Re:[Ibm-netrexx] Bugs (was :Puzzle)

[billfen]

Yes, MFC's ability to write error free code is truly amazing.  The only
time I ever saw anything that resembled a bug in his code (20+ years ago)
was actually a PL/I compiler optimization issue with a null pointer test.

If he could detail his approach, techniques, etc. in a book, it should be
recommended reading for all programmers, and might save the industry
countless $.  Maybe there is no magic except hard work (and brilliance),
but perhaps he knows and practices things the rest of us do not.

In my code comments I questioned if the problem was my understanding of how
+0 in a parse template works or an actual bug.  I would always assume that
Mikes code is more likely to be correct than my understanding.

Still, bugs are always possible - just not yet found.  But your point is
well taken.  I probably should have said "... , since changes to NetRexx
will probably not occur soon, if ever :("


On 4/29/2011 4:32 AM, Marc Remes wrote:
> >>Not that it matters, since bugs in
> >>NetRexx will probably not be fixed anytime soon, if ever :(
>
> I have not yet encountered a bug in NetRexx, unless perhaps Thomas'
strange unwanted .nrx~ compiling, which to me looks more like an OS bug
than a NetRexx bug.
>
> Thomas, you mention you get - not more precisely documented - compilation
issues with very large projects. Could that be related to the java heap
running OutOfMemory? Try adding -Xms32m -Xmx512m (or more) to the java
invocation in NetRexxC.sh.

>
> Best regards
>
> Marc Remes
>
> IBM Certified IT Specialist
> IBM Global Technology Services
> Mobile: [phone]
> mailto:[hidden email]
>
>
>
> Tenzij hierboven anders aangegeven: / Sauf indication contraire

ci-dessus: / Unless otherwise stated above:
>
> International Business Machines of Belgium sprl / bvba
> Siège social / Maatschappelijke zetel: Avenue du Bourget 42 Bourgetlaan,
B-1130 Bruxelles/Brussel
> N° d'entreprise / Ondernemingsnr: TVA / BTW BE 0405 912 336
> RPM Bruxelles / RPR Brussel
>
>
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]

--------------------------------------------------------------------
mail2web - Check your email from the web at
http://link.mail2web.com/mail2web



_______________________________________________
Ibm-netrexx mailing list
[hidden email]

[Patric Bechtel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Bill,

as much as I appreciate MFC's abilities, but NetRexx is not error free.
It's as error free as any software project of any reasonable size...

The first three errors we found in the compiler were fixed within a few
weeks (that was in 2000); the next badge of errors, some of which are
quite annoying, was never fixed. No, not the kind of errors Thomas
found, these are silly. I mean real errors, like the unability to
resolve dependencies in large number of files within a directory
depending upon the sorting of the files by the file system. Or the
ability to resolve the right signature of a function in deeply nested
interface structures (like java.lang.StringBuilder). The ability to
resolv minor static classes (like javax.sound.LineEvent.Type). Errors
which make you redesign perfectly fine program structures to work around
a compiler bug.

I could continue here, but I just want to stop the impression, NetRexx
was almost perfect. It's not. I love NetRexx, though, but the continued
stall of development is something I'm working around for about a decade
now. And I've written almost 800k lines of NetRexx code meanwhile (I
don't really know the number, but that's the net amount of lines
currently, excluding the edits).

The bottom line is: NetRexx has bugs, and they need to be fixed. It has
missing features, which needs to be fixed. Soon. And for my
understanding it doesn't help if we pretend that NetRexx is bug free or
even almost bug free. As if there's no real pressure to open source it.

The pressure it there. Let's not ignore it. We don't have time to waste.
Not. Any. More.

[hidden email] schrieb am 29.04.2011 18:31:

> Yes, MFC's ability to write error free code is truly amazing.  The only
> time I ever saw anything that resembled a bug in his code (20+ years ago)
> was actually a PL/I compiler optimization issue with a null pointer test.
>
> If he could detail his approach, techniques, etc. in a book, it should be
> recommended reading for all programmers, and might save the industry
> countless $.  Maybe there is no magic except hard work (and brilliance),
> but perhaps he knows and practices things the rest of us do not.
>
> In my code comments I questioned if the problem was my understanding of how
> +0 in a parse template works or an actual bug.  I would always assume that
> Mikes code is more likely to be correct than my understanding.
>
> Still, bugs are always possible - just not yet found.  But your point is
> well taken.  I probably should have said "... , since changes to NetRexx
> will probably not occur soon, if ever :("
>
>
> On 4/29/2011 4:32 AM, Marc Remes wrote:
>>>> Not that it matters, since bugs in
>>>> NetRexx will probably not be fixed anytime soon, if ever :(
>>
>> I have not yet encountered a bug in NetRexx, unless perhaps Thomas'
> strange unwanted .nrx~ compiling, which to me looks more like an OS bug
> than a NetRexx bug.
>>
>> Thomas, you mention you get - not more precisely documented - compilation
> issues with very large projects. Could that be related to the java heap
> running OutOfMemory? Try adding -Xms32m -Xmx512m (or more) to the java
> invocation in NetRexxC.sh.

- --
cu, Patric
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: GnuPT 2.5.2

iEYEARECAAYFAk27L9sACgkQfGgGu8y7ypBX3wCeLU55FaPYcwNjgdRDM6osF12Q
XzsAoOax/+6kRLYzW7bGtkFvnXsMaDwz
=4t4T
-----END PGP SIGNATURE-----
_______________________________________________
Ibm-netrexx mailing list
[hidden email]

[ThSITC]

Hello Patrick,
     thanks for you info :-)

I do think that some of the problems I do have are related to the
dependencies
issue as well.

I'l have to say that I did port my software now so many times, and that
it is still not
fully object oriented.

What I do (personally) find annoying, thet we do have a languge NetRexx
and nobody
seem's to be responsible/capable to improve it.

I, for my part, am going my own way, now ;-)
Thomas.
====================================================================
Am 29.04.2011 23:38, schrieb Patric Bechtel:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hi Bill,
>
> as much as I appreciate MFC's abilities, but NetRexx is not error free.
> It's as error free as any software project of any reasonable size...
>
> The first three errors we found in the compiler were fixed within a few
> weeks (that was in 2000); the next badge of errors, some of which are
> quite annoying, was never fixed. No, not the kind of errors Thomas
> found, these are silly. I mean real errors, like the unability to
> resolve dependencies in large number of files within a directory
> depending upon the sorting of the files by the file system. Or the
> ability to resolve the right signature of a function in deeply nested
> interface structures (like java.lang.StringBuilder). The ability to
> resolv minor static classes (like javax.sound.LineEvent.Type). Errors
> which make you redesign perfectly fine program structures to work around
> a compiler bug.
>
> I could continue here, but I just want to stop the impression, NetRexx
> was almost perfect. It's not. I love NetRexx, though, but the continued
> stall of development is something I'm working around for about a decade
> now. And I've written almost 800k lines of NetRexx code meanwhile (I
> don't really know the number, but that's the net amount of lines
> currently, excluding the edits).
>
> The bottom line is: NetRexx has bugs, and they need to be fixed. It has
> missing features, which needs to be fixed. Soon. And for my
> understanding it doesn't help if we pretend that NetRexx is bug free or
> even almost bug free. As if there's no real pressure to open source it.
>
> The pressure it there. Let's not ignore it. We don't have time to waste.
> Not. Any. More.
>
> [hidden email] schrieb am 29.04.2011 18:31:
>> Yes, MFC's ability to write error free code is truly amazing.  The only
>> time I ever saw anything that resembled a bug in his code (20+ years ago)
>> was actually a PL/I compiler optimization issue with a null pointer test.
>>
>> If he could detail his approach, techniques, etc. in a book, it should be
>> recommended reading for all programmers, and might save the industry
>> countless $.  Maybe there is no magic except hard work (and brilliance),
>> but perhaps he knows and practices things the rest of us do not.
>>
>> In my code comments I questioned if the problem was my understanding of how
>> +0 in a parse template works or an actual bug.  I would always assume that
>> Mikes code is more likely to be correct than my understanding.
>>
>> Still, bugs are always possible - just not yet found.  But your point is
>> well taken.  I probably should have said "... , since changes to NetRexx
>> will probably not occur soon, if ever :("
>>
>>
>> On 4/29/2011 4:32 AM, Marc Remes wrote:
>>>>> Not that it matters, since bugs in
>>>>> NetRexx will probably not be fixed anytime soon, if ever :(
>>> I have not yet encountered a bug in NetRexx, unless perhaps Thomas'
>> strange unwanted .nrx~ compiling, which to me looks more like an OS bug
>> than a NetRexx bug.
>>> Thomas, you mention you get - not more precisely documented - compilation
>> issues with very large projects. Could that be related to the java heap
>> running OutOfMemory? Try adding -Xms32m -Xmx512m (or more) to the java
>> invocation in NetRexxC.sh.
> - --
> cu, Patric
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.10 (GNU/Linux)
> Comment: GnuPT 2.5.2
>
> iEYEARECAAYFAk27L9sACgkQfGgGu8y7ypBX3wCeLU55FaPYcwNjgdRDM6osF12Q
> XzsAoOax/+6kRLYzW7bGtkFvnXsMaDwz
> =4t4T
> -----END PGP SIGNATURE-----
> _______________________________________________
> Ibm-netrexx mailing list
> [hidden email]
>
>

--
Thomas Schneider (www.thsitc.com)
_______________________________________________
Ibm-netrexx mailing list
[hidden email]

[George Hovey-2]

Patric,
The received wisdom is that we are powerless to do anything but wait for IBM to move.  Do you have any other ideas?
George

On Fri, Apr 29, 2011 at 5:38 PM, Patric Bechtel <[hidden email]> wrote:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Bill,

as much as I appreciate MFC's abilities, but NetRexx is not error free.
It's as error free as any software project of any reasonable size...

The first three errors we found in the compiler were fixed within a few
weeks (that was in 2000); the next badge of errors, some of which are
quite annoying, was never fixed. No, not the kind of errors Thomas
found, these are silly. I mean real errors, like the unability to
resolve dependencies in large number of files within a directory
depending upon the sorting of the files by the file system. Or the
ability to resolve the right signature of a function in deeply nested
interface structures (like java.lang.StringBuilder). The ability to
resolv minor static classes (like javax.sound.LineEvent.Type). Errors
which make you redesign perfectly fine program structures to work around
a compiler bug.

I could continue here, but I just want to stop the impression, NetRexx
was almost perfect. It's not. I love NetRexx, though, but the continued
stall of development is something I'm working around for about a decade
now. And I've written almost 800k lines of NetRexx code meanwhile (I
don't really know the number, but that's the net amount of lines
currently, excluding the edits).

The bottom line is: NetRexx has bugs, and they need to be fixed. It has
missing features, which needs to be fixed. Soon. And for my
understanding it doesn't help if we pretend that NetRexx is bug free or
even almost bug free. As if there's no real pressure to open source it.

The pressure it there. Let's not ignore it. We don't have time to waste.
Not. Any. More.

[hidden email] schrieb am 29.04.2011 18:31:

> Yes, MFC's ability to write error free code is truly amazing.  The only
> time I ever saw anything that resembled a bug in his code (20+ years ago)
> was actually a PL/I compiler optimization issue with a null pointer test.
>
> If he could detail his approach, techniques, etc. in a book, it should be
> recommended reading for all programmers, and might save the industry
> countless $.  Maybe there is no magic except hard work (and brilliance),
> but perhaps he knows and practices things the rest of us do not.
>
> In my code comments I questioned if the problem was my understanding of how
> +0 in a parse template works or an actual bug.  I would always assume that
> Mikes code is more likely to be correct than my understanding.
>
> Still, bugs are always possible - just not yet found.  But your point is
> well taken.  I probably should have said "... , since changes to NetRexx
> will probably not occur soon, if ever :("
>
>
> On 4/29/2011 4:32 AM, Marc Remes wrote:
>>>> Not that it matters, since bugs in
>>>> NetRexx will probably not be fixed anytime soon, if ever :(
>>
>> I have not yet encountered a bug in NetRexx, unless perhaps Thomas'
> strange unwanted .nrx~ compiling, which to me looks more like an OS bug
> than a NetRexx bug.
>>
>> Thomas, you mention you get - not more precisely documented - compilation
> issues with very large projects. Could that be related to the java heap
> running OutOfMemory? Try adding -Xms32m -Xmx512m (or more) to the java
> invocation in NetRexxC.sh.

- --
cu, Patric
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: GnuPT 2.5.2

iEYEARECAAYFAk27L9sACgkQfGgGu8y7ypBX3wCeLU55FaPYcwNjgdRDM6osF12Q
XzsAoOax/+6kRLYzW7bGtkFvnXsMaDwz
=4t4T
-----END PGP SIGNATURE-----

_______________________________________________
Ibm-netrexx mailing list
[hidden email]

_______________________________________________
Ibm-netrexx mailing list
[hidden email]