Oops, 5.8.1 broke my program

[Xaonon]

Or, more likely, it was broken before in some way that wasn't apparent until
I installed 5.8.1. The program in question is a signature generator, which
runs as a daemon and produces random quips via a named pipe. The main loop
looks like this:

while( 1 )
{
-p $options{outfile}
or die "$0: `$options{outfile}' disappeared\n";
sysopen FIFO, $options{outfile}, O_WRONLY
or die "$0: $options{outfile}: $!\n";

print FIFO randsig();
close FIFO or die "$0: $options{outfile}: $!\n";

# Sleep briefly to let the reader see EOF.
select undef, undef, undef, 0.1;
}

Now, though, it seems that the reader doesn't ever see an EOF. For example,
running "cat sig.fifo" spits out signature after signature until I kill the
process, as opposed to just printing one and finishing. I note that running
the program with 5.6.1 still works, but I want a more rigorous solution.
Can anyone identify the problem?

 

[Wally Sanford]

Xaonon ([email protected]) wrote on MMMDCLXXXII September MCMXCIII in
<URL:{} Or, more likely, it was broken before in some way that wasn't
apparent until {} I installed 5.8.1. The program in question is a
signature generator, which {} runs as a daemon and produces random
quips via a named pipe. The main loop {} looks like this:
{}

[Snip more NON STANDARD quoting...]

Abigail, coudl you please refrain from using non standard quoting. It makes
it more difficult for people who use parsers to sort out quoting levels
(color coding, etc.) groups.google.com is one such place that does this.
Thank you.


[Heres an old post relating to this, indicating I'm not the first to ask
either.]
[Google:
http://groups.google.com/groups?sel...0035.dyn.interpath.net&oe=UTF-8&output=gplain ]

From: (e-mail address removed) (J. Moreno)
Subject: Re: Statistics for comp.lang.perl.misc
Date: 1999/06/22
Message-ID: <[email protected]>#1/1
References: <[email protected]>
<[email protected]> <[email protected]>
<[email protected]> <[email protected]>
<[email protected]> <[email protected]>
<[email protected]> <[email protected]>
<[email protected]>
Mail-Copies-To: nobody
X-ACLA-Charter:
<http://www.dejanews.com/msgid.xp?MID=<phenix-121119981748546229@roxboro0-
023.dyn.interpath.net%3E&ST=&AH=1>
X-Face: 48Sy^\</dKZe1J<x|w^od|[iJ}v*!!%."nNxwca8S2[}H-$<)+4]aq(v<uv
:jQ5(aeN*L~ayLVWcoPoFtB7M7v}N,+YkX*__,I8RqhT
Organization: (e-mail address removed)
X-FAQs-Read: perl, pascal.mac
User-Agent: MacSOUP/2.4
Newsgroups: comp.lang.perl.misc

Greg> That's beatable too. What if she uses different prefixes for each
Greg> quoted line?

Maybe Abi deserves to be taken out of the normal mix, and given an
honorary "automatically historically excessive" OCR place in the
standings. Item # 0 or something.

Maybe what I'm really saying is I don't understand if Abi is using
different prefixes just to foul up statistics, or what. And if so,
the stats should just ignore her. And I don't like having that kind
of confusion in my life.

print "Just another confused Perl hacker,"

I'd agree with you totally, if you didn't use a non standard style
yourself.

The more people that use the standard quoting style the easier it is to
convince newsreader authors that they should recognize that there /is/ a
standard.

And ditto for posters -- non-standard quoting just encourages them to
ignore the fact that their newsreaders spew out garbage; after all why
should they take the trouble to switch to something better or fix it by
hand when the old timers/big names are using something non-standard.

 

[Wally Sanford]

Wally Sanford ([email protected]) wrote on MMMDCLXXXIII
September MCMXCIII in
<URL:== Abigail
wrote: == > Xaonon ([email protected]) wrote on MMMDCLXXXII September
MCMXCIII in == > <URL:== > {} Or, more likely, it was broken before in some way that wasn't
== > apparent until {} I installed 5.8.1. The program in question
is a == > signature generator, which {} runs as a daemon and
produces random == > quips via a named pipe. The main loop {} looks
like this: == > {}
== [Snip more NON STANDARD quoting...]
==
== Abigail, coudl you please refrain from using non standard
quoting. It makes == it more difficult for people who use parsers to
sort out quoting levels == (color coding, etc.) groups.google.com is
one such place that does this. == Thank you.
==
==
== [Heres an old post relating to this, indicating I'm not the first
to ask == either.]
== [Google:
==
== From: (e-mail address removed) (J. Moreno)
== Subject: Re: Statistics for comp.lang.perl.misc
== Date: 1999/06/22


Let's see, two requests in 4 years. I can't really be impressed by
those numbers.

Then you didn't search very hard. There have been numerous threads on the
subject, and the net conclusion as always come to that std is the way to go;
it's just that developers that create readers don't consider the
*standards*. And it gets worse when the "veterans" of usenet don't follow
the standards themselves. Not the kind of example you want to set for newer
folks.

You have a few options:
- Make your parser smarter. It'll be fairly trivial in Perl to read
a posting of mine and determine that posts quoting prefix.
- Killfile me.

The latter is the easiest option.

This is entirely beside the point. Even if *I* killfiled you, many others
would se it. New comers would catch the same bad habit and it spreads. Why
do you think this subject has been brought up all over usenet for years? And
not to mention virtually all content still ends up in Google.

 

[Xaonon]

Xaonon ([email protected]) wrote on MMMDCLXXXII September MCMXCIII in


What does your program do with 5.8.0?

I don't know. I had 5.6.1 and went directly to 5.8.1 from there.

 

[Xaonon]

Xaonon ([email protected]) wrote on MMMDCLXXXIII September MCMXCIII in


Well, you were claiming that 5.8.1 broke your program. And before anyone
is looking at the new things in 5.8.1, or are put of upgrading from
5.8.0 to 5.8.1, it would be useful to find out how your program behaves
in 5.8.0. Especially considering that 5.8.0 defaults to using 'perlio'
instead of 'stdio' as default.

Well, I'm not building 5.8.0 just to see what happens. I can just run the
program explicitly with 5.6.1 if necessary, I'd just like to know why it
works differently now.

Also, the new documentation misspelled "nitroglycerin".

And you also might want to try setting the environment variable
PERLIO to 'perlio' and 'stdio' repectively.

I did that, and it didn't seem to have any immediate (visible) effect.

 

[Randal L. Schwartz]

Purl> My testing, experiments, exemplify Perl 5.8.x is the most buggy
Purl> and the most poorly written version of Perl to date.

Really? And how many of these bugs have you reported, with
rt.perl.org, or during the public beta and gamma testing?

That would have been the responsible thing to do.

I can support the claim that 5.8.1 *fixes* far more bugs
than it *introduces*.

I will admit that new bugs have been introduced. This is the nature
of change. But the net progress is forward, not backward.

 

[Al MacHonahey]

Wally Sanford ([email protected]) wrote on MMMDCLXXXIII September
MCMXCIII in <URL:== Abigail wrote:
== > Xaonon ([email protected]) wrote on MMMDCLXXXII September MCMXCIII in
== > <URL:== > {} Or, more likely, it was broken before in some way that wasn't
== > apparent until {} I installed 5.8.1. The program in question is a
== > signature generator, which {} runs as a daemon and produces random
== > quips via a named pipe. The main loop {} looks like this:
== > {}
== [Snip more NON STANDARD quoting...]
==
== Abigail, coudl you please refrain from using non standard quoting. It makes
== it more difficult for people who use parsers to sort out quoting levels
== (color coding, etc.) groups.google.com is one such place that does this.
== Thank you.
==
==
== [Heres an old post relating to this, indicating I'm not the first to ask
== either.]
== [Google:
==
== From: (e-mail address removed) (J. Moreno)
== Subject: Re: Statistics for comp.lang.perl.misc
== Date: 1999/06/22


Let's see, two requests in 4 years. I can't really be impressed by
those numbers.

You have a few options:
- Make your parser smarter. It'll be fairly trivial in Perl to read a
posting of mine and determine that posts quoting prefix.
- Killfile me.

The latter is the easiest option.


Abigail

You know, its really ironic how you folks complain profusely if some
goes against Usenet tradition/standards of not top-posting or quoting
entire posts while adding one line of oc, but you feel to go against
another, seemingly because you feel "above" everyone else who tries to
follow the standard/ countless Usenet guidelines that you pretend to
cherish.

You see, using > for quoting characters as been the Usenet standard
(or plain text medium standard for that matter, i.e. email) for
quoting text for over a decade. Many article processors use this, as
have millions of Usenet users do, but you instead for what ever reason
continue your own way, preventing countless article processors to miss
parse your articles, such as stat keepers.

Maybe we should use Abigail's example and start posting using html, as
she has shown us that Usenet tradition/guidelines/standards do not
need to be followed.