David Garamond wrote:
> Robert Treat wrote:
>
>> Given that the cygwin version is currently labeled as not ready for
>> production I would say you are right. The truth is that many will never
>> declare win32 good for production simply because of the OS it runs on,
>> but we still want to make it as solid as possible.
>
>
> People _do_ use postgresql+cygwin in production environments though (see
> the pgsql-cygwin archive).
>
> And I suspect people _will_ use 7.5 for win32 in production, despite the
> release notes and the website clearly saying it's not production ready.
> Why?
>
> 1) The version number is "7.5" and many people will presume the ports
> are more or less equal in quality/maturity since they have the same
> version number;
>
> 2) People don't read release notes. See the various reviews on the
> recently released Fedora Core 2, complaining about how it doesn't
> support MP3 or DVD playback, despite the [legal] issues having been
> known and documented since Red Hat 8. Strangely enough, these people
> (who don't read release notes) _do_ write public reviews. They will
> badmouth PostgreSQL, saying it's unstable, crashes a lot, MySQL being
> much much more rock solid, etc etc.
>
> I suggest we label the win32 port as "7.5 ALPHA" or "7.5 DANGEROUS" :-)
My concern is about the fact the fact that Postgresql rely on the OS about
his ability to optimize memory access. Do we have any possibility at this
stage to compare Postgresql performance on top of Win32 with other products
for this platform ?
I think that Postgresql with this version is going to address a specific
class of final users that right now are using what ? I don't think the
response is: "users that now are using postgresql+cygwin". Can the first
version of postgresql for Win32 compared with other products? I really
hope yes, you know: there is no a *second* possibility to do a *first* good
impression.
Regards
Gaetano Mendola
