Dear Ferret users,
This is to document a minor, known bug in Ferret and PyFerret, which seems to surface only very rarely. It's already reported and confirmed. At present, it's not known when it will be fixed. I also present a few workarounds we have found (Thanks to Ansley).
Suppose we have a Ferret script like this:
. . .
ppl dfltfnt TR
. . .
set view myview
. . .
repeat/L=1:3000 ( \
. . . ; \
fill . . . ; \
. . .; \
frame/file=image`L,zw=4`.png )
The timestamp on the image is eventually corrupted. See the attached picture. This happens only in some obscure situations.
1) One workaround is to reset the viewport at each iteration:
repeat . . . ( \
. . .
cancel view; set view myview )
2) (PyFerret) Another workaround is to use a different font
ppl dfltfnt TR ! → something else
3) A third is to stop the iteration before the corruption happens:
pyferret -nodisplay -script plot.jnl 1 1000 !! image Nos.1–1000
pyferret -nodisplay -script plot.jnl 1001 2000 !! image Nos.1001–2000
. . .
(Breaking up the loop within a single pyferret session doesn't seem to fix the problem. Hence the multiple sessions as above.)
Best regards,
Ryo