Thread: win32 open item

win32 open item

From
Magnus Hagander
Date:
The open items list for 8.4 has

*  problems with Windows global namespace


I propose we move this over to the TODO list instead. We are no worse
off than we are in previous versions, and we've had *zero* bug reports
related to this issue.

Basically, we have a small window of failed protection if a user that
has permissions, manually starts the database engine from two different
terminal server sessions at the same time. Given that AFAIK *every
single* production deployment would be running the server as a service,
is probably why we have had zero reports about it.

The issue that caused the server to not start properly has already been
fixed in both 8.3 and head.


I don't think this is anywhere near important enough to hold up a
release over.

//Magnus


Re: win32 open item

From
Merlin Moncure
Date:
On Sat, Mar 21, 2009 at 12:11 PM, Magnus Hagander <magnus@hagander.net> wrote:
> The open items list for 8.4 has

*) "PQinitSSL broken in some use casesf"
This should be addressed.  Andrew C proposed a few different ways to
do it and supplied patches.  Pick one and apply it.

*) Re: [HACKERS] patch to fix client only builds
defer to 8.5

merlin


Re: win32 open item

From
Tom Lane
Date:
Magnus Hagander <magnus@hagander.net> writes:
> The open items list for 8.4 has
> *  problems with Windows global namespace

> I propose we move this over to the TODO list instead.

Agreed.  In general, any issue that already exists in released versions
should not be considered a blocker for 8.4.  Certainly it wouldn't be
a blocker for beta release.
        regards, tom lane


Re: win32 open item

From
Merlin Moncure
Date:
On Sat, Mar 21, 2009 at 12:15 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Magnus Hagander <magnus@hagander.net> writes:
>> The open items list for 8.4 has
>> *  problems with Windows global namespace
>
>> I propose we move this over to the TODO list instead.
>
> Agreed.  In general, any issue that already exists in released versions
> should not be considered a blocker for 8.4.  Certainly it wouldn't be
> a blocker for beta release.

The 'ssl use cases thing' fits under that umbrella...so defer that
too...maybe we can sneak a doc patch in later...to busy now to worry
about it.

merlin