Trolltech Home | Qt-interest Home | Recent Threads | All Threads | Author | Date
All threads index page 1

Qt-interest Archive, April 2007
need something like QRandomCrash


Message 1 in thread

Hi,

I've got a slight problem with some of my customers: they are used to MFC 
applications that crash once in a while. The usual cafeteria talk used be 
along the lines "can you imagine - it crashed twice already 
today!" - "yeah, mine too!". Due to the more stable Qt programs the office 
climate is deteriorating rapidly ("how's your husband? Still working 
for..." - "Don't ask!"; or "..and your son? I heard he's got some trouble 
w..." - "shutup!"). We can't possible revert to the old software, so I'd 
like to remedy the problem in the new one.

Is anyone here aware of some class that crashes the application randomly 
once in a while? Could you point me to its source?

It would also be great if I could tune it to the users needs and configure 
the average and standard deviation of the crash frequency somewhere. The 
actual error message doesn't matter - over the years the users developed a 
reflex to click any message box away before the message can hurt their 
feelings.



	Konrad

Attachment:

Attachment: pgpfsF0VtbmPA.pgp
Description: PGP signature


Message 2 in thread

Hmm, don't you have a project leader? Usually it is his responsibility
to make sure a program crashes sufficiently. A good project leader can 
wonderfully fine tune be instability of a program by using general 
management tools like 'feature creep', 'overambitious time frames', or 
'changing requirements'. Ok, with Qt it is a little bit more difficult, 
but after a short time a project leader needs to  adjust to the 
greater robustness of Qt, those tools still work perfectly fine.

> It would also be great if I could tune it to the users needs and configure 
> the average and standard deviation of the crash frequency somewhere. The 
> actual error message doesn't matter - over the years the users developed a 
> reflex to click any message box away before the message can hurt their 
> feelings.

Guido


--
 [ signature omitted ] 

Message 3 in thread

Hmm... just wondering, are there any such management tools like that?
I mean, these kinds of 'feature creep', 'overambitious time frames', or 
'changing requirements'.

Doni


Guido Seifert wrote:
> Hmm, don't you have a project leader? Usually it is his responsibility
> to make sure a program crashes sufficiently. A good project leader can 
> wonderfully fine tune be instability of a program by using general 
> management tools like 'feature creep', 'overambitious time frames', or 
> 'changing requirements'. Ok, with Qt it is a little bit more difficult, 
> but after a short time a project leader needs to  adjust to the 
> greater robustness of Qt, those tools still work perfectly fine.

> Guido
>   
>
> --
> To unsubscribe - send a mail to qt-interest-request@xxxxxxxxxxxxx with "unsubscribe" in the subject or the body.
> List archive and information: http://lists.trolltech.com/qt-interest/
>
>
>   

--
 [ signature omitted ] 

Message 4 in thread

It is my experience that the managers can do this just fine without tools.

On Sunday 01 April 2007 12:09, Doni Andri Cahyono wrote:
> Hmm... just wondering, are there any such management tools like that?
> I mean, these kinds of 'feature creep', 'overambitious time frames', or
> 'changing requirements'.
>
> Doni
>
> Guido Seifert wrote:
> > Hmm, don't you have a project leader? Usually it is his responsibility
> > to make sure a program crashes sufficiently. A good project leader can
> > wonderfully fine tune be instability of a program by using general
> > management tools like 'feature creep', 'overambitious time frames', or
> > 'changing requirements'. Ok, with Qt it is a little bit more difficult,
> > but after a short time a project leader needs to  adjust to the
> > greater robustness of Qt, those tools still work perfectly fine.
> >
> > Guido
> >
> >
> > --
> > To unsubscribe - send a mail to qt-interest-request@xxxxxxxxxxxxx with
> > "unsubscribe" in the subject or the body. List archive and information:
> > http://lists.trolltech.com/qt-interest/
>
> --
> To unsubscribe - send a mail to qt-interest-request@xxxxxxxxxxxxx with
> "unsubscribe" in the subject or the body. List archive and information:
> http://lists.trolltech.com/qt-interest/

-- 
 [ signature omitted ] 

Attachment: pgpGlaRcMIwdr.pgp
Description: PGP signature


Message 5 in thread

Konrad Rosenbaum wrote:

> Hi,
> 
> I've got a slight problem with some of my customers: they are used to MFC
> applications that crash once in a while. The usual cafeteria talk used be
> along the lines "can you imagine - it crashed twice already
> today!" - "yeah, mine too!". Due to the more stable Qt programs the office
> climate is deteriorating rapidly ("how's your husband? Still working
> for..." - "Don't ask!"; or "..and your son? I heard he's got some trouble
> w..." - "shutup!"). We can't possible revert to the old software, so I'd
> like to remedy the problem in the new one.
> 
> Is anyone here aware of some class that crashes the application randomly
> once in a while? Could you point me to its source?
> 
> It would also be great if I could tune it to the users needs and configure
> the average and standard deviation of the crash frequency somewhere. The
> actual error message doesn't matter - over the years the users developed a
> reflex to click any message box away before the message can hurt their
> feelings.
> 
> 
> 
> Konrad

Oh, that's pretty straightforward, by subclassing QConstantCrash.
I do this every day ...

--
 [ signature omitted ] 

Message 6 in thread

Konrad Rosenbaum wrote:
> ...snip...
> Is anyone here aware of some class that crashes the application randomly
> once in a while? Could you point me to its source?
> ...snip...

I seem to have a LOT of classes like this, and would happily share the
source.  But in order to pick the right one, I need some clarification.

a) What do you mean by "randomly".  Would a pseudo-random event class based
on rand() be sufficient for your needs, or do you need a more rigorously
unpredictable event generator?

b) How long is "a while"?  You might try making the duration of "a while" be
itself a random value.

c) You do realize, I hope, that by definition, constraining supposedly
random events so that they occur only "once" within any given "while", they
become non-randon?  You should consider removing this constraint.

--
 [ signature omitted ] 

Message 7 in thread

On Sunday 01 April 2007, Larry Bristol wrote:
> Konrad Rosenbaum wrote:
> > ...snip...
> > Is anyone here aware of some class that crashes the application
> > randomly once in a while? Could you point me to its source?
> > ...snip...
>
> I seem to have a LOT of classes like this, and would happily share the
> source.  But in order to pick the right one, I need some clarification.

Wow. I didn't imagine that there are more than a few in the Qt universe!

> a) What do you mean by "randomly".  Would a pseudo-random event class
> based on rand() be sufficient for your needs, or do you need a more
> rigorously unpredictable event generator?

If it has a good seeding then yes, a pseudo random function will be fine - 
the users won't know the difference.

> b) How long is "a while"?  You might try making the duration of "a while"
> be itself a random value.

I'm terribly sorry. I should have defined this.

A while (or more correctly "eine Weile") is an old German imperial measure 
of time which can be anything between 5min 23sec and 2months, 1 week, 3 
days, 4:55:48 hrs. Although the latter is a bit on the extreme side, most 
of the time it was below one day, 50% of the time below one hour. What it 
was exactly heavily depended on state, the current states head (king, 
baron, duke, etc.pp.), his wife or, more often than not, his current 
mistress, and any number of arcane and subtle factors.

Since I'm from Saxony and a huge fan of August der Starke, I most often 
refer to a "Weile" being exactly 20min 9.6sec - today commonly known as a 
millifortnight - with the notable difference that in baroque times it 
differed by give or take 10 minutes depending on the time of day and the 
beauty of Augusts current female companion.


> c) You do realize, I hope, that by definition, constraining supposedly
> random events so that they occur only "once" within any given "while",
> they become non-randon?  You should consider removing this constraint.

Of course I meant "once" in the sense that any given instance of a program 
cannot crash twice at the exact same moment - although it might get two 
crash commands at the same time, it can only exit once (barring some freak 
implementation of a process scheduler).




	Happy april fools day,
	Konrad

Attachment:

Attachment: pgppFLBzilVRq.pgp
Description: PGP signature


Message 8 in thread

Konrad Rosenbaum wrote:

> Wow. I didn't imagine that there are more than a few in the Qt universe!
> ...<lot's of good stuff snipped>...

A friend of mine once stated, "If ever anyone actually SHOULD figure out the
answer to the ultimate question as to the meaning of life, and understood
exactly why we are here and what we should do, the universe would instantly
transform itself into something even more bizarre and inexplicable."

Apparently, this happened once again early yesterday.

--
 [ signature omitted ] 

Message 9 in thread

I don't understand what the fuss is about.
Any programmer should know there is "rand()" ( and variants)
and there is "(int *)NULL = 42;"

Anything more you need????

-- Gert.


> -----Original Message-----
> From: qt-interest-request@xxxxxxxxxxxxx
> [mailto:qt-interest-request@xxxxxxxxxxxxx]On Behalf Of Konrad Rosenbaum
> Sent: 01 April 2007 08:21
> To: qt-interest@xxxxxxxxxxxxx
> Subject: need something like QRandomCrash
>
>
> Hi,
>
> I've got a slight problem with some of my customers: they are used to MFC
> applications that crash once in a while. The usual cafeteria talk used be
> along the lines "can you imagine - it crashed twice already
> today!" - "yeah, mine too!". Due to the more stable Qt programs
> the office
> climate is deteriorating rapidly ("how's your husband? Still working
> for..." - "Don't ask!"; or "..and your son? I heard he's got some trouble
> w..." - "shutup!"). We can't possible revert to the old software, so I'd
> like to remedy the problem in the new one.
>
> Is anyone here aware of some class that crashes the application randomly
> once in a while? Could you point me to its source?
>
> It would also be great if I could tune it to the users needs and
> configure
> the average and standard deviation of the crash frequency somewhere. The
> actual error message doesn't matter - over the years the users
> developed a
> reflex to click any message box away before the message can hurt their
> feelings.
>
>
>
> 	Konrad
>


--
 [ signature omitted ]