bug-cvs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Windows Build Broken: Function "piped_child" New Argument "bool f


From: Derek Price
Subject: Re: Windows Build Broken: Function "piped_child" New Argument "bool fix_stderr"
Date: Fri, 30 Sep 2005 01:41:56 -0400
User-agent: Mozilla Thunderbird 1.0.6 (Windows/20050716)

Conrad T. Pino wrote:

>>From: Derek Price
>>
>>Conrad T. Pino wrote:
>>
>>    
>>
>>>I'm adding the argument to Windows "piped_child" function.
>>>
>>>What if anything should be done with the new argument?
>>>      
>>>
>>I don't know if the problem that that argument is designed to work
>>around exists on Windows.  When the problem exists on UNIX, the
>>sshstdio-6 test in sanity.sh fails, if you'd like to try and construct a
>>test case on Windows.
>>    
>>
>
>This suggestion implies knowledge I don't have.  Unless specific direction
>appears, this is as far as I take this one.
>  
>

Search for "sshstdio-6" test in the src/sanity.sh file.  It's just a
series of Bourne shell commands, and it likely depends on a few earlier
setup commands.  If you can find your way around a Solaris command
prompt, you should be able to interpret the commands you find.

Basically, and I'm going to get some of these details wrong, when CVS is
exec'd `cvs co -p somefile >logfile 2>&1', and somefile is REALLY big,
and :ext:/ssh access method is being used, then data gets lost. 
(because SSH sets its stderr to nonblocking and stderr and stdout have
been set to the same descriptor - this discussion has been going on for
years
<http://groups.google.com/group/gnu.cvs.bug/msg/8618402c11a7e981?dmode=source>).

Even the shell redirection syntax is the same on Windows.  If you want
to investigate, you should be able to.

Regards,

Derek

-- 
Derek R. Price
CVS Solutions Architect
Ximbiot <http://ximbiot.com>
v: +1 717.579.6168
f: +1 717.234.3125
<mailto:derek@ximbiot.com>






reply via email to

[Prev in Thread] Current Thread [Next in Thread]