--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-develo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_K%2B3ksnSYePOQyWA9EWV2T%3DcwrDJvcxoZ6UrN4ggFK%2BEg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAJZOyhxbsaDSmS2EO8JR0opga7f-tRA5A5iG0TdAfQzjp_iJcg%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_JOX-fbG9gCk4DrBtn%2BmW4O%2BWGeT2%2BsHLRhND-GeAbMiQ%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAGg1j-CL5T9xrYPqN%3D_HuMUe8S30nPJYZ2p7VTPbM5HBcahYbw%40mail.gmail.com.
☆PhistucK
☆PhistucK
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_K%2B3ksnSYePOQyWA9EWV2T%3DcwrDJvcxoZ6UrN4ggFK%2BEg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAJZOyhxbsaDSmS2EO8JR0opga7f-tRA5A5iG0TdAfQzjp_iJcg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_JOX-fbG9gCk4DrBtn%2BmW4O%2BWGeT2%2BsHLRhND-GeAbMiQ%40mail.gmail.com.
---alexander
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/b7e2f291-fa71-44ae-af66-57c5048f81cf%40googlegroups.com.To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-develo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAGg1j-Be30F0jWO-%3DnRO-seMUK4%3DEo-QUkMEadAL9wVUqwx_ig%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_Lp8egQRaC0i8ahU3Ykw0c5wHushDh7jpoB2wUfb%2BMJSQ%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAGg1j-Ae8y2sf_mJSxuwo1rWfa%2B%2BdvV5MZsptAc%2Bsqb%3DTYsjcQ%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_%2Bb-_dQaruyFr1Wo4mF6zVPOif0zx3rt7Fxxo-7E4429A%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAEX3KhasQjv-4z9m6_eXP-a2nbqx9UBS-V53h3tVkqaDoeMVtA%40mail.gmail.com.
☆PhistucK
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/b7e2f291-fa71-44ae-af66-57c5048f81cf%40googlegroups.com.
---alexander---alexander
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAGg1j-Be30F0jWO-%3DnRO-seMUK4%3DEo-QUkMEadAL9wVUqwx_ig%40mail.gmail.com.
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_Lp8egQRaC0i8ahU3Ykw0c5wHushDh7jpoB2wUfb%2BMJSQ%40mail.gmail.com.
---alexander
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAGg1j-Ae8y2sf_mJSxuwo1rWfa%2B%2BdvV5MZsptAc%2Bsqb%3DTYsjcQ%40mail.gmail.com.
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_%2Bb-_dQaruyFr1Wo4mF6zVPOif0zx3rt7Fxxo-7E4429A%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAEX3KhasQjv-4z9m6_eXP-a2nbqx9UBS-V53h3tVkqaDoeMVtA%40mail.gmail.com.
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_%2BvZHmcL%2B60T8KZxmE9ZDwJ93RO0zwF8-ON0VxFxxuPjw%40mail.gmail.com.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-develo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAK-PPf2dQ5HEduZqe5G2uHNdEoH-NiB%2BF2h1nEz6sWjdf7pYDg%40mail.gmail.com.
[off topic: Paul, you have a huge open source community as followers on twitter. Can you tweet about it? Maybe we get some additional attention and help-willing people :)]
You received this message because you are subscribed to a topic in the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/google-chrome-developer-tools/5iLsW7T4m90/unsubscribe.
To unsubscribe from this group and all its topics, send an email to google-chrome-develo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_JT9YVqRYUyuPQYd5qDom5PqyLo5rBbwMtQozFgaxQcOw%40mail.gmail.com.
☆PhistucK
☆PhistucK
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_K%2B3ksnSYePOQyWA9EWV2T%3DcwrDJvcxoZ6UrN4ggFK%2BEg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAJZOyhxbsaDSmS2EO8JR0opga7f-tRA5A5iG0TdAfQzjp_iJcg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_JOX-fbG9gCk4DrBtn%2BmW4O%2BWGeT2%2BsHLRhND-GeAbMiQ%40mail.gmail.com.
---alexander
--
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-developer-tools+unsub...@googlegroups.com.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-develo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAK-PPf2dQ5HEduZqe5G2uHNdEoH-NiB%2BF2h1nEz6sWjdf7pYDg%40mail.gmail.com.
tl;dr, my concerns are bug tracking, review process and regression testing.>> When Travis comes back green, a manual reviewer (e.g. me) applies a "ready-for-rietveld" label if the code is reasonable, it looks good, crbug created, etc.We should not fork the bug repository. It should be clear where the bug goes and it is crbug.com today. Filing a bug does not require building Chromium, so it does not contribute to the high contribution barrier.
>> Some automation transforms the PR into a Rietveld issue.>> Subsequent commits on the PR can be sent to form new patch sets.I don't think it is worth it: it is expensive to implement and maintain;
I don't think that Rietveld contributes to the high barrier much either, it is in the cloud.
>> Pull requests kick off Travis CI work to do our devtools presubmit tests (and maybe the Blink ones)This one is the most challenging. DevTools tests are LayoutTest-runner-based. Running them externally against some LKGR build is feasible and is on the critical path to the success of this project. I'd basically start with implementing it and if it succeeds, we can proceed further.
See my comments inline.☆PhistucKOn Thu, Jun 18, 2015 at 7:34 PM, Pavel Feldman <pfel...@chromium.org> wrote:tl;dr, my concerns are bug tracking, review process and regression testing.>> When Travis comes back green, a manual reviewer (e.g. me) applies a "ready-for-rietveld" label if the code is reasonable, it looks good, crbug created, etc.We should not fork the bug repository. It should be clear where the bug goes and it is crbug.com today. Filing a bug does not require building Chromium, so it does not contribute to the high contribution barrier.But no one suggested to fork the bug repository. Where did you get that impression?The label is on a pull request (a GitHub patch), not on an issue (a GitHub bug).Even more, we can explicitly disable the issue section (I just did it - https://github.com/phistuck/chromium-developer-tools-frontend).
>> Some automation transforms the PR into a Rietveld issue.>> Subsequent commits on the PR can be sent to form new patch sets.I don't think it is worth it: it is expensive to implement and maintain;I am willing to implement this automation and maintain it and hopefully, if I do it right (using a public Rietveld API, say), there will basically be nothing to maintain.
I don't think that Rietveld contributes to the high barrier much either, it is in the cloud.Of course it contributes to the high barrier (at least it does for me), it (at least, currently) requires a long git clone/pull (I tried to only clone the devtools folder and it took more than an hour instead of a minute). How can you upload patches without it?
A subtree (as Robert suggested in the other thread) will make it much much shorter, but is it feasible?Besides, it requires depot_tools whereas GitHub does not require anything other than a browser (you can change files and create pull requests (patches) using a web interface, which I have done many times).You can even code using a mobile device. ;)>> Pull requests kick off Travis CI work to do our devtools presubmit tests (and maybe the Blink ones)This one is the most challenging. DevTools tests are LayoutTest-runner-based. Running them externally against some LKGR build is feasible and is on the critical path to the success of this project. I'd basically start with implementing it and if it succeeds, we can proceed further.I never suggested to run layout tests (though it might be feasible), I only suggested to run compile_frontend.py (well, and the presubmit tests, which do not include layout tests, as far as I know).What prevents anyone from currently cloning the repository and uploading a patch without ever running a single layout test?
When I fixed bugs (or implemented a feature) in the Developer Tools when it was still a part of WebKit, I did not touch layout tests, nor was I asked to touch them (that might have been wrong, of course, but this is how it went).
I fixed those bugs and implemented those features because the barrier was fairly low - I just needed SVN. I did the rest manually (I created a patch using SVN and a changelog entry by hand).GitHub lowers the barrier even more, which is always welcome when you want to contribute.If you absolutely require running layout tests (which I am not sure we should, the try bots can run them once the code compiles using Travis CI and is approved in GitHub by Paul and the like and uploaded to Rietveld), I can look into it.
On Thu, Jun 18, 2015 at 8:35 PM, PhistucK <phis...@gmail.com> wrote:See my comments inline.☆PhistucKOn Thu, Jun 18, 2015 at 7:34 PM, Pavel Feldman <pfel...@chromium.org> wrote:tl;dr, my concerns are bug tracking, review process and regression testing.>> When Travis comes back green, a manual reviewer (e.g. me) applies a "ready-for-rietveld" label if the code is reasonable, it looks good, crbug created, etc.We should not fork the bug repository. It should be clear where the bug goes and it is crbug.com today. Filing a bug does not require building Chromium, so it does not contribute to the high contribution barrier.But no one suggested to fork the bug repository. Where did you get that impression?The label is on a pull request (a GitHub patch), not on an issue (a GitHub bug).Even more, we can explicitly disable the issue section (I just did it - https://github.com/phistuck/chromium-developer-tools-frontend).Disabling it sounds good. As for impression, the manual review process implied that the work could have started with no associated crbug and I thought that it would be based on the GitHub bug.
>> Some automation transforms the PR into a Rietveld issue.>> Subsequent commits on the PR can be sent to form new patch sets.I don't think it is worth it: it is expensive to implement and maintain;I am willing to implement this automation and maintain it and hopefully, if I do it right (using a public Rietveld API, say), there will basically be nothing to maintain.You might end up using parts of the API that change, but if you are willing to give it a try, who am I to say no. It is just that you'll need to sync comments and patch sets since they could be linkable, etc.
I don't think that Rietveld contributes to the high barrier much either, it is in the cloud.Of course it contributes to the high barrier (at least it does for me), it (at least, currently) requires a long git clone/pull (I tried to only clone the devtools folder and it took more than an hour instead of a minute). How can you upload patches without it?Not sure if we are talking about the same thing. I meant to reuse the Rietveld against the GitHub mirror. Yes it does require depot_tools, but that does not sound like a high barrier to me comparing to what we have today.
A subtree (as Robert suggested in the other thread) will make it much much shorter, but is it feasible?Besides, it requires depot_tools whereas GitHub does not require anything other than a browser (you can change files and create pull requests (patches) using a web interface, which I have done many times).You can even code using a mobile device. ;)>> Pull requests kick off Travis CI work to do our devtools presubmit tests (and maybe the Blink ones)This one is the most challenging. DevTools tests are LayoutTest-runner-based. Running them externally against some LKGR build is feasible and is on the critical path to the success of this project. I'd basically start with implementing it and if it succeeds, we can proceed further.I never suggested to run layout tests (though it might be feasible), I only suggested to run compile_frontend.py (well, and the presubmit tests, which do not include layout tests, as far as I know).What prevents anyone from currently cloning the repository and uploading a patch without ever running a single layout test?Hm. So lets say I have a patch ready, but one of the tests fails or flakes with it. It can't be landed, so what do I do? I can't fall back to the heavyweight process of building Chromium - it would defeat the purpose of this project. And this happens with every patch, not mentioning that most of the changes should be adding / changing the tests.
When I fixed bugs (or implemented a feature) in the Developer Tools when it was still a part of WebKit, I did not touch layout tests, nor was I asked to touch them (that might have been wrong, of course, but this is how it went).Yes, I introduced the testing harness for WebKit inspector when I joined the project, back in 2009. We've been writing tests since then :)
I fixed those bugs and implemented those features because the barrier was fairly low - I just needed SVN. I did the rest manually (I created a patch using SVN and a changelog entry by hand).GitHub lowers the barrier even more, which is always welcome when you want to contribute.If you absolutely require running layout tests (which I am not sure we should, the try bots can run them once the code compiles using Travis CI and is approved in GitHub by Paul and the like and uploaded to Rietveld), I can look into it.I wonder if my example above makes you change your mind wrt this one.
(Tell me if you prefer to fork the thread, or just fork it ;))Can you share some information regarding running the layout tests for the Developer Tools?What is currently needed for that?content_shell?
Is there a command line flag that can make the test binary take the files of the Developer Tools from a path?
--☆PhistucK
You received this message because you are subscribed to the Google Groups "Google Chrome Developer Tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-chrome-develo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CABc02_L8UDh%2BmAc7Xe8sEw2_ET0%3D0U23sSft51kZpc_TF61Pvw%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-chrome-developer-tools/CAJZOyhw_so-O3Bm7NWL9_FBF%3DGwttuSKy-9%2BZhNuUM%3DuKkowvg%40mail.gmail.com.