Hi all,
I've recently read in the changelog for v2.1.0 [1] that there is a new "C/C++ library for developers", that "encapsulates pgRouting related functions [and] Boost.Graph graphs".
However, I could not find any details on this library in the docs, but I'm curious what exactly this means and what new options may evolve from the presence of this library.
For instance, does it allow for using pgRouting functions without a database? Can anyone judge if using the C/C++ lib directly would resolve the well-known performance issues? I recently dropped pgRouting for a "real-world application" where a graph with approx. 7.5 mio edges had to be routed, due to bad performance (even with optimising Postgres configuration, RAMDISK etc.).
A general explanation of what is the major benefit of the newly created library and what can be done with it would help me understand a lot!
Many thanks in advance and all the best,
Matthias
---
[1] - https://github.com/pgRouting/pgrouting/releases/tag/pgrouting-2.1.0
On 10/8/2015 9:49 AM, Matthias Loeks wrote:
Hi all,
I've recently read in the changelog for v2.1.0 [1] that there is a new
"C/C++ library for developers", that "encapsulates pgRouting related
functions [and] Boost.Graph graphs".
However, I could not find any details on this library in the docs, but
I'm curious what exactly this means and what new options may evolve from
the presence of this library.
For instance, does it allow for using pgRouting functions without a
database? Can anyone judge if using the C/C++ lib directly would resolve
the well-known performance issues? I recently dropped pgRouting for a
"real-world application" where a graph with approx. 7.5 mio edges had to
be routed, due to bad performance (even with optimising Postgres
configuration, RAMDISK etc.).
A general explanation of what is the major benefit of the newly created
library and what can be done with it would help me understand a lot!
Mattias,
Vicky has been working very hard to rationalize the internal code to pgrouting. Part of this effort has been to rewrite a lot of the code and to leverage reusable components by placing them in a library. We are not supporting the library as a public API because it is undergoing lots of changes as we migrate more of the pgrouting code into this library.
Ideally, at some point in the future it would be nice to make the library available via a stable API, but that is not funded and it would be way in the future.
So good things happening internally, mainly to the benefit of the existing pgrouting code to make it more stable by reusing tested components. We would be happy if other C++ developers want to help with this effort or if you were interested in seeing if you can use the library in some other code.
Vicky might have more to say on this.
-Steve
Mathias:
Sorry for my late reply, I was on a trip.
As Steve mentions its an undergoing development, its not complete,
new bits and pieces will be emerging on the 2.x vesions. And when its
all incorporated to the new library, the documentation for developers
will also be there.
Vicky
To: pgrouting-users@lists.osgeo.org
From: woodbri@swoodbridge.com
Date: Thu, 8 Oct 2015 10:24:32 -0400
Subject: Re: [pgrouting-users] pgRouting 2.1.0 C library
On 10/8/2015 9:49 AM, Matthias Loeks wrote:
Hi all,
I’ve recently read in the changelog for v2.1.0 [1] that there is a new
“C/C++ library for developers”, that “encapsulates pgRouting related
functions [and] Boost.Graph graphs”.
However, I could not find any details on this library in the docs, but
I’m curious what exactly this means and what new options may evolve from
the presence of this library.
For instance, does it allow for using pgRouting functions without a
database? Can anyone judge if using the C/C++ lib directly would resolve
the well-known performance issues? I recently dropped pgRouting for a
“real-world application” where a graph with approx. 7.5 mio edges had to
be routed, due to bad performance (even with optimising Postgres
configuration, RAMDISK etc.).
A general explanation of what is the major benefit of the newly created
library and what can be done with it would help me understand a lot!
Mattias,
Vicky has been working very hard to rationalize the internal code to
pgrouting. Part of this effort has been to rewrite a lot of the code and
to leverage reusable components by placing them in a library. We are not
supporting the library as a public API because it is undergoing lots of
changes as we migrate more of the pgrouting code into this library.
Ideally, at some point in the future it would be nice to make the
library available via a stable API, but that is not funded and it would
be way in the future.
So good things happening internally, mainly to the benefit of the
existing pgrouting code to make it more stable by reusing tested
components. We would be happy if other C++ developers want to help with
this effort or if you were interested in seeing if you can use the
library in some other code.
Vicky might have more to say on this.
-Steve
Pgrouting-users mailing list
Pgrouting-users@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/pgrouting-users