Re: Luxasm news

Johannes Kroll <j-krollNO-SPAM-BLAH-BLAH@xxxxxx> écrivait news:d5d8it$98h
$1@xxxxxxxxxxxxxxxxxx:

> You mean if RosAsm crashes the error must be in the user's source code?

No. I mean that, once a tool is considered "pretty
stable", whatever that means, when the whole thingie
hangs, this is much probably not with a trivial and
standard Source, and that, in most case, this will
be because of some weird and unregular use of something
in the users's Source.

The fact that such events help at pointing out real
bugs, in the programming tool is not the problem, and
is another story. The real problem is that, when
implementing out-of-reason securities for the user
typings, you put the user's source in danger, because
the worst enemy of a source is the programmer himself,
and certainaly not the Tool he is using.

For RosAsm, we have:

* Source saved into the Executable, at each successful
Compilation, what takes almost no time, so that the
user could verify after a couple of Lines that his
Source is a valid one.

* The [Ctrl]/[S] feature, that offer 3 different Source(s)
saving Options.

* The [Ctrl]/[K] feature, that is an Automatic Back-Up
with a complete naming Management, and with user defintion
of number of Back-Ups (for deletion of the older version
and renaming of the recent ones...)

* The IncInclude Pre-Parser that does sophisticated
updates of the Included Files, back and forth.

* The final Exception Handler that saves the whole Source
to disk before the hangs shut-down.

Well... i _never_ use any of these (but point 1, of course).
Maybe because i am an old *** who payed to learn how to
keep his sources alive under DOS, at times when each small
error was equal to a hard re-boot, by the Power Button...,
but i am quite sure that the younger RosAsm users might feel
happy with all of those securities, and that they certainaly
do not need to save the errors they introduce into their own
Sources, when the result of these faults might push the Tool
to hang.

This is not a problem of "Whose fault?". The problem is that,
as long as the bug, inside the programming Tool, will not be
fixed, the Source would be made simply _unusable_. Way simpler
is to effectively loose the last "faultive or not faultive"
modifications of the Source, and to restart writing something
that the Tool could "eat".


Betov.

< http://rosasm.org >






.