[prev in list] [next in list] [prev in thread] [next in thread] 

List:       taskjuggler
Subject:    Re: [taskjuggler] tjx2gantt problem
From:       Stefan Janecek <stefan.janecek () jku ! at>
Date:       2003-10-17 11:27:23
[Download RAW message or body]


On Fri, 2003-10-17 at 09:00, Remo Behn wrote:
> * Stefan Janecek <stefan.janecek@jku.at> [031016 16:25]:
> > hi taskjuggler people,
> > 
> > I am using taskjugler 1.9.2 and recently encountered a strange problem
> > with tjx2gantt. After doing minor changes in my project file, tjx2gantt
> > prints the following error messages:
> > 
> > Use of uninitialized value in subtraction (-) at
> > /usr/local/bin/tjx2gantt line 303.
> > Use of uninitialized value in subtraction (-) at
> > /usr/local/bin/tjx2gantt line 309.
> > Use of uninitialized value in subtraction (-) at
> > /usr/local/bin/tjx2gantt line 309.
> > Use of uninitialized value in subtraction (-) at
> > /usr/local/bin/tjx2gantt line 318.
> > 
> > and gv also complains about some errors in the resulting postscript, but
> > still displays the chart without any of the text. According to my RCS
> > diff, the only difference between the working version of the project
> > file and the one resulting in the errors are:
> > - one task was removed
> > - some 'complete 0' directives have been changed to 'complete 100'
> > (finally ;-)
> > Thanks god I am keeping my project files in RCS....
> > 
> > I have to admit I know nothing about perl (and after reading perl code,
> > moreover consider it to be an ugly language), but tried to shed a bit of
> > light onto the problem by adding the following statement in line 294:
> > 
> >    print "t=$t, x1=$x1, y1=$y1, task_height=$task_height\n---------\n";
> >  
> > The following lines seem to be relevant to the problem:
> > 
> > t=exams.groupth, x1=465, y1=57.5, task_height=5
> > ---------
> > Use of uninitialized value in concatenation (.) or string at
> > /usr/local/bin/mytjx2gantt line 294.
> > Use of uninitialized value in concatenation (.) or string at
> > /usr/local/bin/mytjx2gantt line 294.
> > Use of uninitialized value in concatenation (.) or string at
> > /usr/local/bin/mytjx2gantt line 294.
> > t=ex, x1=, y1=, task_height=5
>         ^^^^^^^^^
> 
> the task has no length, no start, no end ???

In the project file, its length is defined, and start time is defined by
depending on its predecessor: 

>    task thermo "Thermodynamik" {
>        depends !groupth
>        effort 160h
>        complete 0
>    }

The relevant part of the tjx-file is:

   <Task Id="exams.thermo" >
    <Index>44</Index>
    <Name>Thermodynamik</Name>
    <ProjectID>main</ProjectID>
    <Priority>500</Priority>
    <complete>0.0</complete>
    <Type>Task</Type>
    <ParentTask>exams</ParentTask>
    <actualStart humanReadable="2003-12-16 15:00 CET"
>1071583200</actualStart>
    <actualEnd humanReadable="2004-01-09 19:00 CET"
>1073671200</actualEnd>
    <planStart humanReadable="2003-12-16 15:00 CET"
>1071583200</planStart>
    <planEnd humanReadable="2004-01-09 19:00 CET" >1073671200</planEnd>
    <Previous>exams.groupth</Previous>
    <Follower>exams.hltech</Follower>
    <Follower>exams.milestones.finalbest</Follower>
    <Allocation ResourceID="me_uni" >
     <Load>0</Load>
     <Persistent>No</Persistent>
    </Allocation>
    <Resource Id="me_uni" >Arbeitszeit Uni</Resource>
   </Task>

which seems to look quite OK.
> 
> send me your ???.xml for examination ...

I'll send you the whole tjx file in private mail.

> 
> taskjuggler version : ???
> on plattform        : ???
taskjuggler 1.9.2, on Debian Linux testing 'sarge'

Meanwhile, I probably found out something new: I renamed the Resource Id
'me_uni' to 'meuni', and now things are working again. However, I've
used the underscore in the Resource Id since ever, without encountering
any problems so far....

One explanation that comes to my mind is that per chance some perl
module tjx2gantt is depending on got updated by the last 'apt-get
update' (remember, I am using this on a Debian machine), breaking 
tjx2gantt. Whether it also fails on my tjx file on another machine or
not could prove or rebut this theory - so I am curious to hear about
your results ...


many thanks,
Stefan



-- 
"We have joy, we have fun, we have Linux on a Sun ..."
__________________________________________________________________________
Stefan Janecek
Institute of Semiconductor & Solid State Physics
Universtity of Linz/Austria                 -----==-
work phone: +43 (0)732 2468-9612            ----==--
Fax         +43 (0)732 2468-8650            ---==---(_)__  __ ____  __
mail: stefan.janecek@jku.at                 --==---/ / _ \/ // /\ \/ /
                                            -=====/_/_//_/\_,_/ /_/\_\
ICQ:  segfault/#52065279                    The choice of a GNU generation
--------------------------------------------------------------------------
GPG: 1024D/F37F4243 2002-05-07 Stefan Janecek <stefan.janecek@jku.at>
     Key fingerprint = F9A7 AC8E BF0D FF85 17A2  8E88 8955 A3F0 F37F 4243

["signature.asc" (application/pgp-signature)]

[prev in list] [next in list] [prev in thread] [next in thread]