iraf.net

Hi,I am debugging a problem in IRAF 2.15.1a plotting via PyRAF. I know you are not in the business of PyRAF support (I am), but I am curious about changes to IRAF graphics tasks between 2.14 and 2.15 so I thought I'd ask.With 2.15, the command to plot/append (prow dev$pix 333 append+) gives:
[code:1:0d81da1e1f]
851, "Cannot restore graphics world coordinate systems"
[/code:1:0d81da1e1f]this does not occur in the CL, only in PyRAF. But this works with 2.14 and the calling code in PyRAF has not changed. And I note that the GKI output from a simple plot (prow dev$pix 333) is different in 2.15 than it was in 2.14:[code:1:0d81da1e1f]
% diff plot.214.mc.txt plot.215.mc.txt
3c3
< set_wcs nwords=352
---
> set_wcs nwords=384
7c7
< set_wcs nwords=352
---
> set_wcs nwords=384
46c46
< set_wcs nwords=352
---
> set_wcs nwords=384
[/code:1:0d81da1e1f]Looking at the IPC has been unhelpful so far. Do you have any suggestions? I figure if there was a change to the graphics code, then the CL itself was also modified to handle it.All of the above is for 32-bit.Thanks,
Chris

I think what you're seeing is the effect of this change:[code:1:c69bfadd20]
lib/gio.h
sys/gio/cursor/gtr.h
Increased the size of the WCS buffer. The size was previously calculated
as being 17 structure elements by assuming the SZ_INT was 2. Increased
to accomodate 64-bit sizes and will live with the wasted space.
[/code:1:c69bfadd20]The specific change is to the GP_WCSPTR in gio.h and LEN_TRSTRUCT in gtr.h (diff the files for details).

Thanks a lot. That directed me to the change in gio.h where LEN_WCS was changed from 11 to 12. In the end, that was what I need to work around in PyRAF. The size of the WCS metacode burst is now 384 instead of 352.For any PyRAF users reading this post, the graphics changes to IRAF are now being correctly handled (at least for 32-bit) in the current version of PyRAF (i.e. the nightly tar-ball until v1.11 is out).Thanks again,
Chris

Good, thanks for the info.

Chris,What about 64 bit version? Have just tried latest IRAF x86_64 with today's nightly build of pyraf and pstselect fails with this error. Any chance to get this fixed or better rollback to 2.14?--
Best regards,
Ivan