cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Joshua McKenzie (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-9795) Fix cqlsh dtests on windows
Date Wed, 15 Jul 2015 17:49:04 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-9795?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14628437#comment-14628437
] 

Joshua McKenzie commented on CASSANDRA-9795:
--------------------------------------------

Haven't reviewed yet since I have other things active, but I ran tests locally and we're at
25/26 passing on cqlsh_copy_tests.py and 34/38 on cqlsh_tests.py. I'd like to see a run of
dtests on 2.1 w/these 3 branches on linux to confirm that there's no regression there; given
the current state of Windows dtests I'm ok being a bit more cavalier with committing on that
side w/out a full dtest run as we're still stabilizing it.

[~mshuler]: Would it be possible for us to get a dtest run w/the linked C*, ccm, and dtest
branches on linux to confirm no regression?

I'll get started on review of this shortly as I'd prefer to get this in as close to the 2.2.0
release as possible.

> Fix cqlsh dtests on windows
> ---------------------------
>
>                 Key: CASSANDRA-9795
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9795
>             Project: Cassandra
>          Issue Type: Sub-task
>            Reporter: T Jake Luciani
>            Assignee: T Jake Luciani
>             Fix For: 2.2.x
>
>
> There are a number of portability problems with python on win32 as I've learned over
the past few days.  
>   * Our use of multiprocess is broken in cqlsh for windows.  
> https://docs.python.org/2/library/multiprocessing.html#multiprocessing-programming
> The code was passing self to the sub-process which on windows must be pickleable (it's
not).  So I refactored to be a class which is initialized in the parent.
> Also, when the windows process starts it needs to load our cqlsh as a module. So I moved
cqlsh -> cqlsh.py and added a tiny wrapper for bin/cqlsh 
>   * Our use of strftime is broken on windows
> The default timezone information %z in strftime isn't valid on windows.  I added code
to the date format parser in C* to support windows timezone labels.
>   * We have a number of file access issues in dtest
>   * csv import/export is broken on windows and requires all file be opened with mode
'wb' or 'rb'
>  
> http://stackoverflow.com/questions/1170214/pythons-csv-writer-produces-wrong-line-terminator/1170297#1170297
>   * CCM's use of popen required the univeral_newline=True flag to work on windows



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message