Jeremy Faden

BasicStroke on Windows

Note that any data source which maintains a state will potentially have this problem. A better solution might be to have "reload" create a new DataSource object.

I've fixed the CDF metadata as well.

The problem was the CDF reader had a feature that it would try to only read the attributes once, making it stateful when other codes expect them to be stateless. Reading the metadata takes less than a millisecond and is not worth caching. However this showed me a potential error, where when the metadata was cached it would do actions which did not occur when reading for the first time. This needs further study.

I finally met with Ivar and we did verify that changing CDF files does not update. See sftp://draylax.physics.uiowa.edu:/home/jbf/ct/autoplot/u/2026/ivar/20260903/ cp ts1_l2_ace_def_20260401_v1.1.0.cdf to file.cdf plot file.cdf?Epoch cp ts2_l2_ace_def_20260401_v1.1.1.cdf to file.cdf reload all data The metadata tab does not update.

I notice that TimeParser only allows one field type per URI, so data_$Y-$m-$d_$(enum;id=sc;values=A,B)_$(enum;id=mode;values=slow,fast).dat will not work presently. I'll correct this soon.

Chris pointed out an area that Larry has made which I could use for testing on campus: \\spdr0.physics.uiowa.edu\home\jbf\data\istp\omni\omni_cdaweb\hro_5min\2007\omni_hro_5min_20070301_v01.cdf ChatGPT suggests I use file://spdr0.physics.uiowa.edu/home/jbf/data/istp/omni/omni_cdaweb/hro_5min/2007/omni_hro_5min_20070301_v01.cdf as the URI for this. Since there are two slashes, this would work.

Chris pointed out an area that Larry has made which I could use for testing on campus: \\spdr0.physics.uiowa.edu\home\jbf\data\istp\omni\omni_cdaweb\hro_5min\2007\omni_hro_5min_20070301_v01.cdf ChatGPT suggests I use file://spdr0.physics.uiowa.edu/home/jbf/data/istp/omni/omni_cdaweb/hro_5min/2007/omni_hro_5min_20070301_v01.cdf as the URI for this. Since there are two slashes, this would work.