graphics window problem on Macs?
Philip Massey wrote on Apr 19, 2007
Hi, Mike et al.
One very annoying and inconsistent problem I have in splot (say): if I expand
plot a spectrum using "splot" and then use the "E" key to expand, and then try to print the graphics window by hitting "=",
I SOMETIMES get an "ERROR: sgmentation violation" and no print. (I do see the echo ".snap - " in the graphics window, but not ".snap - done", as it should
say. Subsequent efforts to run "splot" fail to bring up the plot at all even though the graphics window is still there; instead I just immediately get a prompt back. The only solution I've found is to log out of the cl and back.
Doing exactly the same steps again will then (usually) work.
I'm running this on a Power Mac G5 with Tiger 10.4.9. I have the identical problem on my Power Book G4 machine.
I realize that the fact this only happens SOME of the time probably makes this tough to diagnose. Any ideas?
thanks,
phil
One very annoying and inconsistent problem I have in splot (say): if I expand
plot a spectrum using "splot" and then use the "E" key to expand, and then try to print the graphics window by hitting "=",
I SOMETIMES get an "ERROR: sgmentation violation" and no print. (I do see the echo ".snap - " in the graphics window, but not ".snap - done", as it should
say. Subsequent efforts to run "splot" fail to bring up the plot at all even though the graphics window is still there; instead I just immediately get a prompt back. The only solution I've found is to log out of the cl and back.
Doing exactly the same steps again will then (usually) work.
I'm running this on a Power Mac G5 with Tiger 10.4.9. I have the identical problem on my Power Book G4 machine.
I realize that the fact this only happens SOME of the time probably makes this tough to diagnose. Any ideas?
thanks,
phil
Mike Fitzpatrick wrote on Apr 19, 2007
Hi Phil,
I played with it a bit but couldn't reproduce the problem at all. It sounds as if the segvio is coming from SPLOT and not the graphics printing code (since it hangs afterwards) and so a simple flpr should avoid having to log back in. To fix it though I'll need more information, e.g. do you use the 'E' key to expand in some extreme fashion or just "normally"? Is this an especially large or complicated spectrum where increasing the size of the 'cmbuflen' (graphics buffer defined in login.cl) makes the problem go away? Any chance you can find a reproducible case? Is this happening a few times a day or a few times a month?
Cheers,
-Mike
I played with it a bit but couldn't reproduce the problem at all. It sounds as if the segvio is coming from SPLOT and not the graphics printing code (since it hangs afterwards) and so a simple flpr should avoid having to log back in. To fix it though I'll need more information, e.g. do you use the 'E' key to expand in some extreme fashion or just "normally"? Is this an especially large or complicated spectrum where increasing the size of the 'cmbuflen' (graphics buffer defined in login.cl) makes the problem go away? Any chance you can find a reproducible case? Is this happening a few times a day or a few times a month?
Cheers,
-Mike
Philip Massey wrote on Apr 19, 2007
Hi, Mike---
A "flpr" or even a couple (followed by a flpr 0) doesn't help.
I'm expanding normally, not anything extreme.
The spectra I saw this on yesterday were hydra spectra that had been scopy'd
down to individual one-d format; they are only about 2000 pixels in size.
I would say that it happens once every few days. I tried very hard to find a reproducible example yesterday, but gave up in disgust when I found that the same spectrum failed to cause the same problem when I had done the exact same steps. It made me wonder if logging out the cl and back in had cleared up other problems.
Sorry this is all so vague and useless...
cheers,
phil
A "flpr" or even a couple (followed by a flpr 0) doesn't help.
I'm expanding normally, not anything extreme.
The spectra I saw this on yesterday were hydra spectra that had been scopy'd
down to individual one-d format; they are only about 2000 pixels in size.
I would say that it happens once every few days. I tried very hard to find a reproducible example yesterday, but gave up in disgust when I found that the same spectrum failed to cause the same problem when I had done the exact same steps. It made me wonder if logging out the cl and back in had cleared up other problems.
Sorry this is all so vague and useless...
cheers,
phil
Last post on Apr 19, 2007