[comp.windows.x] X11R3 and X11R4 compatibility under unicos

jones@ut-emx.UUCP (William L. Jones) (01/09/90)

There seems to be a problem with X11R3 and X11R4 compatibility under unicos.

I get the following message when I try to run a X11R3 client on unicos with
a X11R4 server.

>/usr/adm/bin/crayperf
>X Protocol error:  BadLength, poly request too large or internal Xlib length e\
rr
>or
>  Major opcode of failed request:  45 (X_OpenFont)
>  Minor opcode of failed request:  0
>  Resource id in failed request:  0xb00002
>  Serial number of failed request:  13
>  Current serial number in output stream:  13

Does anybody have a fix for this yet.

I am running with "xset bc".


Bill Jones

jdn@misc.Sun.COM (Jeff Nisewanger) (01/10/90)

In article <23045@ut-emx.UUCP> jones@ut-emx.UUCP (William L. Jones) writes:
>There seems to be a problem with X11R3 and X11R4 compatibility under unicos.
>
>I get the following message when I try to run a X11R3 client on unicos with
>a X11R4 server.
>
>>/usr/adm/bin/crayperf
>>X Protocol error:  BadLength, poly request too large or internal Xlib length e\
>rr
>>or
>>  Major opcode of failed request:  45 (X_OpenFont)
>>  Minor opcode of failed request:  0
>>  Resource id in failed request:  0xb00002
>>  Serial number of failed request:  13
>>  Current serial number in output stream:  13
>
>Does anybody have a fix for this yet.
>
>I am running with "xset bc".
>
>
>Bill Jones

	We ran across this problem recently between a cray X client
and the OpenWindows server. What we found was that the client
was sending an X OpenFont protocol request that had 7 trailing pad
bytes instead of the legal maximum of 3.


Jeff Nisewanger                         ARPA: jdn@sun.com
Window Systems Group                    UUCP: ...!sun!jdn
Sun Microsystems, Inc.                  415/336-5743

rws@EXPO.LCS.MIT.EDU (Bob Scheifler) (01/10/90)

    There seems to be a problem with X11R3 and X11R4 compatibility under unicos.

The R4 server checks for a correct length on OpenFont requests (and other
requests); the R3 server did not.  Apparently something you are using in
the R3 client generates a bad length (probably including the NUL in the
length of the string?).

    Does anybody have a fix for this yet.

Rebuild the client?

    I am running with "xset bc".

This only "fixes" a few specific things that we knew would be big problems
for many R3 clients.  It was not intended as a door to reinstate all possible
R3 bugs.