I can do it.I think the best way is to decouple telemtry charts retrieval and display:
* A daily task put all telemetry charts (png files) in a fixed directory under hackyDevSite's control. The daily task is responsible for overwrites the png files (i.e. update the chart) periodically.
* All png files must be of the same size, so that I know how to lay out them on a web page nicely.
* The name of the png files are used as chart descriptions (saving the trouble of using an xml file to describe what the chart is).
* hackyDevSite does not have to how many png files are there in the directory. Each time the telemetry chart page is requested, it scans the directory, and dynamically generates the web page containing all the charts. No caching, no optimization will be used, since hackyDevSite is a low volume web site.
Any comments.
Thanks.
Cedric
Aaron Kagawa wrote:
Cedric,
I'll defer this task to you, since you're the "Telemetry Guy".
If you decline, at the very least please send me a list of Charts available using the TelemetryChartAnalysis that you feel is "interesting". I'm not sure but, you might want to stay away from Charts that provide Member level analyses.
thanks, aaron
At 10:15 AM 5/12/2005, you wrote:
--On Wednesday, May 11, 2005 11:29 PM -1000 Aaron Kagawa <[EMAIL PROTECTED]> wrote:
Philip: do you want to use this Ant code anywhere in CSDL's build process?
ABSOLUTELY!!!!
What about the following:
- Update public server so that Ant task will work. - Update the hackyDevSite package to replace the "Build Archive" link in the navigation bar with a link entitled "Telemetry" which goes to a new page that contains a selection of interesting telemetry charts (review the Telemetry Wall charts to find some interesting candidates). - Configure a new Ant task to run once a night on hackydev which generates the set of charts you find interesting for hackystat and dumps them into a directory available to the web page pointed to by the "Telemetry" link.
The only tricky part is that the hackyDev webapp gets checked out and redeployed each night automatically. This means that you might want to actually make the chart dumper task a part of the redeployment process, occuring right after the webapp gets deployed. It might be best to encapsulate the chart dumping in its own build.xml file, and call out to it from the build.xml file associated with building the hackydevsite webapp.
Anyone see any problems with this approach?
Cheers, Philip
