Open Bug 850040 Opened 11 years ago Updated 2 years ago

make -C jstestbrowser work

Categories

(Testing :: Reftest, defect)

defect

Tracking

(Not tracked)

People

(Reporter: bc, Unassigned)

Details

Attachments

(1 file)

Attached patch patchSplinter Review
decoder ran into this error this morning trying to run the jsreftests from the command line:

make -C firefox-debug jstestbrowser

rm -f ./jstestbrowser.log && /mozilla/builds/nightly/mozilla/firefox-debug/_virtualenv/bin/python _tests/reftest/runreftest.py --extra-profile-file=./dist/plugins --symbols-path=./dist/crashreporter-symbols  "./dist/test-package-stage/jsreftest/tests//jstests.list --extra-profile-file=./dist/test-package-stage/jsreftest/tests/user.js" | tee ./jstestbrowser.log

...

REFTEST TEST-UNEXPECTED-FAIL | | EXCEPTION: [Exception... "Component returned failure code: 0x80520012 (NS_ERROR_FILE_NOT_FOUND) [nsIChannel.open]"  nsresult: "0x80520012 (NS_ERROR_FILE_NOT_FOUND)"  location: "JS frame :: chrome://reftest/content/reftest.js :: ReadManifest :: line 684"  data: no]
make: *** [jstestbrowser] Error 1

This is after package-tests and the file dist/test-package-stage/jsreftest/tests//jstests.list does exist.

Looking at testing/testsuite-targets.mk and the tbpl logs for jsreftests, 

jstestbrowser: TESTS_PATH?=test-package-stage/jsreftest/tests/
jstestbrowser:
        $(MAKE) -C $(DEPTH)/config
        $(MAKE) -C $(DEPTH)/js/src/config
        $(MAKE) stage-jstests
        $(call RUN_REFTEST,$(DIST)/$(TESTS_PATH)/jstests.list --extra-profile-file=$(DIST)/test-package-stage/jsreftest/tests/user.js)
        $(CHECK_TEST_ERROR)

the $(1) argument ends up including both the manifest and the --extra-profile-file... in a single argument as in:

 "./dist/test-package-stage/jsreftest/tests//jstests.list --extra-profile-file=./dist/test-package-stage/jsreftest/tests/user.js"

To fix this we can add an additional argument to RUN_REFTEST so that the manifest and additional arguments are treated separately.

This patch works for me locally, but I am not sure this is the best approach. 

Also, bug 615386 added similar calls for reftest-ipc-gpu, crashtest-ipc, etc.

cjones: do make reftest-ipc-gpu etc work for you?

should this patch include the other RUN_REFTEST calls which have additional options specified?
Attachment #723701 - Flags: feedback?(ted)
Flags: needinfo?(cjones.bugs)
Attachment #723701 - Flags: feedback?(ted) → feedback+
I haven't tested reftest-ipc-gpu for a long time, but AFAIK nothing has changed to break it.  Not sure if that helps.
Flags: needinfo?(cjones.bugs)
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: