On 2009-03-02 17:42-0800 Alan W. Irwin wrote: > I believe most users want their time handling to be done correctly, but > they prefer it to be done by libraries rather than struggling with, e.g., > leap seconds, themselves. Thus, I am quite pleased with these python example > 29 results because I think they demonstrate correct libqsastime and PLplot > time handling for the case of both positive and negative leap intervals with > TAI as the independent variable. > > As soon as I have finished the leap-interval changes for ctimeqsas in > libqsastime (which will allow broken-down UTC to be transformed to continous > TAI), then I plan to expand python example 29 to demonstrate that capability > as well with UTC being the independent variable.
I had plans to finish this all up for the release, but I got distracted by other stuff. What remains to be done is non-trivial not only in the programming sense but also in the understanding of leap-interval transformations which can make your head ache dealing with the discontinuities properly. Thus, I think it would be a big mistake to attempt to finish this in a last-minute rush. That's a perfect recipe for injecting bugs into a release. Therefore, I am going to hold off on more changes until after the release this coming weekend. The partial result we have now (leap intervals implemented for the transformation from continuous to broken-down time and for time formatting, but not for the transformation from broken-down time to continuous time) works for me and is probably okay, but has only been tested on Linux by me as far as I know. Even testing that ignores all the outer linear and leap-second interval transformations and which is therefore confined to just the formal transformation between continuous and broken-down time that is the innermost basis for libqsastime has been quite limited so far. A noble effort to do that "innermost" testing was made on Windows, but that was ultimately thwarted by an inability to find any software there that could be used to act as a basis for comparison with the libqsastime innermost results there. I am thinking of a different kind of test for Windows (and all other platforms) now where we compare libqsastime platform results with a file created on Linux where we _know_ the formal transformation results are correct via our comparisons with glibc time results for Linux.) The python version of example 29 which exercises some of the new stuff does work for me, but is far from finished. Thus, it should not be propagated to C and the other supported languages until after the release. In sum, I just feel a whole lot more comfortable delaying any more changes to the time results until we have had a full release cycle to test those. Therefore, I plan to release with the time-handling status quo with essentially zero publicity about the partial result we have now. I think the currently existing warning to expect more change in this area covers the situation. Alan __________________________ Alan W. Irwin Astronomical research affiliation with Department of Physics and Astronomy, University of Victoria (astrowww.phys.uvic.ca). Programming affiliations with the FreeEOS equation-of-state implementation for stellar interiors (freeeos.sf.net); PLplot scientific plotting software package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net); and the Linux Brochure Project (lbproject.sf.net). __________________________ Linux-powered Science __________________________ ------------------------------------------------------------------------------ Register Now & Save for Velocity, the Web Performance & Operations Conference from O'Reilly Media. Velocity features a full day of expert-led, hands-on workshops and two days of sessions from industry leaders in dedicated Performance & Operations tracks. Use code vel09scf and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf _______________________________________________ Plplot-devel mailing list Plplot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/plplot-devel
