[5584] in Athena Bugs
vax 7.0F: X11r4 libX11.a, XGetDefault() function.
daemon@ATHENA.MIT.EDU (vanharen@ATHENA.MIT.EDU)
Fri Jul 20 10:32:36 1990
From: vanharen@ATHENA.MIT.EDU
To: bugs@ATHENA.MIT.EDU
Date: Fri, 20 Jul 90 10:32:21 EDT
System name: fries
Type and version: CVAXSTAR 7.0F (2 update(s) to same version)
Display type: SM
What were you trying to do?
Get an X default from my app-defaults file using XGetDefault().
What's wrong:
Under R4, it is not possible to specify defaults with periods in
them. Under R3, this was possible.
What should have happened:
XGetDefault() should allow you to get defaults with periods in
them, or the manual should mention that this is not possible.
Please describe any relevant documentation references:
XGetDefault() manpage, X11 manual, etc.
Probable cause:
Under R3, XGetDefault contained this code (XGetDflt.c, lines 145-148):
sprintf(temp, "%s.%s", progname, name);
XrmGetResource(dpy->db, temp, "Program.Name", &type, &result);
return (result.addr);
R4 has this instead (XGetDflt.c, lines 152-159):
names[0] = XrmStringToName(progname);
names[1] = XrmStringToName(name);
names[2] = NULLQUARK;
classes[0] = XrmStringToClass("Program");
classes[1] = XrmStringToClass("Name");
classes[2] = NULLQUARK;
(void)XrmQGetResource(dpy->db, names, classes, &fromType, &result);
return (result.addr);
XrmQGetResource and XrmGetResource are essentially unchanged from R3 to
R4, so the differences in the code above is probably the culprit.
Possible fix:
Put the R3 code back, or substitute something like this:
sprintf(temp, "%s.%s", progname, name);
XrmStringToNameList(temp, names);
classes[0] = XrmStringToClass("Program");
classes[1] = XrmStringToClass("Name");
classes[2] = NULLQUARK;
(void)XrmQGetResource(dpy->db, names, classes, &fromType, &result);
return (result.addr);
You'd also have to make "names" larger, since it is only an array of 3
right now. Note that I have not tested this patch, I'm just suggesting
what I think might be right off the top of my head.
I'm not sure why the R3 code was changed to the way it is for R4. I
suppose there is one less function call performed in the R4 version, but
it causes the problem I mentioned above. If the R4 code is left as-is,
please make sure that this incompatibility is documented somewhere.
