69089 2011-09-29 09:40:11 -0700 EWS bots shouldn't die if a buggy patch with r? modifies the source code of the EWS 2019-09-03 12:39:59 -0700 1 1 1 Unclassified WebKit Tools / Tests 528+ (Nightly build) All All RESOLVED FIXED InRadar P2 Normal --- 1 ossy webkit-unassigned abarth ap eric ews-watchlist galpeter levin ossy webkit-bug-importer oldest_to_newest 475465 0 ossy 2011-09-29 09:40:11 -0700 https://bugs.webkit.org/attachment.cgi?id=109089 in https://bugs.webkit.org/show_bug.cgi?id=68973 killed all EWS bots, because the proposed patch updated the code of the EWS, and it was incorrect. I think we should find a way to protect EWS from buggy python patches in bugzilla with r? . error log: Failed to run "['/storage/WebKit-qt-ews/Tools/Scripts/webkit-patch', '--status-host=queues.webkit.org', '--bot-i..." exit_code: 1 Traceback (most recent call last): File "/storage/WebKit-qt-ews/Tools/Scripts/webkit-patch", line 74, in <module> main() File "/storage/WebKit-qt-ews/Tools/Scripts/webkit-patch", line 67, in main from webkitpy.tool.main import WebKitPatch File "/storage/WebKit-qt-ews/Tools/Scripts/webkitpy/tool/main.py", line 39, in <module> from webkitpy.tool import commands File "/storage/WebKit-qt-ews/Tools/Scripts/webkitpy/tool/commands/__init__.py", line 6, in <module> from webkitpy.tool.commands.download import * File "/storage/WebKit-qt-ews/Tools/Scripts/webkitpy/tool/commands/download.py", line 32, in <module> from webkitpy.tool import steps File "/storage/WebKit-qt-ews/Tools/Scripts/webkitpy/tool/steps/__init__.py", line 32, in <module> from webkitpy.tool.steps.applywatchlisttopatch import ApplyWatchListToPatch ImportError: No module named applywatchlisttopatch 475482 1 abarth 2011-09-29 09:57:30 -0700 Any thoughts on how to accomplish this? 475486 2 levin 2011-09-29 10:01:02 -0700 (In reply to comment #1) > Any thoughts on how to accomplish this? Perhaps the ews could run from a different enlistment that the enlistment where it applies the patch, but still sync its own enlistment every so often. Or perhaps the ews could copy itself somewhere else and run from there (and update that location -- once in a while). 475491 3 abarth 2011-09-29 10:04:31 -0700 > Perhaps the ews could run from a different enlistment that the enlistment where it applies the patch, but still sync its own enlistment every so often. ^^^ That sounds like a good approach. 1567466 4 ews-watchlist 2019-09-03 12:39:22 -0700 We have re-written EWS, and with buildbot-based EWS, this isn't an issue anymore. 1567467 5 webkit-bug-importer 2019-09-03 12:39:59 -0700 <rdar://problem/54986401>