm-v2/lib/python2.5/site-
packages/ReviewBoard-1.0.5.1-py2.5.egg/reviewboard/cmdline/rbsite.py",
line 352, in run_manage_command
execute_manager(reviewboard.settings, [__file__, cmd] + params)
File "/projects/mob_tools/xampp/1.6.4-brcm-v2/lib/python2.5/site-
packages/Django-1.1.1-py2.5.egg/django/core/management/__init__.py",
line 362, in execute_manager
utility.execute()
File "/projects/mob_tools/xampp/1.6.4-brcm-v2/lib/python2.5/site-
packages/Django-1.1.1-py2.5.egg/django/core/management/__init__.py",
line 303, in execute
self.fetch_command(subcommand).run_from_argv(self.argv)
File "/projects/mob_tools/xampp/1.6.4-brcm-v2/lib/python2.5/site-
packages/Django-1.1.1-py2.5.egg/django/core/management/base.py", line
195, in run_from_argv
self.execute(*args, **options.__dict__)
File "/projects/mob_tools/xampp/1.6.4-brcm-v2/lib/python2.5/site-
packages/Django-1.1.1-py2.5.egg/django/core/management/base.py", line
222, in execute
output = self.handle(*args, **options)
File "/projects/mob_tools/xampp/1.6.4-brcm-v2/lib/python2.5/site-
packages/Django-1.1.1-py2.5.egg/django/core/management/base.py", line
351, in handle
return self.handle_noargs(**options)
File "/projects/mob_tools/xampp/1.6.4-brcm-v2/lib/python2.5/site-
packages/ReviewBoard-1.0.5.1-py2.5.egg/reviewboard/reviews/management/
commands/index.py", line 68, in handle_noargs
store = lucene.FSDirectory.getDirectory(store_dir, False)
AttributeError: type object 'FSDirectory' has no attribute
'getDirectory'
Any idea?
Thanks
Kunjal
--
Want to help the Review Board project? Donate today at
http://www.reviewboard.org/donate/
Happy user? Let us know at http://www.reviewboard.org/users/
-~--~~~~--~~--~--~---
To unsubscribe from this group, send email to
reviewboard+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/reviewboard?hl=en
--
Want to help the Review Board project? Donate today at
http://www.reviewboard.org/donate/
Happy user? Let us know at http://www.reviewboard.org/users/
-~--~~~~--~~--~--~---
To unsubscribe from this group, send email to
reviewboard+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/reviewboard?hl=en
Re: One or more fields had errors: fields: 'path': 'This field is required.'
On Tue, Dec 15, 2009 at 15:28, Thilo-Alexander Ginkel wrote: > The question is, how can I calculate the length of this string in > bytes, not characters, taking the wire encoding into account? A patch is available at: http://reviews.reviewboard.org/r/1298/ Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: One or more fields had errors: fields: 'path': 'This field is required.'
On Mon, Dec 14, 2009 at 23:11, Christian Hammond wrote: > It should just be the field constraints from those FileFields. We don't do > any custom validation checks in those forms that I can see. > > It could potentially fail if the diff itself is empty, even though the name > is populated. Are they uploading using post-review? Might be worth seeing > what post-review --output-diff shows. Or otherwise, just making sure the > diff is non-empty. > > Short of that, I wouldn't be able to see off-hand without seeing the code > and doing my own debugging. After a lengthy debugging session I think I figured out what is going on, although the solution still is not 100% clear to me: - In json.new_diff the request.POST field is completely empty, so Django is completely right when complaining about the absence of some fields. - On the client side, though, the request looks sensible and the MIME body is correctly built. - In post-review's http_post method, however, the Content-Length header is calculated as len(body), which will calculate the length of the body incorrectly if it contains non-ASCII characters. So, in the end the sent Content-Length header is shorter than the actual request body, which will cause Django not to find the terminating MIME boundary thus ignoring the body. The question is, how can I calculate the length of this string in bytes, not characters, taking the wire encoding into account? > Though it's probably not the cause, it also might be worth checking if there > are any proxy settings getting in the way. We used to run into some issues > with users who were going through the proxy for the Review Board server > (despite it being within the network), and some stuff broke along the way. > Though I don't believe we hit file upload issues. I disabled proxies in our custom post-review implementation, so that's not an issue. Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: One or more fields had errors: fields: 'path': 'This field is required.'
On Monday 14 December 2009 22:28:35 Christian Hammond wrote: > Which version of Review Board is this running right now? I've seen this > recently on the 1.1 alpha nightlies. It's a slightly patched 1.0.5.1. > Also, what repository type is this? It could indicate a missing dependency. > The error checking/reporting for that on 1.0.x is pretty bad, but improved > in 1.1. It's a custom SCM, so I cannot rule out that the bug is in my code. Dependencies should not be an issue, though, as the SCM client code is mostly self-contained and does all of its communication with the SCM via HTTP. Do you have an idea what may be causing form.is_valid() to return false? Is it just the mandatory field constraints, which are being checked? The weird thing is, that according to my client-side dump the 'path' attribute is populated correctly. > Is this only happening to some people and not everyone? Yes. Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
One or more fields had errors: fields: 'path': 'This field is required.'
Hello everyone,
I am currently somewhat stuck figuring out the cause of an error that
some colleagues are getting in response to the
/api/json/reviewrequests//diff/new/ call when posting a new review
request:
{u'fields': {u'path': [u'This field is required.']}, u'stat': u'fail',
u'err': {u'msg': u'One or more fields had errors', u'code': 105}}
I did some debugging on the server side and it seems that in new_diff
in json.py the form.is_valid() call returns false. I also enhanced
post-review with some debugging code and can see that the following
request is dispatched to the server:
>>> Posting API request: path=api/json/reviewrequests//diff/new/,
>>> fields={}, files={'path': {'content': u'', 'filename': 'diff'}}
Any idea what may be wrong here?
Thanks,
Thilo
--
Want to help the Review Board project? Donate today at
http://www.reviewboard.org/donate/
Happy user? Let us know at http://www.reviewboard.org/users/
-~--~~~~--~~--~--~---
To unsubscribe from this group, send email to
reviewboard+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/reviewboard?hl=en
Re: post-review with bypass proxy settings in Internet Explorer
On Wednesday 02 December 2009 22:08:26 Chris Clark wrote: > Modifying the registry and then restoring is not a great idea. I can see > why you are doing it but I'd encourage you to NOT do this. There is a > potential here for a background web app to fail (e.g. web browser based > IM tool). Not only that, but there is an ugly race condition hidden in that pattern: Start post-review twice in parallel and you might end up with no configured proxy if you have the following execution order: Instance 1Instance 2 ---- p := read setting disable proxy p2 := read setting disable proxy set proxy <- p set proxy <- p2 I fixed the issue for my installation using the approach suggested by Chris in <1eb5631b0911241645m59efcbe0i6c5de6c600313...@mail.gmail.com>, which works like a charm. Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: RB and Browser Compatibilty.
On Wednesday 02 December 2009 02:18:17 Kunjal wrote: > We are on verge of rolling out RB 1.0.5 to 300+ developers. > Some of our developers use IE and some of them use FireFox. > Does RB 1.0.5 work with both browser seamless? > During my testing I did not find the difference but I might have > missed something. There is one noticeable difference: For IE, lines do not wrap for diffs wider than your screen. Firefox wraps those lines. At least that's my observation. I do, however, run a couple of extensions, which may have an influence on this behavior. I like the wrapped lines better, but that's just my personal preference. IE6 tends to freeze when executing RB's JavaScript from time to time. > Pl. let me know if there is any preference of browser we should > enforce to our developers. You might want to ask your developers what they prefer instead of enforcing something... Regards, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Review Board diff issues: table.rows[endRow] is undefined
Hello everyone, I just stumbled across a problem while using Review Board 1.0.5.1 that seems to be related to the retrieval of diffs while the diff viewer is loading. The retrieval of the third file in my diff list seems to trigger a Javascript error, which I can see in my browser's error console and aborts the retrieval process for all subsequent diff items: -- 8< -- Error: table.rows[endRow] is undefined Source File: https://reviewboard./media/rb/js/diffviewer.js?1256731641 Line: 869 -- 8< -- Unfortunately, that's all the info I have right now. Has anybody seen this kind of error before? If required, I could try to debug the Javascript to figure out what's going wrong... Thanks, Thilo -- Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en
Re: Error uploading diff: One or more fields had errors (105)
On Monday 16 November 2009 11:50:36 developer 236 wrote: > i installed p4 tool also @ the server m/c, but still getting the same > error. The p4 is also in path environment variable. Is the PATH modification also visible to the Apache process hosting Review Board? I don't know which OS you are running, but under Windows changing the System PATH in the Control Panel is required, but not sufficient to update the search path for a service. Typically at least a service (if not server) restart is needed to propagate the environment modification to the service process. Regards, Thilo --~--~-~--~~~---~--~~ Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: Integrated authentication with Active Directory
On Friday 13 November 2009 17:36:17 Akhilesh wrote: > Thanks Thilo for reply. I see that "Find DC from DNS" remains > disabled. How do I enable it? IIRC you need to install pydns (e.g., using easy_install). Regards, Thilo --~--~-~--~~~---~--~~ Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: Integrated authentication with Active Directory
On Fri, Nov 13, 2009 at 03:37, Akhilesh wrote: > I'm having some trouble setting up integrated authentication with AD. > Here is what I have done- > 1. joined VM hosting ReviewBoard site to company.com > 2. From Admin menu, set "Authentication method" = Active Directory > 3. Set domain name = Company.com > 4. installed python_ldap-2.3.7-py2.5.egg-info > > When I try to login using my Company.com\Akhilesh credentials, it says > invalid login or password. > Apache Error log says: > > WARNING:root:Active Directory: Domain controller is down > ERROR:root:Active Directory error: Could not contact any domain > controller servers Did you tick the "Find DC from DNS" checkbox or did you manually specify the domain controller? Regards, Thilo --~--~-~--~~~---~--~~ Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: Review Board 1.0.5 released
On Thu, Oct 22, 2009 at 14:12, Paolo Borelli wrote: > I upgraded to 1.0.5 and I am having problems... Creating new review > requests from the web interface fails: after selecting the file, when > you press the "Create Review Request" simply nothing happens, no error > message or even log message. Do you get any errors in the JavaScript error console? Regards, Thilo --~--~-~--~~~---~--~~ Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: Question about AD integration and users/groups
On Wed, Oct 14, 2009 at 19:18, Alfred von Campe wrote: > SMTPException: No suitable authentication method found. > > I have looked but can't figure out where I am supposed to set an > authentication method. As far as I know, our internal SMTP server > doesn't even require a username and password, but I did supply one on > the email configuration page just in case. Did you try without setting the credentials? Regards, Thilo --~--~-~--~~~---~--~~ Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: Question about AD integration and users/groups
On Wed, Oct 14, 2009 at 15:07, Alfred von Campe wrote: > Speaking of email, is there a way to > configure RB to send out email whenever a review is published or > updated? That's actually the standard behavior. Did you maintain an SMTP server in the settings? Regards, Thilo --~--~-~--~~~---~--~~ Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Installing PyLucene on Windows
Hello everyone, I was thinking about adding search support to our Review Board installation, which would require the installation of PyLucene. Has somebody succeeded in getting this working on a Windows-based platform? Thanks, Thilo --~--~-~--~~~---~--~~ Want to help the Review Board project? Donate today at http://www.reviewboard.org/donate/ Happy user? Let us know at http://www.reviewboard.org/users/ -~--~~~~--~~--~--~--- To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: "Expand" broken since 1.0.3?
On Mon, Sep 28, 2009 at 12:20, Thilo-Alexander Ginkel wrote: > On Fri, Sep 25, 2009 at 22:21, Christian Hammond wrote: >> That's odd.. I'll try to reproduce that tonight or this weekend. > > I just tried reproducing the issue, but was not successful. So, this > probably was a one-time glitch. I'll keep an eye on it to see whether > it happens again. Now I was able to figure out the conditions under which the issue manifests itself. Further details are available at: http://code.google.com/p/reviewboard/issues/detail?id=1334 Regards, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: Problem installing on Ubuntu Hardy / Git / GitHub
On Mon, Sep 28, 2009 at 12:53, Neil wrote: > I've just installed ReviewBoard on a plain Hardy install, and have > created a repository pointed at GitHub. > > However, when I try to create a new review request I get an http 500, > plus the following in the logs: > > Git is installed on the server and working. Is the git binary in the PATH for the user that is running the RB web server? Regards, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: "Expand" broken since 1.0.3?
On Fri, Sep 25, 2009 at 22:21, Christian Hammond wrote: > That's odd.. I'll try to reproduce that tonight or this weekend. I just tried reproducing the issue, but was not successful. So, this probably was a one-time glitch. I'll keep an eye on it to see whether it happens again. @Tom: Are you able to reproduce this issue in a consistent fashion? Regards, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: "Expand" broken since 1.0.3?
On Friday 25 September 2009 22:11:01 Christian Hammond wrote: > Were you trying to expand a diff that was still attached to a draft of a > review request? Or was it actually a published diff visible to everyone? It was a published diff visible to everyone (actually a colleague pointed me to the issue while he was reviewing some of my changes). The issue was in effect for both me as the review submitter as well as the reviewer. All that has changed since then are a couple of comments (some already present when the issue surfaced) as well as a "Ship It" added to the request. I can try reproducing the issue (probably only on Monday) to figure out the exact circumstances. Regards, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: "Expand" broken since 1.0.3?
On Friday 25 September 2009 21:38:35 Christian Hammond wrote: > This is a standard diff, not an interdiff? Yes, it's a standard diff. One interesting thing is that for requests created prior to 1.0.3 the fragment expansion still seems to work (I just tried a couple of requests, though). Only for newly created requests the diff # gets created incorrectly. If I specify "diff/1" instead of "diff/0" in the URL I can correctly retrieve the fragment. What is even more interesting (but also confusing) is that the expansion magically started working for the request # 280 I mentioned in my original message (I just checked it via VPN from home to get the exact URL for the bug report). The diff # in the JavaScript source has changed to 1... Any idea? Thanks, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: "Expand" broken since 1.0.3?
On Fri, Sep 25, 2009 at 17:31, Thilo-Alexander Ginkel wrote: > can somebody confirm my observation that clicking on any of the "Expand" > hyperlinks (to expand collapsed source code sections) in the RB diff viewer > no longer works with 1.0.3? > Just a brief update: The root cause seems to be a 404 error: Page not found (404) Request Method: GET Request URL: https://reviewboard ./r/280/diff/0/fragment/2785/chunk/8/ Regards, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
"Expand" broken since 1.0.3?
Hello everyone, can somebody confirm my observation that clicking on any of the "Expand" hyperlinks (to expand collapsed source code sections) in the RB diff viewer no longer works with 1.0.3? Thanks, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: Issue 513 in reviewboard: Improvements for authentication using REMOTE_USER
On Thu, Aug 20, 2009 at 23:37, Christian Hammond wrote: > Good question. I don't know that anyone's tested that or worked on it. I > imagine it doesn't work right today, but could be done. You'd probably have > to deal with two password prompts. Shouldn't it be possible to implement this with a similar authentication middleware that Nathan H. proposed for the X.509 authentication (see [1])? The env variable will be a different one and you will probably also need your own auth backend, but the basic concept should remain the same. Regards, Thilo [1] http://reviews.review-board.org/r/938/diff/1/?file=4943#file4943line45 --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Best practices for supporting multiple projects with a RB installation
Hello everybody, I was wondering whether there are any best practices or experiences on how to support multiple projects (within the same organization) with a Review Board installation (the developers working on each project are typically different people). Would you try to fit everything into the same site or rather install a RB site for each project? Thanks, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: Windows post-review .exe
On Wednesday 19 August 2009 23:03:33 Dana Lacoste wrote:
> So, I have everything working (from trunk, with a patch that I've
> uploaded for review :) to get my perforce users working with
> reviewboard. That is, I can, from Windows, run post-review and create
> a review with a diff.
>
> WooHoo!
>
> The only catch is that I've been asked to make it not require so many
> installs: specifically, I've been asked to make a single .zip
> containing all of the .dll and .exe files needed for this to run.
>
> "OK" says I : that should be do-able. I've done it before for other
> python scripts.
>
> But I can't for the life of me figure out how to get around the .egg
> files (for rbtools and for json). I can build a post-review-
> script.exe that can't find RBTools, and I can build a postreview.exe
> that can't find simple json, but I can't build what they're asking
> for.
How are you building the .exe? I am using py2exe with an older, but highly
customized version of post-review (so I do not know whether this works with
RBTools) and have been successful with the following setup.py script (you need
to adjust the icon/data file references to make this work with the vanilla
post-review):
-- 8< --
from distutils.core import setup
import py2exe
setup(
windows = [{"script": 'post-review.py', "icon_resources": [(1,
"gui/icons/review.ico")], 'dest_base': "post-review-gui"}],
console = [{"script": 'post-review.py', "icon_resources": [(1,
"gui/icons/review.ico")]}],
options = {"py2exe": {
"compressed": 1,
"optimize": 2
},
},
data_files = [("gui/icons",
["gui/icons/review.png", "gui/icons/review.ico"]),
("bin",
["msys/rbdiff.exe", "msys/msys-1.0.dll"]),
],
)
-- 8< --
Regards,
Thilo
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"reviewboard" group.
To post to this group, send email to reviewboard@googlegroups.com
To unsubscribe from this group, send email to
reviewboard+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/reviewboard?hl=en
-~--~~~~--~~--~--~---
File deletion diffs not applying cleanly
Hello everybody, while this is most likely not 100% related to the core Review Board implementation (but rather to the format of the diff I am using) you probably have an idea why my patches related to file deletions do not apply cleanly and thus cause Review Board to bail out when trying to display the diff using the "View Diff" button in the review request. An example error message: -- 8< -- The patch to 'SomeFile.java' didn't apply cleanly. The temporary files have been left in 'c:\windows\temp\reviewboard.fuosla' for debugging purposes. `patch` returned: patching file 'c:\windows\temp\reviewboard.fuosla\tmpje8luv' Reversed (or previously applied) patch detected! Assume -R? [n] Apply anyway? [n] Skipping patch. 1 out of 1 hunk ignored -- saving rejects to file 'c:\windows\temp\reviewboard.fuosla\tmpje8luv-new.rej' -- 8< -- The diff: -- 8< -- --- SomeFile.java /repos/path/to/SomeFile.java#1 +++ SomeFile.java 2009-08-14 18:19:44 @@ -1,28 +0,0 @@ -The quick brown fox jumps over the lazy dog... -Lorem ipsum... -- 8< -- And finally the content of tmpje8luv: -- 8< -- The quick brown fox jumps over the lazy dog... Lorem ipsum... -- 8< -- As far as I can see the command line equivalent of the patch command executed by Review Board in this example would be: patch -o tmpje8luv-new tmpje8luv < SomeFile.java.diff which comes up with the same reverse patch message when I run it on the command line. I cannot see any obvious errors with the patch, yet, something seems to be wrong. Any idea? Thanks, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: Trouble posting reviews since 1.0.1 upgrade
On Wed, Aug 12, 2009 at 23:53, Christian Hammond wrote: > It sounds like it's actually using the ClearCase SCM instead of your own. > Are you sure your repository entry is still mapping to your custom SCM? > > In 1.0, we accidentally left out the database entry for the ClearCase SCM, > and added it back. When you did the rb-site upgrade, it probably replaced > your entry with the ClearCase one. > > This is pretty much the problem with using the database for these > registrations. I'm looking into doing some introspection to find all the > available SCMs instead of looking them up from a database so that this won't > be an issue in the future. You're right. Our SCM was replaced by ClearCase. After adding a new row for our SCM and fixing the configuration for the respective repository everything is working again. Thanks for the speedy help! Regards, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: Trouble posting reviews since 1.0.1 upgrade
On Wednesday 12 August 2009 22:55:44 Christian Hammond wrote:
> Given that this error has been seen on both yours and ClearCase, it's
> possible it's somewhere internal to Review Board.
>
> Try this... Edit reviewboard/webapi/json.py, scroll down to the new_diff
> function, and locate the "except Exception, e:" line.
>
> Remove that whole block for that exception (including the except: line) and
> then restart your web server.
>
> If you run post-review with --debug again, you should see an error page
> instead (assuming you still have DEBUG set to True). I imagine this is
> where the error is being caught and turned into a field error.
>
> I'd really like to know what the results are of this test, and where the
> problem is.
Strangely, RB seems to route the diff upload through the ClearCase SCM. The
error happens at:
[...]
reviewboard\scmtools\core.py:51
reviewboard\scmtools\clearcase.py:43
reviewboard\scmtools\clearcase.py:81:
elem_path = elem_path[elem_path.rindex("vobs/")+5:]
Vars:
drive ''
elem_path u'/MessageConfiguration.java'
path u'/MessageConfiguration.java'
Thanks,
Thilo
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"reviewboard" group.
To post to this group, send email to reviewboard@googlegroups.com
To unsubscribe from this group, send email to
reviewboard+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/reviewboard?hl=en
-~--~~~~--~~--~--~---
Re: Trouble posting reviews since 1.0.1 upgrade
Hello Christian, On Wednesday 12 August 2009 22:25:03 Christian Hammond wrote: > There's another thread on this problem as well. Are you using ClearCase? no, this happens with a proprietary SCM, for which I added support to RB and post-review. So, I cannot rule out that my scm implementation is subtly broken and the error was masked in versions up to 1.0rc2. Regards, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Trouble posting reviews since 1.0.1 upgrade
Hello everybody,
since upgrading from 1.0rc2 to 1.0.1 I have trouble posting reviews to
RB using a custom post-review implementation. The problem is that the
JSON response does not contain too many error details, just:
{"fields": {"path": ["substring not found"]}, "stat": "fail", "err":
{"msg": "One or more fields had errors", "code": 105}}
The log level is set to DEBUG, but nothing meaningful shows up in the log.
I am slightly lost how to determine the root cause of this issue. Any
advice how to diagnose it is highly appreciated.
Thanks,
Thilo
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"reviewboard" group.
To post to this group, send email to reviewboard@googlegroups.com
To unsubscribe from this group, send email to
reviewboard+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/reviewboard?hl=en
-~--~~~~--~~--~--~---
Re: X.509 Certificate Authentication
Hi Nathan, On Tue, Aug 11, 2009 at 19:24, Thilo-Alexander Ginkel wrote: >> I just uploaded my patch to http://reviews.review-board.org/r/938/. > > Thanks! I will try your changes once I have updated my RB installation > to a current version. just a brief update: I tried your patch and it works like a charm. Thanks a lot! Regards, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: X.509 Certificate Authentication
On Mon, Aug 3, 2009 at 06:44, Nathan Heijermans wrote: >> This basically sound like what I intend to implement. Would you mind sharing >> some patches? > > I just uploaded my patch to http://reviews.review-board.org/r/938/. Thanks! I will try your changes once I have updated my RB installation to a current version. Regards, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: X.509 Certificate Authentication
On Friday 31 July 2009 23:19:01 Nathan Heijermans wrote: > I have an implementation of x.509 authentication working that bypasses > the "Login" screen. I had to add a middleware class to the django site > settings (I copied the MIDDLEWARE_CLASSES from settings.py to my > settings_local.py and added my own middleware class). My middleware > class has a process_request() function that uses the mod_ssl-set > environment variables to figure out the user name, and then directly > calls the login() function in my backend. I also haven't figured out how > to get multiple authentication backends working nicely with Django; it > seemed that when I didn't explicitly call the backend login() > implementation I wanted, I randomly got the login screen anyway without > being able to log in using my password. This basically sound like what I intend to implement. Would you mind sharing some patches? > There is a remaining problem that I have yet to resolve, and that is > authenticating with my certificate from the post-review tool. I'd like > to use a password-protected, but Python doesn't make that easy at all; I > end up having to enter my password each time python makes a web request. > Having post-review is definitely nice, but the web UI doesn't seem as > opaque any more :). I would probably prefer to continue to use user/password authentication with post-review as manual setup would be required to get hold of the X.509 client key/cert from post-review whereas the certificate comes pre-installed in the web browser's secure store for all users. So, I hope your statement about the usage of multiple authentication backends being tricky does not apply to this scenario. ;-) Regards, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: X.509 Certificate Authentication
On Fri, Jul 31, 2009 at 13:27, rupert.thurner wrote: > just as a side note, edgewall trac supports it by just taking the > webservers authentication, see: > * http://trac.edgewall.org/browser/trunk/trac/web/auth.py. > * > http://trac-hacks.org/browser/sslauthenticationplugin/0.11/sslauthentication/__init__.py > * http://trac-hacks.org/wiki/SslAuthenticationPlugin Thanks for the pointer. I agree, using the Apache mod_ssl client authentication feature to do the dirty work definitely makes sense (and this is already working for me). I am however still somewhat lost with regards to the authentication backend implementation. While I could theoretically implement a backend, which just evaluates the environment variables set by mod_ssl (and ignores the password supplied to the 'authenticate' method), this would probably still require the user to click on the "Login" hyperlink in the RB web UI. What I would prefer is some implementation, which is mostly transparent and automagically signs in the user when he performs the first page hit. Do you think that's technically feasible with the current authentication architecture (or would it require a major rewrite)? Regards, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
X.509 Certificate Authentication
Hello everybody, I was wondering whether anybody has set up his Review Board installation to make use of SSL/TLS X.509 client certificate-based authentication instead of using passwords. A quick search did not bring up any related documents on the web site, so maybe it is just not supported, but I thought I'd better ask before giving up or attempting to extend RB to support it. ;-) Thanks, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Re: Integrating a custom SCM / non-numeric change IDs
On May 11, 9:55 pm, Christian Hammond wrote: > With this SCM, is the change identifier a server-stored ID that contains the > description and other information for Review Board to parse? Or is it more > like an atomic ID representing that change that gets pushed to the server > when committed? The SCM supports both committed and uncommitted changes and in both cases the change ID provides a means to retrieve meta-information about the change (such as a description, list of involved files, etc.) from the SCM server. > If the former, and if you want to support pulling that information from the > server, then we'll need a patch to make Review Board less strict about the > type of input, and move validation of the ID into the SCMTools. > > [...] I actually got a first version to work by declaring the repository not to support change sets and moving the summary / description retrieval into post-review where the full change id is still available. Nevertheless, it would be nice to be able to support this on the server side leaving less hacks in post-review (and the positive side- effect that submitting the same change a second time would update an existing review). About patching RB to be less strict: I think the exception happens somewhere in the relational database mapping, so changing this would affect both the DB schema (which currently declares the changeset ID as an int) as well as plenty other places in the code which I currently cannot yet overlook being a novice to the RB code. Thanks, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Integrating a custom SCM / non-numeric change IDs
Hello everybody, while attempting to add some basic support for a proprietary SCM to Review Board / post-review I stumbled over one issue that I'm currently uncertain how to solve: The SCM I'm integrating uses change identifiers, which are not numeric. Instead arbitrary strings are used, which (at least with alpha4) leads to the following exception when attempting to create the review request: - invalid literal for int() with base 10: '' Are there other SCMs supported by RB that also have to deal with this issue? If so, how was the problem solved for these SCMs? Any input is highly appreciated! Thanks, Thilo --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "reviewboard" group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
