On Saturday, July 23, 2011, Christopher Browne <cbbrowne(at)gmail(dot)com> wrote:
> I generally agree, Josh, but I think readline is getting pointed at a bit too much. Yeah, it's a bad one, but we also include other stuff like zlib that doesn't commonly come up as an issue.
> I'd argue something just a wee bit different...
> By the time we would add in:
> - autoconf rules to detect it,
> - makefile rules to link it in
> - include file changes
> - wrappers to ensure use of pmalloc
> - Debian guys add build dependancies
> - rpm dependencies get added
> - BSD ports dependencies
> That is likely rather more code than 1 not terribly large file of C needed to do it ourselves. And this code is rather worse, as it is in a bunch of languages, spread all over.
> If we were gaining a lot of extra functionality "for free" it would be one thing. That is true for libssl, and likely zlib, but not here.
Also consider if the library is widely available on common distros or
not. If not, packagers are going to have to start packaging that
first, in order to build the PostgreSQL packages. This is a *huge*
issue for use if we want to use wxWidgets addon libraries with
EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
In response to
pgsql-hackers by date
|Next:||From: Albe Laurenz||Date: 2011-07-23 06:54:26|
|Subject: Re: Questions and experiences writing a Foreign Data Wrapper|
|Previous:||From: Josh Kupershmidt||Date: 2011-07-23 02:44:30|
|Subject: Re: psql: display of object comments|