Deeply sorry about noise on GitHub... If you see this comment: "Can one of the admins verify this patch?" you can ignore... I'm trying to setup a machine to trigger the testsuite on each PRs so that we know if it breaks the testsuite before it's applied. And this bot says this comment... I'll create a new user to make it clearer.. Thomas _______________________________________________ gatein-dev mailing list gatein-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/gatein-dev

Good job.

> With this it should help gate keepers to refuse a PR if it doesn't pass > the testsuite before even looking into it. that's the most important thing to retain : As a gate keeper a PR that passes does not implies it is a valid one.

On Sep 26, 2012, at 6:31 PM, Thomas Heute <theute(a)redhat.com&gt; wrote: > I will try again soon to setup the environment. > > A bot called gatein-bot will say things like: > "Can one of the admins verify this patch?" on pull requests > > It means that a PR request was made by a person who is not whitelisted, > an "admin" would then need to check that there is no malicious code in > the PR and trigger the Jenkins job by commenting "This is ok to test". > > The bot will trigger the job and give results of the testsuite, in the > pull request. > > With this it should help gate keepers to refuse a PR if it doesn't pass > the testsuite before even looking into it. > > For now the jenkins server is my private machine which is not available > outside so you won't be able to see Jenkins results (only if it passed > or failed). If it works ok, I will try to change it to have the job public. > > The configuration is very sensitive and I tried couple of times on test > repositories, I hope that it won't be too noisy... > > You can see here an example that passed: > https://github.com/theute-test/test/pull/1 > > Here if the requester is whitelisted: > https://github.com/theute-test/test/pull/2 > > Thomas > > > On 09/21/2012 03:44 PM, Thomas Heute wrote: >> Deeply sorry about noise on GitHub... >> >> If you see this comment: "Can one of the admins verify this patch?" you can ignore... >> >> I'm trying to setup a machine to trigger the testsuite on each PRs so that we know if it breaks the testsuite before it's applied. >> >> And this bot says this comment... >> >> I'll create a new user to make it clearer.. >> >> Thomas >> _______________________________________________ >> gatein-dev mailing list >> gatein-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/gatein-dev > _______________________________________________ > gatein-dev mailing list > gatein-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/gatein-dev

is it normal that the server http://10.33.144.247 cannot be accessed ? On Sep 27, 2012, at 9:55 AM, Thomas Heute <theute(a)redhat.com&gt; wrote: > On 09/26/2012 06:55 PM, Julien Viet wrote: >> Good job. > I only setup the stuff, AS7 team and an intern Honza Brázdil did all the work :) >>> With this it should help gate keepers to refuse a PR if it doesn't pass >>> the testsuite before even looking into it. >> that's the most important thing to retain : >> >> As a gate keeper a PR that passes does not implies it is a valid one. > Definitely not, and the bot doesn't close the PR in any case. > It's just for information since a PR must pass the testsuite as a condition but not the only one :) > (Condition necessaire mais pas suffisante) > > Few more info on the setup: > - Running JDK7 > - Package only with JBoss AS 5 for now > > Few results from yesterday: > https://github.com/gatein/gatein-portal/pull/147 worked beautifully, see the green bar "Good to merge — Build finished" > > https://github.com/gatein/gatein-portal/pull/124 and https://github.com/gatein/gatein-portal/pull/138 failed, but it's because they are old and don't include a patch that was ported later for JDK7 support. So the result doesn't mean anything here. > > https://github.com/gatein/gatein-portal/pull/146 has been a bit chaotic, but you can see a first commit that failed: 1f29f97 has an orange dot next to it, and later mwringe added a new commit: f1f68ae that passed (green dot). > > There are 2 more open PRs that I won't execute because they include the JDK7 issue. > > Things to add, when an admin says "ok to test" not only it generates the job but also add the PR author to the whitelist (so that next time the job is triggered right after the PR). > I will report the issue as it can be a security issue. > > So as of now we have the following on the white list: > stianst mwringe kiennguyen vstorm83 > I'm adding mstruk nscavell mto theute trongtt kenfinnigan metacosm haint kiennguyen mposolda ppalaga > based on the last PRs > > I'm adding bdaw and vietj as admin > Again the 2 things for admins are: > - comment "ok to test" when someone is not whitelisted to execute the job and add the author in whitelist for subsequent PRs > - comment "retest this please" to reinitiate a test (after a build env failure for instance) > > Thomas > > >> On Sep 26, 2012, at 6:31 PM, Thomas Heute <theute(a)redhat.com&gt; wrote: >> >>> I will try again soon to setup the environment. >>> >>> A bot called gatein-bot will say things like: >>> "Can one of the admins verify this patch?" on pull requests >>> >>> It means that a PR request was made by a person who is not whitelisted, >>> an "admin" would then need to check that there is no malicious code in >>> the PR and trigger the Jenkins job by commenting "This is ok to test". >>> >>> The bot will trigger the job and give results of the testsuite, in the >>> pull request. >>> >>> With this it should help gate keepers to refuse a PR if it doesn't pass >>> the testsuite before even looking into it. >>> >>> For now the jenkins server is my private machine which is not available >>> outside so you won't be able to see Jenkins results (only if it passed >>> or failed). If it works ok, I will try to change it to have the job public. >>> >>> The configuration is very sensitive and I tried couple of times on test >>> repositories, I hope that it won't be too noisy... >>> >>> You can see here an example that passed: >>> https://github.com/theute-test/test/pull/1 >>> >>> Here if the requester is whitelisted: >>> https://github.com/theute-test/test/pull/2 >>> >>> Thomas >>> >>> >>> On 09/21/2012 03:44 PM, Thomas Heute wrote: >>>> Deeply sorry about noise on GitHub... >>>> >>>> If you see this comment: "Can one of the admins verify this patch?" you can ignore... >>>> >>>> I'm trying to setup a machine to trigger the testsuite on each PRs so that we know if it breaks the testsuite before it's applied. >>>> >>>> And this bot says this comment... >>>> >>>> I'll create a new user to make it clearer.. >>>> >>>> Thomas >>>> _______________________________________________ >>>> gatein-dev mailing list >>>> gatein-dev(a)lists.jboss.org >>>> https://lists.jboss.org/mailman/listinfo/gatein-dev >>> _______________________________________________ >>> gatein-dev mailing list >>> gatein-dev(a)lists.jboss.org >>> https://lists.jboss.org/mailman/listinfo/gatein-dev

On Sep 27, 2012, at 10:19 PM, Thomas Heute <theute(a)redhat.com&gt; wrote: > Yes I mentioned it :) > > For now I don't have a completely working solution, but builds are published here once they are finished: > https://jenkins-gateinportal.rhcloud.com/job/gatein-portal-pull-request/ ok, would be cool indeed to monitor the WIP.

now I see it in action it's very cool, provides an extra check. which github integration feature does it use ?

> I'm currently modifying the plugin so that when it's done it points to this server URL instead (Your current build will be available here soon: https://jenkins-gateinportal.rhcloud.com/job/gatein-portal-pull-request/10/ ), but during build it still won't be accessible (running in the office behind the VPN). I wanted to move to OpenShift but the version of Jenkins there is a bit too old ok > Thomas > > > On 09/27/2012 10:11 PM, Julien Viet wrote: >> is it normal that the server http://10.33.144.247 cannot be accessed ? >> >> On Sep 27, 2012, at 9:55 AM, Thomas Heute <theute(a)redhat.com&gt; wrote: >> >>> On 09/26/2012 06:55 PM, Julien Viet wrote: >>>> Good job. >>> I only setup the stuff, AS7 team and an intern Honza Brázdil did all the work :) >>>>> With this it should help gate keepers to refuse a PR if it doesn't pass >>>>> the testsuite before even looking into it. >>>> that's the most important thing to retain : >>>> >>>> As a gate keeper a PR that passes does not implies it is a valid one. >>> Definitely not, and the bot doesn't close the PR in any case. >>> It's just for information since a PR must pass the testsuite as a condition but not the only one :) >>> (Condition necessaire mais pas suffisante) >>> >>> Few more info on the setup: >>> - Running JDK7 >>> - Package only with JBoss AS 5 for now >>> >>> Few results from yesterday: >>> https://github.com/gatein/gatein-portal/pull/147 worked beautifully, see the green bar "Good to merge — Build finished" >>> >>> https://github.com/gatein/gatein-portal/pull/124 and https://github.com/gatein/gatein-portal/pull/138 failed, but it's because they are old and don't include a patch that was ported later for JDK7 support. So the result doesn't mean anything here. >>> >>> https://github.com/gatein/gatein-portal/pull/146 has been a bit chaotic, but you can see a first commit that failed: 1f29f97 has an orange dot next to it, and later mwringe added a new commit: f1f68ae that passed (green dot). >>> >>> There are 2 more open PRs that I won't execute because they include the JDK7 issue. >>> >>> Things to add, when an admin says "ok to test" not only it generates the job but also add the PR author to the whitelist (so that next time the job is triggered right after the PR). >>> I will report the issue as it can be a security issue. >>> >>> So as of now we have the following on the white list: >>> stianst mwringe kiennguyen vstorm83 >>> I'm adding mstruk nscavell mto theute trongtt kenfinnigan metacosm haint kiennguyen mposolda ppalaga >>> based on the last PRs >>> >>> I'm adding bdaw and vietj as admin >>> Again the 2 things for admins are: >>> - comment "ok to test" when someone is not whitelisted to execute the job and add the author in whitelist for subsequent PRs >>> - comment "retest this please" to reinitiate a test (after a build env failure for instance) >>> >>> Thomas >>> >>> >>>> On Sep 26, 2012, at 6:31 PM, Thomas Heute <theute(a)redhat.com&gt; wrote: >>>> >>>>> I will try again soon to setup the environment. >>>>> >>>>> A bot called gatein-bot will say things like: >>>>> "Can one of the admins verify this patch?" on pull requests >>>>> >>>>> It means that a PR request was made by a person who is not whitelisted, >>>>> an "admin" would then need to check that there is no malicious code in >>>>> the PR and trigger the Jenkins job by commenting "This is ok to test". >>>>> >>>>> The bot will trigger the job and give results of the testsuite, in the >>>>> pull request. >>>>> >>>>> With this it should help gate keepers to refuse a PR if it doesn't pass >>>>> the testsuite before even looking into it. >>>>> >>>>> For now the jenkins server is my private machine which is not available >>>>> outside so you won't be able to see Jenkins results (only if it passed >>>>> or failed). If it works ok, I will try to change it to have the job public. >>>>> >>>>> The configuration is very sensitive and I tried couple of times on test >>>>> repositories, I hope that it won't be too noisy... >>>>> >>>>> You can see here an example that passed: >>>>> https://github.com/theute-test/test/pull/1 >>>>> >>>>> Here if the requester is whitelisted: >>>>> https://github.com/theute-test/test/pull/2 >>>>> >>>>> Thomas >>>>> >>>>> >>>>> On 09/21/2012 03:44 PM, Thomas Heute wrote: >>>>>> Deeply sorry about noise on GitHub... >>>>>> >>>>>> If you see this comment: "Can one of the admins verify this patch?" you can ignore... >>>>>> >>>>>> I'm trying to setup a machine to trigger the testsuite on each PRs so that we know if it breaks the testsuite before it's applied. >>>>>> >>>>>> And this bot says this comment... >>>>>> >>>>>> I'll create a new user to make it clearer.. >>>>>> >>>>>> Thomas >>>>>> _______________________________________________ >>>>>> gatein-dev mailing list >>>>>> gatein-dev(a)lists.jboss.org >>>>>> https://lists.jboss.org/mailman/listinfo/gatein-dev >>>>> _______________________________________________ >>>>> gatein-dev mailing list >>>>> gatein-dev(a)lists.jboss.org >>>>> https://lists.jboss.org/mailman/listinfo/gatein-dev