[Scons-dev] Status of QMTest framework support

[Gary Oberbrunner]

On Mon, Dec 17, 2012 at 10:18 AM, anatoly techtonik <techtonik at gmail.com>wrote:


> On Mon, Dec 17, 2012 at 4:39 PM, Gary Oberbrunner <garyo at oberbrunner.com>wrote:

>

>>

>> On Mon, Dec 17, 2012 at 8:03 AM, anatoly techtonik <techtonik at gmail.com>wrote:

>>

>>> Hi,

>>>

>>> For a long time I thought that QMTest is bundled with SCons. Now I see

>>> that it is some external framework, which downloads are no longer

>>> available. So, the questions are:

>>>

>>> 1. What the code in QMTest directory is for?

>>> 2. Are there pieces from GPL'ed QMTest framework?

>>> 3. What things are we still missing from it?

>>> http://www.scons.org/wiki/SConsTestingRevisions

>>>

>>> And what is this AegisBatchStream and "test result stream" in general?

>>>

>>> Aegis was used originally by Stephen for packaging I believe.  It hasn't

>> been used in many years.  It should be flushed; see

>> http://comments.gmane.org/gmane.comp.programming.tools.scons.devel/10707for instance.

>>

>> QMTest was thought to be a good test framework for Python programs; we

>> got pretty deeply into it for a while, but eventually discovered we had

>> "extended" it so far we weren't really using QMTest itself hardly at all.

>>  The tests *might* still run under QMTest but I don't think anyone cares

>> anymore.

>>

>

> So, what about this directory? Should it be renamed or purged?

>


It's got the whole test framework in it (TestCmd, TestSCons etc.) so it's
important -- it should be renamed.  Maybe "test-framework"?



>

> And btw, I just added -jN to runtest.py, so per the above wiki page, we

>> can now run the tests in parallel (with no QMTest).  With -j10 on my

>> average dual-core laptop, the tests take about 15 minutes now, and the test

>> results seem OK.  And we also have test timing and time reporting now, as

>> well as out-of-core tool testing and test fixtures.  So at this point I

>> think there's no further reason to consider QMTest.  I'll mark that wiki

>> page as pretty much all complete now.

>>

>

> I guess the next bottleneck is to speed up disk operations. Something to

> dig about ramdisks in user space.  At least Linux looks like a capable OS.

>


Sure, why not!  I'm much happier already, so will probably move into other
areas, but any test speedups are always welcome.  Another possibility would
be to look at the slowest tests (use -t) and see why they are slow.  They
may be doing unnecessary work.

-- 
Gary
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://two.pairlist.net/pipermail/scons-dev/attachments/20121217/3739c85e/attachment.html>