Blank tiles
Stijn Tallir
I have succesfully set up a few tilesets with mod-geocache. Great
software BTW.
There is just a small problem with loading new (unseeded) tiles in
openlayers.
Sometimes a few (randomly) new tiles aren't loaded (I get the pink
image with the broken tile icon in the top left corner). However when
I scroll or zoom in & out the images appear just fine.
Is there some persistency parameter I can adjust? Any other
suggestions?
Stijn
thomas bonfort
http://trac.osgeo.org/openlayers/ticket/447
--
thomas
> --
> You received this message because you are subscribed to the Google Groups "mod-geocache" group.
> To post to this group, send email to mod-ge...@googlegroups.com.
> To unsubscribe from this group, send email to mod-geocache...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/mod-geocache?hl=en.
>
>
Jeff McKenna
OpenLayers viewer on mapserver.org.
I just want to make sure we give our feedback from the experience:
- cache lives on demo.mapserver.org, a shared OSGeo virtual machine
(http://wiki.osgeo.org/wiki/AdhocVM)
- uses disk cache
- I am now running geocache_seed there
- I notice that Thomas made some more changes to it this morning, thanks
(what did you change, other than removing the restrictExtent stuff,
which i also removed but did not checkin yet... please explain)
- using mod_geocache trunk (from yesterday) we kept getting a configure
error:
{{{
./configure: line 1566: syntax error near unexpected token `fi'
./configure: line 1566: `fi'
}}}
Solved this by commenting the "else" on line 1566
- when we ran "autoconf" on that machine, we got many macro warnings:
{{{
jmckenna@adhoc:/osgeo/mapserver/mod-geocache-trunk$ autoconf
autoconf: Undefined macros:
configure.in:124:
AC_HELP_STRING([--with-pkg-config[=PATH]],[path to pkg-config)]),
configure.in:151:
AC_HELP_STRING([--with-cairo[[=ARG]]],[Include Cairo Support
(ARG=yes/no/path to cairo.pc)]),
configure.in:194:
AC_HELP_STRING([--with-pixman[[=ARG]]],[Include Pixman Support
(ARG=yes/no/path to pixman.pc)]),
configure.in:244: AC_HELP_STRING([--enable-memcache],[Enable
memcache backend]),
configure.in:284: AC_HELP_STRING([--with-apu-config], [path to
apu-config program]),
configure.in:341: AC_HELP_STRING([--with-apr-config], [path to
apr-config program]),
configure.in:446: AC_HELP_STRING([--enable-debug],[Enable
debug build]),
configure.in:460:
AC_HELP_STRING([--enable-strict-json],[Make the JSON parser be strict
and reject comments]),
configure.in:480: AC_HELP_STRING([--with-gdal], [path
to gdal-config program]),
configure.in:518: AC_HELP_STRING([--with-geos], [path
to geos-config program]),
configure.in:550:
AC_HELP_STRING([--with-fastcgi=DIR],[Specify where FASTCGI is installed]),
configure.in:601: AC_HELP_STRING([--with-png=DIR],[Specify
where PNG is installed]),
configure.in:653: AC_HELP_STRING([--with-sqlite],
[path to sqlite-config program]),
configure.in:724: AC_HELP_STRING([--with-jpeg=DIR],[Specify
where JPEG is installed]),
configure.in:775:
AC_HELP_STRING([--with-lock-mechanism=file|semaphore],[Specify the
locking mechanism to use, either file based or posix-semaphore based.]),
configure.in:794: AC_HELP_STRING([--enable-version-string],[add
mod-geocache version string to the apache server header]),
configure.in:803: AC_HELP_STRING([--enable-module],[Enable
apache module building]),
configure.in:845:AC_CONFIG_FILES(Makefile.inc)
configure.in:96: AC_HELP_STRING([--with-curl-config], [path to
curl-config program]),
}}}
which then causes the following problem in configure:
{{{
./configure: line 3140: syntax error near unexpected token `Makefile.inc'
./configure: line 3140: `AC_CONFIG_FILES(Makefile.inc)'
}}}
Boy that was a tough one to solve. It turns out that on that shared
virtual machine when you run "autoconf" we're actually running a very
old autoconf2.13 version. Long long story short: we had to uninstall
autoconf2.13 and install a fresh autoconf2.68...then those warnings went
away. phew.
- we also noticed a difference in the map image from a geocache WMS
request to a straight WMS request:
This is unresolved. We tried changing grids from 4326 to google
grids...and other things. That result/problem is still unsolved.
If you have ideas please share. It is an OSGeo machine so you can also
dive right in! :)
thanks.
-jeff&mike
thomas bonfort
> Yesterday Mike and I spent the day configuring caching for the OpenLayers
> viewer on mapserver.org.
>
> I just want to make sure we give our feedback from the experience:
>
> - cache lives on demo.mapserver.org, a shared OSGeo virtual machine
> (http://wiki.osgeo.org/wiki/AdhocVM)
>
> - uses disk cache
>
> - I am now running geocache_seed there
>
> - I notice that Thomas made some more changes to it this morning, thanks
> (what did you change, other than removing the restrictExtent stuff, which i
> also removed but did not checkin yet... please explain)
there's no use in specifying a restricted extent. I also switched the
layer to TMS from WMS to avoid scaling issues (see below)
will have to add a check for autoconf version, or commit a configure
script generated by a recent version of autoconf (the release tarballs
have that one, but not the svn trunk checkouts). Might also want to
check which are the offending macros that fail in older autoconf
versions.
>
> - we also noticed a difference in the map image from a geocache WMS request
> to a straight WMS request:
>
> geocache WMS GetMap request:
> http://demo.mapserver.org:8081/geocache?LAYERS=osm-denver&FORMAT=png&TRANSPARENT=off&SERVICE=WMS&VERSION=1.1.1&REQUEST=GetMap&STYLES=&EXCEPTIONS=application%2Fvnd.ogc.se_inimage&SRS=EPSG%3A4326&BBOX=-105.17578023551,39.77807379252,-105.16765279439,39.786201233642&WIDTH=256&HEIGHT=256
>
> straight WMS GetMap request:
> http://demo.mapserver.org:8080/cgi-bin/foss4g?LAYERS=OSM_Denver&FORMAT=png&TRANSPARENT=off&SERVICE=WMS&VERSION=1.1.1&REQUEST=GetMap&STYLES=&EXCEPTIONS=application%2Fvnd.ogc.se_inimage&SRS=EPSG%3A4326&BBOX=-105.17578023551,39.77807379252,-105.16765279439,39.786201233642&WIDTH=256&HEIGHT=256
>
> This is unresolved. We tried changing grids from 4326 to google grids...and
> other things. That result/problem is still unsolved.
it's "not-a-bug", your openlayers resolutions have to match the grid
resolutions, otherwise you get results that a up/downsampled from the
closest resolution tiles. The demo interface is here to provide the
javascript code you can cut'n'paste into your own javascript files.
--
thomas
>
> If you have ideas please share. It is an OSGeo machine so you can also dive
> right in! :)
>
> thanks.
>
> -jeff&mike
>
>
>
>