whiteknight/io_cleanup1 branch ready for testing

James E Keenan jkeen at verizon.net
Wed Aug 15 02:48:31 UTC 2012


On 8/14/12 5:09 PM, Reini Urban wrote:
> On Tue, Jul 24, 2012 at 8:01 PM, James E Keenan<jkeen at verizon.net>  wrote:
>> On 6/25/12 6:27 PM, James E Keenan wrote:
>>> [snip]
>> On linux/i386, in the io_cleanup1 branch 'make' built successfully with all
>> gcc -- but is still failing tests 14 and 28 in t/pmc/filehandle.t -- even
>> though (as reported in separate post) 'make test' is PASSing on Darwin/i386.
>>
>> http://smolder.parrot.org/app/projects/report_details/30272
>
> I believe io_cleanup1 is now ready to be merged into master.
> There were some problems with /tmp as a swap filesystem and
> interestings semantics outlined in https://github.com/parrot/parrot/issues/811
> but this was an test artefact and is now resolved.
>
> If I remember moritz also gave his okay for rakudo already.

Assuming that test 28 in t/pmc/filehandle.t remains TODO-ed, then I 
agree that the whiteknight/io_cleanup1 branch is ready to be merged into 
master.

Test 28 in t/pmc/filehandle.t is that weird test which, as I've reported 
several times, gives me this output under 'prove -v':

#####
not ok 28 - write after buffered read # TODO GH #811 Write error: No 
space left on device on swap fs

#   Failed (TODO) test 'write after buffered read'
#   at t/pmc/filehandle.t line 927.
# Exited with error code: 1
# Received:
# 10
# Write error: No space left on device
# current instr.: 'test' pc 69 
(/Users/jimk/gitwork/parrot/t/pmc/filehandle_28.pir:18)
#
# Expected:
# 10
# abcde#####klmno
#
#####

It also causes a popup warning on my iBook saying that my startup disk 
is full -- a condition which, if true, is transitory.

It would be nice to get a better diagnosis of the problem with test 28, 
but since the file is completing with that test TODOed, I think we can 
proceed to a merge. See: 
http://smolder.parrot.org/app/projects/report_details/30607

The whiteknight/io_cleanup1 branch is PASSing for me now with the 
all-gcc build on linux/i386.  It also PASSed make fulltest.  And, given 
what I observed over this past weekend, I decided to create a temporary 
branch in which the io_cleanup1 branch was merged into master.  That 
branch PASSed a simple all-gcc build on linux/i386.

Thank you very much.
Jim Keenan



More information about the parrot-dev mailing list