2011/6/15 Guillaume Lelarge
<guillaume@lelarge.info>On Wed, 2011-06-15 at 10:00 +0200, Jasmin Dizdarevic wrote:
> You're right. Patching very old releases will become more difficult.
> But those patches are rare and most of the time not very large, so the
> merge should not be a big problem.
I'm not sure I understand. Do you mean that users will have to backpatch
some bugfixes and compile an old pgAdmin to be able to use old
PostgreSQL releases? I sincerely hope you don't mean that.
No, of course. I wanted to say, that if we change the current code to much, patching older versions with patches developed with the current head would become more complicated - because there will probably be more conflicts and so on.
> Either one actual version of pgAdmin which supports everything back to
> 7.3. or different versions which targets different postgresql
> releases. Currently we have a hybrid - that's not very good, I think.
We don't have both. We currently have a pgAdmin which supports
everything back to 7.3. And you cannot say otherwise, unless you know
some bugs you didn't report.
You don't like the idea to remove the old code from pgAdmin, but we have older versions which are designed to work with older version of postgresql. If there are any poor souls working with < 8.2, in my opinion they should also use an older pgAdmin client - Dave described it in a later mail: they must! So removing the old code wouldn't be so bad and it wouldn't break any functionality.
I'm saying we're not actively doing that, but if someone wants
to do so, or if we want to simplify things when working on a
particular piece of code, then removing the support for older versions
is acceptable.
I agree to Dave's idea.
--