Dave Page wrote:
> On Tue, Aug 12, 2008 at 12:51 PM, Mickael Deloison <mdeloison@gmail.com> wrote:
>> Indeed! A big code review is required and perhaps a lot of more tests.
>> So maybe this way of doing (separate program) is good for now. What's
>> your opinion about that? Should I keep pgScript code integrated into
>> pgAdmin or should I keep it as a separate program?
>
> Well that really is an interesting question. I'm not sure I know the
> answer, so I'll attempt to sum up some pros and cons and see what the
> other hackers think:
>
> Pro integration:
>
> - One less dependency for packages to worry about
> - One build system to maintain
> - Functionality is more responsive
> - No problems with different DLL versions if builds come from
> different machines.
> - Cannot be broken by moving one exe etc.
>
> Pro seperation:
>
> - pgScript can be maintained (and upgraded) independently of pgAdmin
> - No big impact on the pgAdmin source tree (and consequent learning
> curve for other developers).
>
> Despite the difference in numbers, I think those pros and cons are
> roughly equal.
>
> Any other thoughts? (Magnus, Guillaume, Hiroshi etc).
I think the Pros overweigh, so I'd like to see it there.
If the isolation is still pretty good, there is no need for a large
number of "pgadmin developers" to get a learning curve to get over. And
the part in the query tool will need to be done that way *anyway*.
The only "major" con is the upgrade one, but I think that's fairly
survivable.
Since pgscript depends on all the wx stuff anyway, it's not feasible for
it to ever be included in the backend distribution, which is what I
would've liked even more (as a libpgscript thingy). I think bundling it
with pgadmin makes most sense in this case, both in a binary and source
perspective, the same way we do with pgAgent.
//Magnus