Re: Windows CVS 1.12.12.1 Latest + Patch - "version" Command Hangs

[Derek Price]

Conrad T. Pino wrote:

>I'm not so sure disk files are always ready.  They may use this approach
>because I can't find a way to find out if a disk file will block without
>actually launching an asynchronous read/write operation which raises the
>problem of what to do with the data when the operation suceeds!!!
>  
>

Well, if wouldn't work in the select() replacement, but if it were
implemented in fd_buffer_input, then extra data could be stored in the
buffer structure.  I'd prefer the select() replacement, but if needed
this could be used as a fallback.

>Can you describe the general I/O concurrency strategy CVS uses or wants
>to use on POSIX platforms?
>

Hrm?  If there was a select() function that worked to the POSIX
semantics I already forwarded, CVS's needs should be satisfied.

>I assume you've seen my prior message about the bad behaviour existing
>in version 1.12.12 also.  Does it make sense to probe versions earlier
>then 1.12.12 to see if we can revert to that method?
>  
>

No.  There was a good reason for moving to fd_buffers from stdio
buffers, though I can't recall it off the top of my head, and it would
be pretty hard to extricate them now anyhow.

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>