[Live-demo] 4.5alpha3 Built - Mapfish

[Hamish]

Angelos wrote:
> I just did an update on the vm as Cameron suggested and
> Mapfish does not start.
> The script log shows an error about not being able to
> update some db table, being accessed by other users.
>
> After trying the Mapfish icon, I get an error that the page
> is not found.

try editing the icon file:
-Exec=sensible-browser http://localhost/mapfishsample/osgeolive/wsgi/
+Exec=sensible-browser http://localhost/mapfishsample_2.0/osgeolive/wsgi/


if the old dir was present on the author's machine, it would
explain why it would still work for them but not on a new install.

I've just read over the install script (but not tried it).
some comments, the most important one is the change above, and
'svn co' vs 'svn export' (I'm on the hunt for disc space).


unused variable?
  BIN="/usr/bin"

please install to /usr/local/bin:
  INSTALL_DIR="/usr/lib"

if not using /tmp/build_mapfish/ please use 'svn export' not 'svn co' as a
full checkout doubles the required disk space (keeps a full copy of originial
files in .svn/ so offline diffs work):
  svn co http://mapfish.org/svn/mapfish/sample/trunk MapfishSample
probably best to do that in /tmp/build_mapfish/ then copy to destination
rather than directly into the filesystem.

please use /tmp/build_mapfish not /tmp

please use 'psql --quiet' if needed (? create_database.bash)

epsg:900913 is already stored in /usr/share/proj/esri.extra

re. other people's troubles, perhaps stop tomcat at the start of the script,
then start it again at the end instead of just
  /etc/init.d/tomcat6 restart
at the end? (? probably that's really a pgsql error not a tomcat one but who knows...)

do not use 'wget -c' with an unversioned file name. it simply works by
comparing number of bytes and adding more to the end. if upstream file
changes you get a corrupted download. Use "wget -nv" for tiny files.



thanks,
Hamish