Patchwork test-patchbomb.t failing

login
register
mail settings
Submitter Greg Ward
Date April 25, 2014, 1:34 a.m.
Message ID <20140425013425.GB28877@gerg.ca>
Download mbox | patch
Permalink /patch/4437/
State Deferred
Headers show

Comments

Greg Ward - April 25, 2014, 1:34 a.m.
Hey all --

I'm getting a test failure from test-patchbomb.t. It's new in that I
haven't seen it before, and I ran the tests with 3.0-rc-ish a couple
of days ago on a different machine. But OTOH it's not new; I can
reproduce it with 2.9.2 and 2.5.4. So presumably something is funny on
my box that test-patchbomb.t doesn't expect. This is a very vanilla
PC: Ubuntu 14.04, 64-bit Intel.

Hmmmmm. I just upgraded to Ubuntu 14.04 the other day. Must try this
on my Fedora 20 laptop.

Anyways, here is the only interesting part of the failure:

"""
"""

Does this ring any bells?

        Greg
Mads Kiilerich - April 25, 2014, 11:03 a.m.
On 04/25/2014 03:34 AM, Greg Ward wrote:
> Hey all --
>
> I'm getting a test failure from test-patchbomb.t. It's new in that I
> haven't seen it before, and I ran the tests with 3.0-rc-ish a couple
> of days ago on a different machine. But OTOH it's not new; I can
> reproduce it with 2.9.2 and 2.5.4. So presumably something is funny on
> my box that test-patchbomb.t doesn't expect. This is a very vanilla
> PC: Ubuntu 14.04, 64-bit Intel.
>
> Hmmmmm. I just upgraded to Ubuntu 14.04 the other day. Must try this
> on my Fedora 20 laptop.
>
> Anyways, here is the only interesting part of the failure:
>
> """
> --- /home/greg/src/hg-stable/tests/test-patchbomb.t
> +++ /home/greg/src/hg-stable/tests/test-patchbomb.t.err
> @@ -233,6 +233,7 @@
>     sgVa2F19D1tx9+hgbAygLgQwaIqcDdgA4BjQgIiz/AEP72++llgDKhKducqodGE4B0ETqF3JFOFC
>     Q70eyNw=
>     --===*-- (glob)
> +
>   
>   utf-8 patch:
>     $ python -c 'fp = open("utf", "wb"); fp.write("h\xC3\xB6mma!\n"); fp.close();'
> @@ -683,6 +684,7 @@
>     +c
>     
>     --===*-- (glob)
> +
>   
>   
>   test inline for single patch (quoted-printable):
> """
>
> Does this ring any bells?

http://bz.selenic.com/show_bug.cgi?id=4188

-Mads
Matt Mackall - April 29, 2014, 8 p.m.
On Thu, 2014-04-24 at 21:34 -0400, Greg Ward wrote:
> Hey all --
> 
> I'm getting a test failure from test-patchbomb.t. It's new in that I
> haven't seen it before, and I ran the tests with 3.0-rc-ish a couple
> of days ago on a different machine. But OTOH it's not new; I can
> reproduce it with 2.9.2 and 2.5.4. So presumably something is funny on
> my box that test-patchbomb.t doesn't expect. This is a very vanilla
> PC: Ubuntu 14.04, 64-bit Intel.
> 
> Hmmmmm. I just upgraded to Ubuntu 14.04 the other day. Must try this
> on my Fedora 20 laptop.
> 
> Anyways, here is the only interesting part of the failure:
> 
> """
> --- /home/greg/src/hg-stable/tests/test-patchbomb.t
> +++ /home/greg/src/hg-stable/tests/test-patchbomb.t.err
> @@ -233,6 +233,7 @@
>    sgVa2F19D1tx9+hgbAygLgQwaIqcDdgA4BjQgIiz/AEP72++llgDKhKducqodGE4B0ETqF3JFOFC
>    Q70eyNw=
>    --===*-- (glob)
> +  
>  
>  utf-8 patch:
>    $ python -c 'fp = open("utf", "wb"); fp.write("h\xC3\xB6mma!\n"); fp.close();'
> @@ -683,6 +684,7 @@
>    +c
>    
>    --===*-- (glob)
> +  
>  
> 
>  test inline for single patch (quoted-printable):
> """
> 
> Does this ring any bells?

Unfortunately, we don't have a good way to make the test suite ignore
single lines yet.
Mads Kiilerich - April 29, 2014, 9:54 p.m.
On 04/29/2014 10:00 PM, Matt Mackall wrote:
> On Thu, 2014-04-24 at 21:34 -0400, Greg Ward wrote:
>> Hey all --
>>
>> I'm getting a test failure from test-patchbomb.t. It's new in that I
>> haven't seen it before, and I ran the tests with 3.0-rc-ish a couple
>> of days ago on a different machine. But OTOH it's not new; I can
>> reproduce it with 2.9.2 and 2.5.4. So presumably something is funny on
>> my box that test-patchbomb.t doesn't expect. This is a very vanilla
>> PC: Ubuntu 14.04, 64-bit Intel.
>>
>> Hmmmmm. I just upgraded to Ubuntu 14.04 the other day. Must try this
>> on my Fedora 20 laptop.
>>
>> Anyways, here is the only interesting part of the failure:
>>
>> """
>> --- /home/greg/src/hg-stable/tests/test-patchbomb.t
>> +++ /home/greg/src/hg-stable/tests/test-patchbomb.t.err
>> @@ -233,6 +233,7 @@
>>     sgVa2F19D1tx9+hgbAygLgQwaIqcDdgA4BjQgIiz/AEP72++llgDKhKducqodGE4B0ETqF3JFOFC
>>     Q70eyNw=
>>     --===*-- (glob)
>> +
>>   
>>   utf-8 patch:
>>     $ python -c 'fp = open("utf", "wb"); fp.write("h\xC3\xB6mma!\n"); fp.close();'
>> @@ -683,6 +684,7 @@
>>     +c
>>     
>>     --===*-- (glob)
>> +
>>   
>>
>>   test inline for single patch (quoted-printable):
>> """
>>
>> Does this ring any bells?

[ to give context: aka http://bz.selenic.com/show_bug.cgi?id=4188 ]

> Unfortunately, we don't have a good way to make the test suite ignore
> single lines yet.

I guess one of the "simplest" ways to solve this would be to pipe it 
through a careful selection expressions like   sed -e '117{/^$/d}' .

/Mads

Patch

--- /home/greg/src/hg-stable/tests/test-patchbomb.t
+++ /home/greg/src/hg-stable/tests/test-patchbomb.t.err
@@ -233,6 +233,7 @@ 
   sgVa2F19D1tx9+hgbAygLgQwaIqcDdgA4BjQgIiz/AEP72++llgDKhKducqodGE4B0ETqF3JFOFC
   Q70eyNw=
   --===*-- (glob)
+  
 
 utf-8 patch:
   $ python -c 'fp = open("utf", "wb"); fp.write("h\xC3\xB6mma!\n"); fp.close();'
@@ -683,6 +684,7 @@ 
   +c
   
   --===*-- (glob)
+  
 
 
 test inline for single patch (quoted-printable):