Number: 910
Title: fdp catches signal and emits graph but all coords are 0,0
Submitter: Stephen North
Date: Fri Mar 17 09:19:24 2006
Subsys: Output generation
Version: 2.9
System: *-*-
Severity: major
Problem:

Run fdp on a really huge graph.
    fdp -v hugegraph.dot > hugegraph.out

After a while, issue an INTR.


hugegraph.out contains a zillion lines of the form
        "Robert J. Herbold" [pos="0,0", width="1.75", height="0.50"];
        "Agilent Technologies, Inc." [pos="0,0", width="2.42", height="0.50"];
        "Heidi Kunz" [pos="0,0", width="1.28", height="0.50"];
        "David M. Lawrence" [pos="0,0", width="1.92", height="0.50"];
        "Addison Barry Rand" [pos="0,0", width="2.00", height="0.50"];
        "Edward W. Barnholt" [pos="0,0", width="1.97", height="0.50"];
        "Gerald Grinstein" [pos="0,0", width="1.67", height="0.50"];
        "James G. Cullen" [pos="0,0", width="1.67", height="0.50"];
        "Walter B. Hewlett" [pos="0,0", width="1.75", height="0.50"];
        "Robert L. Joss" [pos="0,0", width="1.47", height="0.50"];
        "Koh Boon Hwee" [pos="0,0", width="1.67", height="0.50"];
        "Agilent Technologies" [pos="0,0", width="2.03", height="0.50"];
        "Randall L. Tobias" [pos="0,0", width="1.75", height="0.50"];

In this case it was a two day run so I was hoping to get some useful output.

Maybe I misunderstand?
Comments:
[erg] dot's not so clean any more either. On typing an interrupt:


s) ./dot -m
dot: emit.c:1145: init_job_viewport: Assertion `((g)->u.bb.LL.x == 0) && ((g)->u.bb.LL.y == 0)' failed.
Abort
s)
In any case, it is the generic signal mechanism that is broken. All but dot need to transfer any position information from ND_pos to ND_coord.
Owner: *
Status: *