delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp-workers/1998/03/30/13:56:24

Message-Id: <m0yJi7C-000S3DC@inti.gov.ar>
Comments: Authenticated sender is <salvador AT natacha DOT inti DOT gov DOT ar>
From: "Salvador Eduardo Tropea (SET)" <salvador AT inti DOT gov DOT ar>
Organization: INTI
To: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>, djgpp-workers AT delorie DOT com
Date: Mon, 30 Mar 1998 14:16:33 +0000
MIME-Version: 1.0
Subject: Re: Auto-symified traceback
References: <m0yJghG-000S2kC AT inti DOT gov DOT ar>
In-reply-to: <Pine.SUN.3.91.980330193929.20942A-100000@is>

Eli Zaretskii <eliz AT is DOT elta DOT co DOT il> wrote:
> On Mon, 30 Mar 1998, Salvador Eduardo Tropea (SET) wrote:
> 
> > How do you plan to implement it? calling it from the destroyed program?
> 
> If you mean calling `symify.exe', then no, I didn't think about that.  I 
> thought about copying some of the `symify' code into the library 
> function which prints the traceback.  The code is actually quite simple.

But then the overhead won't be so small, and the same applies: What if the 
program destroys some function needed for that making the the GPF even worst?
In my editor I save all the un-saved stuff when I get the signal, until now I 
got only benefits, but I'm not so sure about how secure is it. The feature can 
be disabled from the command line.
 
> > If the program crashes in the machine of an end user (and beleive me I saw tons 
> > of djgpp programs distributed with debug info) it won't help.
> 
> It was meant as an aid to someone who is debugging the program on the same
> machine where it was developed.  

In this case the user can use RHIDE or other tool to get it automatically, a 
simple batch file can do it (looking at the return code).

> But I also don't understand why are you
> saying that it won't help on other machines.  

I never stated: other machine, I said: the end user.

> `symify' doesn't need
> sources to print the info, so the file names and line numbers will be
> printed even if the sources are long gone, or were never installed.  It is
> much easier to deal with a crash report when it has this info attached,
> even if it crashed on someone else's machine.  For example, imagine the
> situation where a program I wrote crashed at some other machine, while the
> sources I maintain have already changed.  A symified traceback will help a
> whole lot more in this situation. 

1) I supposed you tried to make it using symify so the user must have symify.
2) I ever have copies of the last 3 versions with debug info, so I don't need 
to force the user to download a file with the debug info. My editor is 5Mb with 
debug info and only 327Kb without it (and compressed). I preffer to run symify 
on my copy.

> > Experimented 
> > users can run symify or use a custom methode and RHIDE users (a lot of newbies) 
> > gets it from RHIDE (It traps the GPF and shows the translated trace back). So 
> > doesn't look to be very needed.
> 
> You seem to only think about RHIDE users ;-).

Sorry Eli, sometimes I get biased. But you must agree that most of the newbies 
(and they really need help with symify) uses RHIDE.
 
> There are situations when this could be useful.  

I don't doubt it. Is just that the code bloat isn't good and the cases where it 
could help aren't so much.

> One such case is when a 
> program which you already use for a long time crashes once in a while.  
> If you are unlucky enough and the properties of that program aren't set 
> to not close the DOS box, the traceback is gone after you click "OK".  A 
> symified traceback at least lets you see where did it happen before it 
> disappears for good.

I never start programs in this way (I guess you are making it from a shortcut 
in the W95 desktop, is that correct?).

> Anyway, I won't force anybody to have that feature.  If it isn't 
> required, I have better things to do...

Perhaps it could be done as an add-on, I mean: It's included by request of the 
programmer. For example with a #define before some header (expanded to a symbol 
that forces the module ...). But not as unconditional code.

SET 
------------------------------------ 0 --------------------------------
Visit my home page: http://set-soft.home.ml.org/
or
http://www.geocities.com/SiliconValley/Vista/6552/
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Alternative e-mail: set-soft AT usa DOT net set AT computer DOT org
ICQ: 2951574
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA
TE: +(541) 759 0013

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019