Pedants

[Richard]

On 23 Jun 2008 at 21:06, Eric Sosman wrote:
jacob navia wrote:
[...]
I have tried to initiate such discussions, to show code for a
general container library for instance.

There wasn't a SINGLE comment about my code.
I don't recall seeing it, but I've seen a lot of things I
don't recall. Perhaps I saw it and thought "Good job, Jacob;
nothing I can add."

Call my a cynic, but I find it hard to imagine you have a positive
thought about something written by Jacob.

This may be what Jacob is referring to:
http://groups.google.com/group/comp.lang.c/browse_frm/thread/6e5fdc8d...

Saying there were no comment is a /slight/ exaggeration, but the silence
from the "regulars" is deafening.

"Jacob never posts source code"
A search for that phrase on groups.google.com turns up

Your search - "Jacob never posts source code" - did
not match any documents.

I know it's been a long two days, but try stretching your memory back to
earlier in this very thread:
http://groups.google.com/group/comp.lang.c/msg/3cf5e4b855819f29

Thanks.

Please don't reply to trolls.
<snip>

And you are? Please take your net nannying and play with it
offline. Thank you.

 

[Keith Thompson]

Because "here" there weren't any bug reports.

1) All "bugs" posted here were ironic stuff to show how bad
my compiler is. For instance somebody posted a "bug report"
when he defined a dprintf function. It is not intrinsically
bad, and the correction is easy, but the general tone is:
"I am a new user and I can't understand"

Perhaps the intent was "ironic", or whatever. Since I can't read the
poster's mind, I can't be sure. But regardless of the poster's
intent, it was (if I recall correctly) a genuine bug.

In my opinion, your best response would have been to ignore the tone,
acknowledge the error, and indicate your intent to fix it.

But since (a) lcc-win32 has had a tendency to declare non-standard
identifiers in standard headers, and (b) you have a history of
reacting badly when this is pointed out, a troll who wants to start a
flame war can easily do so. And you take the bait.

Then, Heathfield and co will jump in as if they weren't
involved:

What do you mean by "as if" they weren't involved? If you're
insinuating that the poster was a sock puppet, please provide some
evidence to back up your accusation.

"Of course, that is a horrible compiler", etc.

Whom are you quoting?

2) No problems whatsoever with *real* problems, bad code
generation, etc. The only "problems" are cosmetic.

PEDANTS like those bugs, they make an implicit assertion
about "book knowledge" as cited in dictionary.com when
consulting the definition of "pedant".

I don't know what you're talking about here. Can you provide an
example?

3) Use of undocumented flags (like -ansic89) that they gather
by looking into the executable of linux or whatever.

Yes, somebody complained, for whatever reason, about the behavior of
lcc-win32 with "-ansic89". Most of us had no way of knowing that that
was an undocumented flag. If you had said so at the beginning of that
discussion, there would have been nothing to discuss.

Then,
Heathfield springs in and writes:
"That compiler doesn't conform to *any* standard"

Is he mistaken? If so, to what standard does it conform? If not,
what's the problem?

Since most gcc headers will not compile with other
compilers I could post HUNDREDS of those bugs here,
just to show that gcc is "bad". Obviously I will not
do that, I am not a pedant.

Nobody else would do that either, because it would be absurd. There's
no reason to assume that gcc-specific headers will compile with any
other compiler.

If some gcc-specific header didn't compile with gcc, or if some
lcc-specific header didn't compile with lcc, that would be a bug. In
either case, comp.lang.c might be an appropriate place to ask whether
it's a bug or not.

This group is lacking most of a real discussion about issues
in software development. Any such discussions will be killed by this
people with their "off topic" stuff. Here, only THEY can discuss
anything that goes beyond

int main (void);

or printf directives, or students homework.

Please don't exaggerate; it doens't help your case. ISO C is a big
topic, going well beyond "int main(void)" and printf directives.

I have tried to initiate such discussions, to show code for a
general container library for instance.

There wasn't a SINGLE comment about my code.

In that code I showed a software for developing general containers
with the example of a resizable array.

Nothing.

I'm afraid I don't remember that posting. Can you provide a URL or
Message-ID?

NONE of the people that now tell (in this same group)

"Jacob never posts source code"

Whom are you quoting?

 

[Ben Bacarisse]

Because "here" there weren't any bug reports.

Bug reports would be off topic here so people who are both c.l.c savvy
and who care about topicality will not post them here.

1) All "bugs" posted here were ironic stuff to show how bad
my compiler is. For instance somebody posted a "bug report"
when he defined a dprintf function. It is not intrinsically
bad, and the correction is easy, but the general tone is:
"I am a new user and I can't understand"

There have been some seemingly disingenuous reports, I agree, but then
attributing motive to Usenet posts is very hard. There is a small
chance some of them were genuine.

2) No problems whatsoever with *real* problems, bad code
generation, etc. The only "problems" are cosmetic.

Maybe not here, but I think there may be some serious issues. The
trouble is that if you don't say what your compiler supports, how can
anyone know if it is a bug or not? A number of reports I made
(topically, I think, to comp.compilers.lcc) were answered with "not
implemented". That is fine, but it made me not post several others.
In some sense, every compiler bug is just something that is "not
implemented".

I am not being pedantic about conformance here. A simple list of the
those things that you think are "done" and those that you know need
more work is all that is required. There is little motivation to
prepare a cut-down example to illustrate a potential problem if it is
likely that you already know that you have "not done that bit yet".
For example, if you have implemented C99's ... syntax for
function-like macros, then you have an error in the implementation.
If that is currently "not implemented" they you don't.

I occurs to me now that you might think I am one of these disingenuous
posters. This might explain why my post (again to comp.compilers.lcc)
about VLA parameters went un-answered[1]. If I am honest, I suspect
my motives for reporting bugs are not entirely honourable. I get a
buzz from finding them that has something to do with feeling clever,
but does that mean the information is any less useful to you? If it
helps, I would like to be able say there is another C99 compiler out
there. I am genuinely disappointed at the lack of C99 support.

[1] Message-ID: <[email protected]> The rather feeble "The code
looks fine to me" is just an expression of the fact I don't consider
myself an expert on VLAs, it was not my being disingenuous.

 

[christian.bau]

Since most gcc headers will not compile with other
compilers I could post HUNDREDS of those bugs here,
just to show that gcc is "bad". Obviously I will not
do that, I am not a pedant.

That would just demonstrate that you are clueless. Standard header
files are part of the implementation. The combination of compiler and
standard header file together must produce the effect described by the
C Standard, and the gcc compiler + header files achieve exactly that.
It is _not_ a bug if a file that is part of the implementation cannot
be used with another compiler.

On the other hand, if including a standard header file defines
functions that are in the application name space, that is a serious
bug. gcc + gcc header files don't do that kind of thing.

 

[santosh]

Lets collect those denigrators. Will each one who is determined to
denigrate anything done by Jacob Navia, even if it complies with
the C standard, please answer this. You may also deny. I suggest
ticking one of the following two lines:

Denigrator ........ YES OUI JA
Denigrator ........ NO NON NEIN

YES of course.

 

[santosh]

And he does. However, he neglects to bother to include any means
of identifying the version at hand, equivalent to using "gcc -v" on
gcc. Unless he has changed this practice, in which case I
apologize fo complaining.

Try lcc.exe -v

 

[Serve Lau]

The post cited as an example does not use the quoted phrase,
so if that's the post you intended you've misquoted it. But I
don't think you need get too upset about that post: consider the
source, and decide how much attention is merited.


Your search - "Jacob is paranoic" - did not match any documents.

After correcting the spelling error

Your search - "Jacob is paranoiac" - did not match any documents.

I dont read clc very often but even I know that that is what people say.
Maybe not literally but the regulars sure do. So I'm not sure why you have
to play games now. A little more respect for somebody maintaining a windows
C compiler would not hurt. I'm sure most regulars understand how hard it is
to keep adding features to a software system like this and not introduce
bugs.

 

[jacob navia]

Ben Bacarisse wrote:
... snip ...

And he does. However, he neglects to bother to include any means
of identifying the version at hand, equivalent to using "gcc -v" on
gcc. Unless he has changed this practice, in which case I
apologize fo complaining.

yes yes mr falconer
lcc -v iss too much for you

or just
lcc

that prints all the options including -v

All the versions of lcc are timestamped.

 

[Nick Keighley]

jacob navia wrote:


Lets collect those denigrators.  Will each one who is determined to
denigrate anything done by Jacob Navia, even if it complies with
the C standard, please answer this.  You may also deny.  I suggest
ticking one of the following two lines:

   Denigrator  ........  YES OUI JA
   Denigrator  ........  NO  NON NEIN

what exactly are you trying to prove?

- that you can be funny/ironic/sarcastic?
- that Jacob is wrong, no on eis denegrating him?
- a laugh at Jacob's expense?

Jacob's major problems are that he doesn't like criticism and he
bites
too easily.

 

[Joachim Schmitz]

So rather than admit that you've slandered Jacob, you prefer to change
the problem specification. I've worked for people like you before.
Here's what you said:

No mention of /where/ you want to see Jacob having been polite.

By posting this in CLC and in CLC only, it's crystal clear what I meant. I
don't read or write in comp.compilers.lcc, so can't possible comment on
that.

- if you care then do your own damn research instead of moving the
goalposts.

If I'd care, indeed...

 

[vippstar]

By posting this in CLC and in CLC only, it's crystal clear what I meant. I
don't read or write in comp.compilers.lcc, so can't possible comment on
that.


If I'd care, indeed...

Please don't feed the trolls.

 

[Joachim Schmitz]

Please don't feed the trolls.

Please let me decide myself whom to feed, where and when.

Bye, Jojo

 

[Antoninus Twink]

Off-hand I don't recall anyone calling Jacob paranoid (or a
variation).

I don't believe you. Within the last 24 hours, someone has called Jacob
paranoid in this group
(http://groups.google.com/group/comp.lang.c/msg/09caf140f96977e8), and
searching Google groups for "jacob paranoid" spews out a string as long
as your arm of similar abuse in the past.

I don't really understand you, Eric. You're clearly an intelligent and
experienced programmer who's willing to post thoughtful analyses of
people's code, and pass on the benefit of your knowledge in a lively and
often amusing way. You also often come up with interesting historical
anecdotes and perspectives that help make the group more 3-dimensional,
so to speak.

But then there's some part of you that just can't resist being an
asshat. What fun do you get from playing these stupid word games,
wallowing in Heathfield-esque pedantry, being deliberately obtuse, etc.?

 

[Ralf Damaschke]

Because "here" there weren't any bug reports.

1) All "bugs" posted here were ironic stuff to show how bad
my compiler is. For instance somebody posted a "bug report"
when he defined a dprintf function. It is not intrinsically
bad, and the correction is easy, but the general tone is:
"I am a new user and I can't understand"
Then, Heathfield and co will jump in as if they weren't
involved:
"Of course, that is a horrible compiler", etc.

I remember that I had a word or two with you on the issue that a
standard function to be declared in one standard header should not be
spreaded to other standard headers without explicit (or implicit by
the as-if rule) permission by the standard. To me this dprintf issue
(which I did not follow, I spent my popcorn for other newsgroups this
time) sounds like there are still similar bugs in lcc[-win]. Your
answer in the first mentioned thread was that there is a compiler
option to swich that off. I have no idea why this should not be the
cure for similar issues.

2) No problems whatsoever with *real* problems, bad code
generation, etc. The only "problems" are cosmetic.

PEDANTS like those bugs, they make an implicit assertion
about "book knowledge" as cited in dictionary.com when
consulting the definition of "pedant".

It's the C standard (well, one of them) which defines the C language,
so it's really required to conform to it (or try at best).

3) Use of undocumented flags (like -ansic89) that they gather
by looking into the executable of linux or whatever. Then,
Heathfield springs in and writes:
"That compiler doesn't conform to *any* standard"

No problem here if "-ansic" (that you have suggested formerly) makes
the compiler standard compliant (to whatever you choose; I won't
really mind a bastard if documented properly when there is another
way to restrict the compiler to a specific standard).

Ralf

 

[Keith Thompson]

Lets collect those denigrators. Will each one who is determined to
denigrate anything done by Jacob Navia, even if it complies with
the C standard, please answer this. You may also deny. I suggest
ticking one of the following two lines:

Denigrator ........ YES OUI JA
Denigrator ........ NO NON NEIN

I'm not interested in playing this game.

 

[Keith Thompson]

Please let me decide myself whom to feed, where and when.

Nobody is stopping you. Advice is not censorship.

 

[Antoninus Twink]

I'm not interested in playing this game.

It's hard to imagine who would be.

It's pointless anyway - it's not as if we need someone to publicly
declare "I denigrate Jacob" - we can look at their posts and form a
judgment on their behavior by ourselves. Actually, I suspect that many
of the worse offenders genuinely don't believe that they're guilty of
bullying Jacob, in much the same way as very few alcoholics have the
self-awareness to admit that they have a drink problem.

 

[Kaz Kylheku]

Dear pedantic user

What is a pedant?

According to dictionary.com you are:

1.      a person who makes an excessive or inappropriate display of learning.
2.      a person who overemphasizes rules or minor details.
3.      a person who adheres rigidly to book knowledge without regard to
common sense.

So an example of 3 would be someone who thinks that it's okay for a
repeatedly executed statement block that contains a variable-length
array to leak memory, because the ISO C standard (book knowledge)
doesn't forbid that behavior, whereas common sense says that the
storage should be liberated so that there is no leak.

I am very glad that this flag, done for people like you
works as expected.

I can't see how a flag that causes a compiler to choke on its own
headers is of any use to anyone.

My compiler system is not for pedants, so you can stop using it

But non-pedants write code that relies on extensions (often without
realizing it,
sometimes with good reason).

And the extensions they are relying on are usually not lcc
extensions.

Only a pedant who is currently using another compiler is likely to
have any significant code base that works with lcc without
modification.

(Part of the reason is that he compiles with the most pedantic
diagnostic options of his favorite compiler).

Maybe you should reconsider the target user base?

Don't you think that a pedantic user base is good to have?

Think about the quality and quantity of bug reports from non-pedants
versus pedants. The worst non-pedants just shrug and massage their
code until a problem goes away, blaming themselves.

Ah right, you take bug reports as personal insults. Never mind.

and get a compiler that suits your pedantic needs. Many pedants
here (this group has a lot of them) will point you to their
favorite software.

Machines are pedantic, too. One bad bit, and the show is over.

 

[user923005]

Dear pedantic user

What is a pedant?

According to dictionary.com you are:

1.      a person who makes an excessive or inappropriate display of learning.
2.      a person who overemphasizes rules or minor details.
3.      a person who adheres rigidly to book knowledge without regard to
common sense.

I am very glad that this flag, done for people like you
works as expected.

My compiler system is not for pedants, so you can stop using it
and get a compiler that suits your pedantic needs. Many pedants
here (this group has a lot of them) will point you to their
favorite software.

For testing software and describing problems, there is *nothing*
better than a pedant.
A grammar error in a help file is a serious problem. Someone may try
to use it and misunderstand. In any case, it makes a bad impression.
If your toolset is to be taken seriously, you should be ultra-anal
about correctness. A software product that melts into errors under
the browbeating of pedants needs to be corrected.
Why (for instance) should someone compute the sin() function to over
100 digits {as, for instance, LCC does} unless they are ultra-
concerned about correctness? Certainly, the last 70 digits or so have
literally no practical value (though they will have a certain
mathematical beauty).

I think that you would be more successful if you tried to be nicer
{the origination of this thread (for instance) was an intended jab at
those who aim for correctness in this forum}. I also know a chess
programmer who is very talented, but incredibly ascerbic. He has
tried for years to create a commercial product, but has not been able
to succeed. The problem in his case is not lack of talent but lack of
people skills. Could this shoe fit more than one person?

IMO-YMMV.

 

[Richard]

And then, once you have worked out why you are wrong, please do the
necessary.

It would seem that not only does Heathfield see himself as some kind of
elder statesman and C god, he also sees himself as judge an jury. If a
day could go buy without him telling people (even Chuck) when to
apologise and to whom then the better for this group. The mind boggles
at his own self perceived importance.