xcircuit crashes when using tk +aqua

[watsodw]

Installed xcircuit, but when I try to run it I get the following error:

dwwatson - /Users/dwwatson - 2 -->xcircuit
2012-12-04 13:37:23.768 wish[44904:60b] A call to OSACopyScriptingDefinition() by NSScriptSuiteRegistry returned an error. Is this application's OSAScriptingDefinition Info.plist entry valid? If its value is 'dynamic', is there an 'ascr'/'gsdf' Apple event handler registered?
2012-12-04 13:37:23.809 wish[44904:60b] MADLog - Screens: 
Segmentation fault

It's the same with xcircuit-devel, too. I have tcl @8.5.13_0+corefoundation+threads+universal installed.

[mf2k (Frank Schima)]

In the future, please use WikiFormatting. Cc'ing nomaintainer is not useful since it does not go to anyone.

[daledouglas@…]

Replying to david.w.watson@…:

Installed xcircuit, but when I try to run it I get the following error:

dwwatson - /Users/dwwatson - 2 -->xcircuit
2012-12-04 13:37:23.768 wish[44904:60b] A call to OSACopyScriptingDefinition() by NSScriptSuiteRegistry returned an error. Is this application's OSAScriptingDefinition Info.plist entry valid? If its value is 'dynamic', is there an 'ascr'/'gsdf' Apple event handler registered?
2012-12-04 13:37:23.809 wish[44904:60b] MADLog - Screens: 
Segmentation fault

It's the same with xcircuit-devel, too. I have tcl @8.5.13_0+corefoundation+threads+universal installed.

I confirm the bug. I have not tried xcircuit-devel. For me, wish also crashes, and then xcircuit reports seg faults.

tcl @8.6.0_1+corefoundation+threads
xcircuit @3.7.44_0

[Stochasticus@…]

Cc Me!

[Stochasticus@…]

I'm getting the same problem, but with even less information. The installed version of wish8.6 crashes with a segmentation fault. Seems to be a recurring problem. The last time I had something like this I reinstalled macports after the upgrade to Mountain Lion and it worked again. I'm not sure that will work here as this is happening on a new MacBook pro that had macports and xcircuit installed recently.

[Stochasticus@…]

I just wiped and reinstalled macports and xcircuit 3.7.44 just in case the order I'd done things with this new computer was wrong but I'm still getting the same crash from the version of wish 8.6 installed by the tcl/tk ports. Wish (huh...) I knew what happened to fix it when I reinstalled everything the last time this happened with xcircuit. Maybe this should be reported under tcl/tk rather than here.

[daledouglas@…]

Cc Me!

[daledouglas@…]

This ticket doesn't seem to be making progress -- is there anything I can do? Would reporting this upstream (i.e., the x-circuit dev team, whoever they may be) help? Many thanks!

[larryv (Lawrence Velázquez)]

Replying to daledouglas@…:

This ticket doesn't seem to be making progress -- is there anything I can do? Would reporting this upstream (i.e., the x-circuit dev team, whoever they may be) help? Many thanks!

Yes, you should probably report this upstream. You could also send a message to the macports-users mailing list; you’ll have a better chance of getting help from some Tcl/Tk-savvy folks.

[anddam (Andrea D'Amore)]

Cc Me!

[anddam (Andrea D'Amore)]

I mistakenly replied to this ticket rather than #30787, I see the reporter here already had 8.5.13 and his issue is different than mine anyway.

Since I've already poked the ticket: did any of you solve the issue on his own? Can you still reproduce the issue at all?

[anddam (Andrea D'Amore)]

Please check ticket:30787#comment:13 and see if that solves the issue. That's just a workaround and not a fix.

[cooljeanius (Eric Gallager)]

Cc Me!

[mcglk (Ken McGlothlen)]

Same results, over a year and a half since the bug was first reported. I get a segfault on xcircuit 3.7.50_1 and xcircuit-devel 3.8.40_1, running under OS X 10.9.3 (and tcl/tk 8.6.1).

Anyone had any success with this port yet?

[kurthindenburg (Kurt Hindenburg)]

Cc Me!

[kurthindenburg (Kurt Hindenburg)]

The latest xcircuit 3.8.38 still crashes on Mavericks - not sure if I'll be able to debug it

Exception Type:  EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x0000000000000050

VM Regions Near 0x50:
--> 
    __TEXT                 0000000104f82000-0000000104f86000 [   16K] r-x/rwx SM=COW  /opt/local/bin/wish8.6

[kurthindenburg (Kurt Hindenburg)]

Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0x0000000000000050
0x0000000100379d6b in CmapRecForColormap ()
(gdb) bt
#0  0x0000000100379d6b in CmapRecForColormap ()
#1  0x000000010037a115 in XcmsCCCOfColormap ()
#2  0x0000000100322f33 in XAllocNamedColor ()
#3  0x000000010550a9bc in CvtStringToPixel ()
#4  0x000000010550aa5f in xc_alloccolor ()
#5  0x000000010550385e in build_app_database ()
#6  0x0000000105504250 in GUI_init ()
#7  0x0000000105509010 in xctcl_start ()
#8  0x000000010015c697 in TclNRRunCallbacks ()
#9  0x000000010015d8bb in TclEvalEx ()
#10 0x00000001001ee528 in Tcl_FSEvalFileEx ()
#11 0x00000001001ed33e in Tcl_EvalFile ()
#12 0x00000001001f431e in Tcl_SourceRCFile ()
#13 0x000000010002895c in Tk_MainEx ()
#14 0x0000000100003edb in main ()

I've tried building it --without-tcl and various other configure options, but it won't build - also tried the 3.9 branch w/o any luck

[anddam (Andrea D'Amore)]

Replying to khindenburg@…:

The latest xcircuit 3.8.38 still crashes on Mavericks - not sure if I'll be able to debug it

Out of curiosity where are you getting 3.8.38 from? The stable branch is 3.7 while xcircuit-devel has been 3.8.40 (unstable) for more than a year.

[kurthindenburg (Kurt Hindenburg)]

Replying to and.damore@…:

Replying to khindenburg@…:

The latest xcircuit 3.8.38 still crashes on Mavericks - not sure if I'll be able to debug it

Out of curiosity where are you getting 3.8.38 from? The stable branch is 3.7 while xcircuit-devel has been 3.8.40 (unstable) for more than a year.

​http://opencircuitdesign.com/xcircuit/ - I mistyped - it is 3.8.68

​http://opencircuitdesign.com/xcircuit/archive/xcircuit-3.8.68.tgz

[anddam (Andrea D'Amore)]

I confirm I can reproduce the issue with both 3.8.69 (now stable branch) and 3.9.20 (development) on Mavericks.

[kurthindenburg (Kurt Hindenburg)]

The fix from ​http://www.opencircuitdesign.com/pipermail/magic-dev/2014/000641.html works

sudo port install tk +x11

I'll update the Portfile shortly

[kurthindenburg (Kurt Hindenburg)]

r121693

[anddam (Andrea D'Amore)]

Let's keep this open until solved, forcing +x11 is a workaround.

[SickTeddyBear]

Cc Me!

[chrstphrchvz (Christopher Chavez)]

Replying to anddam:

Let's keep this open until solved, forcing +x11 is a workaround.

According to a ​mailing list post by xcircuit's author, xcircuit relies on "low-level X11/Xlib calls". Compared to other Tk apps, it sounds non-trivial for xcircuit to support Tk Aqua, since it isn't completely relying on Tk's abstractions of whichever windowing system is used. Does it still make sense to leave this ticket open if tk +x11 isn't really a workaround but rather a requirement?

Currently, xcircuit (if the require_active_variants restriction is removed) and xcircuit-devel can't even compile with tk +quartz mainly due to redefinition errors in Xlib.h.

[anddam (Andrea D'Amore)]

Replying to chrstphrchvz:

Does it still make sense to leave this ticket open if tk +x11 isn't really a workaround but rather a requirement?

Considering the little interest in the ticket (two CC's in last five years) and the issues with Aqua I'd say it's safe to close.

[chrstphrchvz (Christopher Chavez)]

In 13a73de8aa212e6358ae39e494ace887b3aa4f47/macports-ports (master):

xcircuit[-devel]: require +x11 variants of tk and cairo

Not compatible with tk +quartz
Copied from portfile for xcircuit
See #37203

Fixes #57666