576026 - runtests.py.in is an interpolated file and shouldn't be

Status: RESOLVED FIXED

(Testing :: Mochitest, defect)

Description

[Jeff Hammel]

http://mxr.mozilla.org/mozilla-central/source/testing/mochitest/runtests.py.in is interpolated.  However, there are no actual interpolation tokens there!  So it is pointless to interpret.  The only thing that happens is that the comments are wiped and CPU cycles are used.

Comment 1

[(not currently active) Ted Mielczarek]

Just nuking this rule and renaming the file should do the trick:
http://mxr.mozilla.org/mozilla-central/source/testing/mochitest/Makefile.in#119

Comment 2

[Justin Wood (:Callek)]

(In reply to comment #1)
> Just nuking this rule and renaming the file should do the trick:
> http://mxr.mozilla.org/mozilla-central/source/testing/mochitest/Makefile.in#119

Well actually we'll need to NSINSTALL the .js for this, no?

Comment 3

[(not currently active) Ted Mielczarek]

If you meant the .py file, then no, it's already listed in that _SERV_FILES variable up above, which gets installed in a rule in that Makefile.

Comment 4

[Edmund Wong (:ewong)]

Attachment: Changed runtests.py.in to runtests.py and removed lines from Makefile.in

Comment 5

[(not currently active) Ted Mielczarek]

Comment on attachment 463753 [details] [diff] [review]
Changed runtests.py.in to runtests.py and removed lines from Makefile.in

You have an extraneous change in browser/locales in this patch.

The Makefile change looks good, but you need to remove runtests.py from GARBAGE as well.

Comment 6

[Edmund Wong (:ewong)]

Attachment: Changed runtests.py.in to runtests.py and removed lines from Makefile.in (v2)

Removed extraneous cruft from patch.  Removed runtests.py from GARBAGE.
Renamed runtests.py.in to runtests.py.

Comment 7

[Justin Wood (:Callek)]

Hmm, does this need approval; it looks like the approval-flags never hit this product/component

Comment 8

[Benjamin Smedberg]

a=me

Comment 9

[[e-mail]]

http://hg.mozilla.org/mozilla-central/rev/2d57f5901aa2

In the future, please follow these suggestions for using checkin-needed:
http://blog.bonardo.net/2010/06/22/so-youre-about-to-use-checkin-needed

It's a pain to have to manually set the user and checkin comment when landing
your patch.

Comment 10

[(not currently active) Ted Mielczarek]

Sounds like this broke running mochitest from the objdir on systems with symlinks, in a similar manner to bug 525047 comment 28. (Python uses the target of the symlink as the directory to use in sys.path for loading modules.)

Comment 11

[Julian Seward [:jseward]]

This causes mochitests to fail at the start as shown below, on
x86_64-Linux (Ubuntu 10.04).  Backing it out by hand gives a working
mochitest.

Could this please be either fixed or backed out?  Right now isn't a
good time for mochitests to be broken.


$ DISPLAY=:3.0 make -C ff-opt mochitest-plain
make: Entering directory `/space2/sewardj/MOZ/MC-23-08-2010/ff-opt'
rm -f ./mochitest-plain.log && /usr/bin/python2.6 _tests/testing/mochitest/runtests.py --autorun --close-when-done --console-level=INFO --log-file=./mochitest-plain.log --file-level=INFO --symbols-path=./dist/crashreporter-symbols  
Traceback (most recent call last):
  File "_tests/testing/mochitest/runtests.py", line 55, in <module>
    from automation import Automation
ImportError: No module named automation
make: *** [mochitest-plain] Error 1
make: Leaving directory `/space2/sewardj/MOZ/MC-23-08-2010/ff-opt'

Comment 12

[Julian Seward [:jseward]]

I wonder if this (hangs, focus weirdness) is related to bug 583699.
That concerns the exact same test case, and Valgrind found an
uninitialised value use in it.

Comment 13

[Julian Seward [:jseward]]

(In reply to comment #12)
Darn.  Wrong bug.  Ignore.  Sorry.

Comment 14

[Justin Lebar (not reading bugmail)]

I'm going to back this out in a moment.

Comment 15

[Justin Lebar (not reading bugmail)]

Backed out: http://hg.mozilla.org/mozilla-central/rev/cbb5793c529e

Comment 16

[Edmund Wong (:ewong)]

Attachment: Renamed runtests.py.in to runtests.py and removed interpolated lines from Makefile.in

Comment 17

[Karl Tomlinson (:karlt)]

http://hg.mozilla.org/mozilla-central/rev/52c2f34cd686