|
|
Created:
4 years, 11 months ago by hbos Modified:
4 years, 11 months ago Reviewers:
kjellander_webrtc CC:
webrtc-reviews_webrtc.org, tterriberry_mozilla.com, qiang.lu, niklas.enbom, yujie_mao (webrtc), peah-webrtc Base URL:
https://chromium.googlesource.com/external/webrtc.git@master Target Ref:
refs/pending/heads/master Project:
webrtc Visibility:
Public. |
Descriptionsetup_links.py fix so that FFmpeg compiles on windows.
Fix creates a symlink tools/generate_stubs -> chromium/src/tools/generate_stubs which is used by ffmpeg on windows.
BUG=468365
NOTRY=True
Committed: https://crrev.com/5602f6565905dea71064f017904a1754f94193ed
Cr-Commit-Position: refs/heads/master@{#11264}
Patch Set 1 : setup_links.py fix, build ffmpeg on all platforms except on iOS #Patch Set 2 : Force gclient sync #Patch Set 3 : Do not always build ffmpeg, revert the sync_chromium.py change which didn't matter #
Messages
Total messages: 19 (6 generated)
Description was changed from ========== setup_links.py fix so that FFmpeg compiles on WIN BUG=468365 ========== to ========== setup_links.py fix so that FFmpeg compiles on windows. Fix creates a symlink tools/generate_stubs -> chromium/src/tools/generate_stubs which is used by ffmpeg on windows. BUG=468365 ==========
hbos@webrtc.org changed reviewers: + kjellander@webrtc.org
This is really strange, PS1: First time I ran win_x64_clang_dbg it failed and second time it succeeded. gclient runhooks says it's created the tools/generate_stubs so what gives with the compile error about the path? Will forcing gclient sync help I wonder... It shouldn't be necessary.
On 2016/01/14 17:39:10, hbos wrote: > This is really strange, PS1: First time I ran win_x64_clang_dbg it failed and > second time it succeeded. gclient runhooks says it's created the > tools/generate_stubs so what gives with the compile error about the path? > > Will forcing gclient sync help I wonder... It shouldn't be necessary. Nope... :/ How can compiling be flakey
Yo kjellander, I think I'm experiencing flakey compiling - wtf? Or am I going mad...
On 2016/01/14 17:57:38, hbos wrote: > Yo kjellander, I think I'm experiencing flakey compiling - wtf? > Or am I going mad... Hm, I haven't seen such problems before. Parallel compilation can fail when a target that depends on something else is built before it's dependency (if the dependencies are not correctly setup). I've seen that on Chromium bots sometimes, but in this case we're creating the dependency (tools/generate_stubs symlink) in the runhooks step, so it cannot be the case here. Since it went green when I retried it, I suggest just commit this and we can investigate further if it keeps reoccurring. Once the link is created it's possible the error will go away (and your green-red-green results are probably caused by different machines picking up the tryjobs each time). The CL is now doing more than the description says - it's building ffmpeg always on non-iOS. I don't think we want that, since it's only needed when building with use_openh264 and use_third_party_h264 is 1, right?
On 2016/01/14 18:35:23, kjellander (webrtc) wrote: > On 2016/01/14 17:57:38, hbos wrote: > > Yo kjellander, I think I'm experiencing flakey compiling - wtf? > > Or am I going mad... > > Hm, I haven't seen such problems before. Parallel compilation can fail when a > target that depends on something else is built before it's dependency (if the > dependencies are not correctly setup). I've seen that on Chromium bots > sometimes, but in this case we're creating the dependency (tools/generate_stubs > symlink) in the runhooks step, so it cannot be the case here. > > Since it went green when I retried it, I suggest just commit this and we can > investigate further if it keeps reoccurring. Once the link is created it's > possible the error will go away (and your green-red-green results are probably > caused by different machines picking up the tryjobs each time). > > The CL is now doing more than the description says - it's building ffmpeg always > on non-iOS. I don't think we want that, since it's only needed when building > with use_openh264 and use_third_party_h264 is 1, right? Yes, I wanted to make sure trybots are green on all supported platforms after tools/generate_stubs fix. I'll remove "always compile" before landing.
PTAL. I can git cl land this without running bots?
Description was changed from ========== setup_links.py fix so that FFmpeg compiles on windows. Fix creates a symlink tools/generate_stubs -> chromium/src/tools/generate_stubs which is used by ffmpeg on windows. BUG=468365 ========== to ========== setup_links.py fix so that FFmpeg compiles on windows. Fix creates a symlink tools/generate_stubs -> chromium/src/tools/generate_stubs which is used by ffmpeg on windows. BUG=468365 NOTRY=True ==========
On 2016/01/15 09:05:32, hbos wrote: > On 2016/01/14 18:35:23, kjellander (webrtc) wrote: > > On 2016/01/14 17:57:38, hbos wrote: > > > Yo kjellander, I think I'm experiencing flakey compiling - wtf? > > > Or am I going mad... > > > > Hm, I haven't seen such problems before. Parallel compilation can fail when a > > target that depends on something else is built before it's dependency (if the > > dependencies are not correctly setup). I've seen that on Chromium bots > > sometimes, but in this case we're creating the dependency > (tools/generate_stubs > > symlink) in the runhooks step, so it cannot be the case here. > > > > Since it went green when I retried it, I suggest just commit this and we can > > investigate further if it keeps reoccurring. Once the link is created it's > > possible the error will go away (and your green-red-green results are probably > > caused by different machines picking up the tryjobs each time). > > > > The CL is now doing more than the description says - it's building ffmpeg > always > > on non-iOS. I don't think we want that, since it's only needed when building > > with use_openh264 and use_third_party_h264 is 1, right? > > Yes, I wanted to make sure trybots are green on all supported platforms after > tools/generate_stubs fix. I'll remove "always compile" before landing. Ah, I see now. I added NOTRY=True to the CL, so you can land it with the CQ skipping the trybots.
On 2016/01/15 09:11:31, hbos wrote: > PTAL. I can git cl land this without running bots? Yes, that's the other way of skipping the CQ bots.
On 2016/01/15 09:15:27, kjellander (webrtc) wrote: > On 2016/01/15 09:11:31, hbos wrote: > > PTAL. I can git cl land this without running bots? > > Yes, that's the other way of skipping the CQ bots. Still need LGTM from you, right?
On 2016/01/15 09:22:26, hbos wrote: > On 2016/01/15 09:15:27, kjellander (webrtc) wrote: > > On 2016/01/15 09:11:31, hbos wrote: > > > PTAL. I can git cl land this without running bots? > > > > Yes, that's the other way of skipping the CQ bots. > > Still need LGTM from you, right? Right, that detail... lgtm ;)
The CQ bit was checked by hbos@webrtc.org
CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/patch-status/1586083003/40001 View timeline at https://chromium-cq-status.appspot.com/patch-timeline/1586083003/40001
Message was sent while issue was closed.
Description was changed from ========== setup_links.py fix so that FFmpeg compiles on windows. Fix creates a symlink tools/generate_stubs -> chromium/src/tools/generate_stubs which is used by ffmpeg on windows. BUG=468365 NOTRY=True ========== to ========== setup_links.py fix so that FFmpeg compiles on windows. Fix creates a symlink tools/generate_stubs -> chromium/src/tools/generate_stubs which is used by ffmpeg on windows. BUG=468365 NOTRY=True ==========
Message was sent while issue was closed.
Committed patchset #3 (id:40001)
Message was sent while issue was closed.
Description was changed from ========== setup_links.py fix so that FFmpeg compiles on windows. Fix creates a symlink tools/generate_stubs -> chromium/src/tools/generate_stubs which is used by ffmpeg on windows. BUG=468365 NOTRY=True ========== to ========== setup_links.py fix so that FFmpeg compiles on windows. Fix creates a symlink tools/generate_stubs -> chromium/src/tools/generate_stubs which is used by ffmpeg on windows. BUG=468365 NOTRY=True Committed: https://crrev.com/5602f6565905dea71064f017904a1754f94193ed Cr-Commit-Position: refs/heads/master@{#11264} ==========
Message was sent while issue was closed.
Patchset 3 (id:??) landed as https://crrev.com/5602f6565905dea71064f017904a1754f94193ed Cr-Commit-Position: refs/heads/master@{#11264} |