Hi,
this is an old issue
(https://code.google.com/p/ndt/issues/detail?id=36), which we should
probably fix right after the release.
Regards,
Jakub.
On 01/28/2015 08:18 PM, Aaron Brown wrote:
> Hey Jordan,
>
> Unhelpfully, the files are compressed using DEFLATE not gzip
(.gz file extension notwithstanding). There may be better ways to
decompress them, but this seems to work:
>
> openssl zlib -d < s2c.ndttrace.gz > s2c.ndttrace
>
> Cheers,
> Aaron
>
>> On Jan 28, 2015, at 2:01 PM, Jordan McCarthy
wrote:
>>
Sure - I've uploaded a couple of samples to:
http://files.opentechinstitute.org/~mccarthy/c2s.ndttrace.gz
http://files.opentechinstitute.org/~mccarthy/s2c.ndttrace.gz
Jordan McCarthy
Open Technology Institute | New America
Public Key: 0xC08D8042 | 4A61 3D39 4125 127D 65EA DDC2 BFBD A2E9
C08D 8042
On 01/28/2015 01:57 PM, Aaron Brown wrote:
>>> Hey Jordan,
>>>
>>> I’m not sure on the ndttrace stuff. Could you send
one?
>>>
>>> Cheers, Aaron
>>>
>>>> On Jan 28, 2015, at 1:51 PM, Jordan McCarthy
>>>> wrote:
>>>>
>>> Hi Aaron, Ah yes - I remember seeing that thread
now. Thanks for
>>> the explanation, and sorry for the noise!
>>>
>>> One other observation, probably also an artifact of
me having
>>> missed a thread somewhere: although the ndttrace
files are back,
>>> they consistently refuse to be decompressed, claiming
to be "not in
>>> gzip format." Am I doing something wrong here?
>>>
>>> (M-Lab is in somewhat urgent need of recent NDT trace
files, so
>>> this issue is weirdly important to us right now).
>>>
>>> Thanks, Jordan
>>>
>>>
>>> Jordan McCarthy Open Technology Institute | New
America Public Key:
>>> 0xC08D8042 | 4A61 3D39 4125 127D 65EA DDC2 BFBD A2E9
C08D 8042
>>>
>>> On 01/28/2015 01:14 PM, Aaron Brown wrote:
>>>>>> Hey Jordan,
>>>>>>
>>>>>> The goal is to move away from fakewww and
start serving the
>>>>>> NDT client using Apache. In furtherance
of that, fakewww is
>>>>>> optional, and isn’t compiled by default.
You need to add
>>>>>> “—enable-fakewww” to get it back.
>>>>>>
>>>>>> Cheers, Aaron
>>>>>>
>>>>>>> On Jan 28, 2015, at 12:11 PM, Jordan
McCarthy
>>>>>>>
wrote:
>>>>>>>
>>>>>> I've also compiled the server - but this
time I used the
>>>>>> standard M-Lab build system to do it, and
successfully run
>>>>>> tests against the test server from both
the CLI and the Flash
>>>>>> client.
>>>>>>
>>>>>> The one glaring problem I'm having is
that the fakewww binary
>>>>>> seems to have completely disappeared.
Has this thing been
>>>>>> renamed or moved, by chance?
>>>>>>
>>>>>> Thanks, Jordan
>>>>>>
>>>>>> Jordan McCarthy Open Technology Institute
| New America
>>>>>> Public Key: 0xC08D8042 | 4A61 3D39 4125
127D 65EA DDC2 BFBD
>>>>>> A2E9 C08D 8042
>>>>>>
>>>>>> On 01/27/2015 01:52 PM, Aaron Brown
wrote:
>>>>>>>>> Excellent, thanks for the
report Peter.
>>>>>>>>>
>>>>>>>>> Cheers, Aaron
>>>>>>>>>
>>>>>>>>>> On Jan 27, 2015, at 11:50
AM, Peter Boothe
>>>>>>>>>> <
>
>>>>>>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>> I can only compile the
server on the same system I
>>>>>>>>>> have been writing
websockets and unit tests
>>>>>>>>>> (different one from the
Ubuntu one I mentioned
>>>>>>>>>> earlier).
>>>>>>>>>>
>>>>>>>>>> The tarball compiles and
the server runs just fine on
>>>>>>>>>> that system and also
passes all the "make check"
>>>>>>>>>> tests. (Linux w/ 2.6
kernel and web100 patches).
>>>>>>>>>>
>>>>>>>>>> -Peter
>>>>>>>>>>
>>>>>>>>>> On Tue, Jan 27, 2015 at
10:37 AM, Aaron Brown
>>>>>>>>>> <
>
>>>>>>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>> Hey Peter,
>>>>>>>>>>
>>>>>>>>>> Thanks for the update.
Does the server work when
>>>>>>>>>> compiled?
>>>>>>>>>>
>>>>>>>>>> Cheers, Aaron
>>>>>>>>>>
>>>>>>>>>>> On Jan 26, 2015, at
10:19 AM, Peter Boothe
>>>>>>>>>>>
< >
>>>>>>>>>>> wrote:
>>>>>>>>>>>
>>>>>>>>>>> The tarball
configured fine and then make worked,
>>>>>>>>>>> but it did not make
web100clt. I cd'd into src and
>>>>>>>>>>> then web100clt made
correctly and was able to run
>>>>>>>>>>> an NDT test to
>>>>>>>>>>>
ndt.iupui.mlab1.lga01.measurement-lab.org
>>>>>>>>>>>
<http://ndt.iupui.mlab1.lga01.measurement-lab.org/>
>>>>>>>>>>> and the resulting
data looked correct.
>>>>>>>>>>>
>>>>>>>>>>> -Peter
>>>>>>>>>>>
>>>>>>>>>>> On Thu, Jan 22, 2015
at 3:36 PM, Aaron Brown
>>>>>>>>>>>
< >
>>>>>>>>>>> wrote:
>>>>>>>>>>>
>>>>>>>>>>> Hey Folks,
>>>>>>>>>>>
>>>>>>>>>>> I’ve updated the NDT
RPMs/tarball. You can grab
>>>>>>>>>>> them at
>>>>>>>>>>>
http://ndb1.internet2.edu/~aaron/ndt-3.7.0-2.el6.x86_64/
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
This should fix the mxmlc error that Peter encountered,
>>>>>>>>>>> along with the issue
that prevented the applet
>>>>>>>>>>> from working with
1.8. Would folks be able to test
>>>>>>>>>>> these out?
>>>>>>>>>>>
>>>>>>>>>>> It’d be good if we
could get the tarball, x86-64
>>>>>>>>>>> and i386 RPMs tested,
using both flash and java
>>>>>>>>>>> clients on OS X,
Windows and Linux, if possible. It
>>>>>>>>>>> doesn’t need to be in
depth testing, just make sure
>>>>>>>>>>> that there’s nothing
anything obviously wrong with
>>>>>>>>>>> the configuration
when you run some tests.
>>>>>>>>>>>
>>>>>>>>>>> Cheers, Aaron
>>>>>>>>>>>
>>>>>>>>>>>> On Jan 9, 2015,
at 12:15 PM, Aaron Brown
>>>>>>>>>>>>
<
>>>>>>>>>>>>
> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>> Hey Folks,
>>>>>>>>>>>>
>>>>>>>>>>>> I’ve merged in
the RPM changes from Issue 162,
>>>>>>>>>>>> fixed a couple
issues and rebuilt the precompiled
>>>>>>>>>>>> flash and java
clients. The tarball and RPMs from
>>>>>>>>>>>> trunk can be
gotten at
>>>>>>>>>>>>
http://ndb1.internet2.edu/~aaron/ndt-3.7.0-1.el6.x86_64/
>>>>>>>>>>>>
>>>>>>>>>>>>
. Could folks test both, and make sure they work? My
>>>>>>>>>>>> scant testing of
the x86-64 RPM seemed to work.
>>>>>>>>>>>>
>>>>>>>>>>>> Cheers, Aaron
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> -- Peter Boothe |
Software Engineer |
>>>>>>>>>>> ɯoɔ˙ǝlƃooƃ@ǝɥʇooqd
| +1
>>>>>>>>>>> 650-889-0879
<tel:650-889-0879> (corp cell)
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> -- Peter Boothe |
Software Engineer |
>>>>>>>>>> ɯoɔ˙ǝlƃooƃ@ǝɥʇooqd
| +1
>>>>>>>>>> 650-889-0879 (corp cell)
>>>>>>>>>
>>>>>>>
>>>>>>
>>>>
>>>
>>
>
|
