search

Perl / perl5

🐪 The Perl programming language
https://dev.perl.org/perl5/
Other
2.07k stars 574 forks source link

Pod::Text doesn't like L<gnuplot F<term/README>> #5707

Closed p5pRT closed 20 years ago

p5pRT commented 22 years ago

Migrated from rt.perl.org#10035 (status was 'resolved')

Searchable as RT10035$

p5pRT commented 22 years ago

From dcd@tc.fluke.com

Created by dcd@tc.fluke.com

the following perl script reports DESCRIPTION *** ERROR​: unterminated F\<...> at line 4 in file (unknown)   "README>" in gnuplot term indicating that it doesn't like L\<gnuplot F\<term/README>> The author of the original pod document indicates this is a bug in the pod2man and pod2text family.

use Pod​::Text; $parser = new Pod​::Text(); $parser->parse_from_filehandle(\*DATA);

__DATA__

=head1 DESCRIPTION

L\<gnuplot F\<term/README>>

=cut

Perl Info ``` Flags: category=library severity=medium Site configuration information for perl v5.8.0: Configured by dcd at Mon Jul 8 20:09:16 PDT 2002. Summary of my perl5 (revision 5.0 version 8 subversion 0 patch 17429) configuration: Platform: osname=linux, osvers=2.4.19-rc1, archname=i686-linux uname='linux dd 2.4.19-rc1 #1 mon jun 24 14:32:06 pdt 2002 i686 ' config_args='-Dmksymlinks -Dinstallusrbinperl -Uversiononly -Dusedevel -Doptimize=-O3 -g -de -Dcf_email=dcd@tc.fluke.com' hint=recommended, useposix=true, d_sigaction=define usethreads=undef use5005threads=undef useithreads=undef usemultiplicity=undef useperlio=define d_sfio=undef uselargefiles=define usesocks=undef use64bitint=undef use64bitall=undef uselongdouble=undef usemymalloc=n, bincompat5005=undef Compiler: cc='cc', ccflags ='-DDEBUGGING -fno-strict-aliasing -I/usr/local/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64', optimize='-O3 -g', cppflags='-DDEBUGGING -fno-strict-aliasing -I/usr/local/include' ccversion='', gccversion='egcs-2.91.66.1 19990314/Linux (egcs-1.1.2 release)', gccosandvers='' intsize=4, longsize=4, ptrsize=4, doublesize=8, byteorder=1234 d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=12 ivtype='long', ivsize=4, nvtype='double', nvsize=8, Off_t='off_t', lseeksize=4 alignbytes=4, prototype=define Linker and Libraries: ld='cc', ldflags =' -L/usr/local/lib' libpth=/usr/local/lib /lib /usr/lib libs=-lgdbm -ldbm -ldb -ldl -lm -lc perllibs=-ldl -lm -lc libc=/lib/libc.so.5.4.44, so=so, useshrplib=false, libperl=libperl.a gnulibc_version='' Dynamic Linking: dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=undef, ccdlflags='-rdynamic' cccdlflags='-fpic', lddlflags='-shared -L/usr/local/lib' Locally applied patches: DEVEL17411 @INC for perl v5.8.0: /usr/local/lib/perl5/5.8.0/i686-linux /usr/local/lib/perl5/5.8.0 /usr/local/lib/perl5/site_perl/5.8.0/i686-linux /usr/local/lib/perl5/site_perl/5.8.0 /usr/local/lib/perl5/site_perl/5.7.3 /usr/local/lib/perl5/site_perl . Environment for perl v5.8.0: HOME=/home/dcd LANG (unset) LANGUAGE (unset) LD_LIBRARY_PATH (unset) LOGDIR (unset) PATH=/home/dcd/bin:/sbin:/usr/local/bin:/bin:/usr/bin:/usr/X11/bin:/usr/games:/usr/local/samba:/home/hobbes/tools/scripts:/home/hobbes/tools/linux:/usr0/hobbes/tools/scripts:/usr0/dcd/bin:/apps/general/bin:/usr/public PERL_BADLANG (unset) SHELL=/bin/bash ```
p5pRT commented 22 years ago

From @tamias

On Tue\, Jul 09\, 2002 at 10​:31​:01AM -0700\, David Dyck wrote​:

the following perl script reports DESCRIPTION *** ERROR​: unterminated F\<...> at line 4 in file (unknown) "README>" in gnuplot term indicating that it doesn't like L\<gnuplot F\<term/README>> The author of the original pod document indicates this is a bug in the pod2man and pod2text family.

What exactly is L\<gnuplot F\<term/README>> supposed to mean?

Ronald

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

On Tue\, 9 Jul 2002 at 13​:40 -0400\, Ronald J Kimball \<rjk@​linguist.thayer.da...​:

On Tue\, Jul 09\, 2002 at 10​:31​:01AM -0700\, David Dyck wrote​:

the following perl script reports DESCRIPTION *** ERROR​: unterminated F\<...> at line 4 in file (unknown) "README>" in gnuplot term indicating that it doesn't like L\<gnuplot F\<term/README>> The author of the original pod document indicates this is a bug in the pod2man and pod2text family.

What exactly is L\<gnuplot F\<term/README>> supposed to mean?

Ronald

I don't know\, I just got the error when installing the Term​::Gnuplot package from CPAN

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

David Dyck \dcd@&#8203;tc\.fluke\.com writes​:

DESCRIPTION *** ERROR​: unterminated F\<...> at line 4 in file (unknown) "README>" in gnuplot term indicating that it doesn't like L\<gnuplot F\<term/README>> The author of the original pod document indicates this is a bug in the pod2man and pod2text family.

That error comes from Pod​::Parser\, not from the Pod​::Text code\, so I can't address it directly in podlators. I'm not sure what's causing it; at a guess\, there some sort of special processing for L\<> that's being tripped up. I'm afraid that I likely won't get a chance to look at it in detail soon (I could if it's in podlators\, but I don't know Pod​::Parser well enough to find bugs quickly).

-- Russ Allbery (rra@​stanford.edu) \<http​://www.eyrie.org/~eagle/>

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

On Tue\, Jul 09\, 2002 at 10​:55​:36AM -0700\, David Dyck wrote​:

What exactly is L\<gnuplot F\<term/README>> supposed to mean?

Be a crosslink to a section named as given...

I don't know\, I just got the error when installing the Term​::Gnuplot package from CPAN

Ilya

p5pRT commented 22 years ago

From @IlyaM

I'm trying to backport Data​::Dumper which comes with latest Perls (I'm using 5.8.0-RC1 - I guess Data​::Dumper in RC2 doesn't differs too much).

My first attempt is trying to make it work on 5.6.1.

With minor patching I could make it compile and even pass 70% of tests. The other 30% cause severe headache for me.

One problem is that Data​::Dumper seems to traverse data structures with circular references in different order in 5.8.0 and 5.6.1.

For example for some data structure Data​::Dumper in 5.8.0 produces

#%b = ( # 'a' => [ # 1\, # {}\, # [ # 'c' # ] # ]\, # 'b' => {}\, # 'c' => [] # ); #$b{'a'}[1] = \%b; #$b{'b'} = \%b; #$b{'c'} = $b{'a'}[2]; #$a = $b{'a'};

and in 5.6.1 it produces

#%b = ( # 'c' => [ # 'c' # ]\, # 'a' => [ # 1\, # {}\, # [] # ]\, # 'b' => {} # ); #$b{'a'}[1] = \%b; #$b{'a'}[2] = $b{'c'}; #$b{'b'} = \%b; #$a = $b{'a'};

These outputs are actually equivalent but Data​::Dumper's test compares output as string with samples inlined in test script so corresponding test fails in 5.6.1.

Now I have a question​: what to do with these tests? Just skip them in old Perls? Or should I add output samples for 5.6.1 (and probably watch them to fail in 5.6.0 and 5.005 :( )? Any other briliant ideas?

Another problem is utf8 support tests but I guess that they can be just skipped in old Perls.

-- Ilya Martynov (http​://martynov.org/)

p5pRT commented 22 years ago

From @tamias

On Tue\, Jul 09\, 2002 at 11​:05​:09AM -0700\, Russ Allbery wrote​:

David Dyck \dcd@&#8203;tc\.fluke\.com writes​:

DESCRIPTION *** ERROR​: unterminated F\<...> at line 4 in file (unknown) "README>" in gnuplot term indicating that it doesn't like L\<gnuplot F\<term/README>> The author of the original pod document indicates this is a bug in the pod2man and pod2text family.

That error comes from Pod​::Parser\, not from the Pod​::Text code\, so I can't address it directly in podlators. I'm not sure what's causing it; at a guess\, there some sort of special processing for L\<> that's being tripped up. I'm afraid that I likely won't get a chance to look at it in detail soon (I could if it's in podlators\, but I don't know Pod​::Parser well enough to find bugs quickly).

It just doesn't look like valid syntax to me.

Ronald

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

On Tue\, 9 Jul 2002\, Ilya Zakharevich wrote​:

On Tue\, Jul 09\, 2002 at 10​:55​:36AM -0700\, David Dyck wrote​:

What exactly is L\<gnuplot F\<term/README>> supposed to mean?

Be a crosslink to a section named as given...

If that's the case then perlpodspec needs updating. But it doesn't look right to me.

-- \<!-- Matt --> \<​:->Get a smart net\</​:->

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

On Tue\, Jul 09\, 2002 at 08​:44​:04PM +0100\, Matt Sergeant wrote​:

On Tue\, 9 Jul 2002\, Ilya Zakharevich wrote​:

On Tue\, Jul 09\, 2002 at 10​:55​:36AM -0700\, David Dyck wrote​:

What exactly is L\<gnuplot F\<term/README>> supposed to mean?

Be a crosslink to a section named as given...

If that's the case then perlpodspec needs updating. But it doesn't look right to me.

*What* does not look right? That one may need to refer to an arbitrary section name?

Ilya

p5pRT commented 22 years ago

From @gsar

On 09 Jul 2002 22​:36​:44 +0400\, Ilya Martynov wrote​:

I'm trying to backport Data​::Dumper which comes with latest Perls (I'm using 5.8.0-RC1 - I guess Data​::Dumper in RC2 doesn't differs too much).

My first attempt is trying to make it work on 5.6.1.

Data​::Dumper has no life outside of the perl sources. This is quite intentional.

I'm ok with someone backporting and maintaining it outside perl as long as they take over its maintenance too. Let me know if you wish to be the maintainer and I can assign rights to the namespace to you in PAUSE.

Sarathy gsar@​ActiveState.com

p5pRT commented 22 years ago

From @rgarcia

On 2002.07.09 20​:36 Ilya Martynov wrote​:

I'm trying to backport Data​::Dumper which comes with latest Perls (I'm using 5.8.0-RC1 - I guess Data​::Dumper in RC2 doesn't differs too much).

My first attempt is trying to make it work on 5.6.1.

With minor patching I could make it compile and even pass 70% of tests. The other 30% cause severe headache for me.

One problem is that Data​::Dumper seems to traverse data structures with circular references in different order in 5.8.0 and 5.6.1.

That's documented in perldelta :

: Although "you shouldn't do that"\, it was possible to write code that : depends on Perl's hashed key order (Data​::Dumper does this). The new : algorithm "One-at-a-Time" produces a different hashed key order.

I think that MLDBM has a similar problem.

Now I have a question​: what to do with these tests? Just skip them in old Perls? Or should I add output samples for 5.6.1 (and probably watch them to fail in 5.6.0 and 5.005 :( )? Any other briliant ideas?

I think that you should check whether $] \<= 5.006 to know what output to expect.

(Sarathy\, will 5.6.2 use the same hashing algorithm as 5.6.1 ?)

p5pRT commented 22 years ago

From @IlyaM

On Tue\, 09 Jul 2002 12​:51​:35 -0700\, Gurusamy Sarathy \gsar@&#8203;ActiveState\.com said​:

GS> On 09 Jul 2002 22​:36​:44 +0400\, Ilya Martynov wrote​:

I'm trying to backport Data​::Dumper which comes with latest Perls (I'm using 5.8.0-RC1 - I guess Data​::Dumper in RC2 doesn't differs too much).

My first attempt is trying to make it work on 5.6.1.

GS> Data​::Dumper has no life outside of the perl sources. This is quite GS> intentional.

One module I use (Test​::Differences) depends on some features in latest Data​::Dumper. This is why I'm trying to backport Data​::Dumper.

GS> I'm ok with someone backporting and maintaining it outside perl as GS> long as they take over its maintenance too. Let me know if you wish GS> to be the maintainer and I can assign rights to the namespace to you GS> in PAUSE.

I don't mind maintaining backport. My CPAN ID is ILYAM.

-- Ilya Martynov (http​://martynov.org/)

p5pRT commented 22 years ago

From @gsar

On Tue\, 09 Jul 2002 21​:56​:01 +0200\, Rafael Garcia-Suarez wrote​:

(Sarathy\, will 5.6.2 use the same hashing algorithm as 5.6.1 ?)

Highly unlikely. I wouldn't want to "break" code that accidentally depends on the hashing order in a maintenance release.

Sarathy gsar@​ActiveState.com

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

Ilya Martynov \ilya@&#8203;martynov\.org writes​:

One problem is that Data​::Dumper seems to traverse data structures with circular references in different order in 5.8.0 and 5.6.1.

That is to be expected the hash function is different.

Now I have a question​: what to do with these tests? Just skip them in old Perls? Or should I add output samples for 5.6.1 (and probably watch them to fail in 5.6.0 and 5.005 :( )? Any other briliant ideas?

Another problem is utf8 support tests but I guess that they can be just skipped in old Perls. -- Nick Ing-Simmons http​://www.ni-s.u-net.com/

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

Ronald J Kimball \rjk@&#8203;linguist\.Thayer\.dartmouth\.edu writes​:

On Tue\, Jul 09\, 2002 at 11​:05​:09AM -0700\, Russ Allbery wrote​:

David Dyck \dcd@&#8203;tc\.fluke\.com writes​:

DESCRIPTION *** ERROR​: unterminated F\<...> at line 4 in file (unknown) "README>" in gnuplot term indicating that it doesn't like L\<gnuplot F\<term/README>> The author of the original pod document indicates this is a bug in the pod2man and pod2text family.

That error comes from Pod​::Parser\, not from the Pod​::Text code\, so I can't address it directly in podlators. I'm not sure what's causing it; at a guess\, there some sort of special processing for L\<> that's being tripped up. I'm afraid that I likely won't get a chance to look at it in detail soon (I could if it's in podlators\, but I don't know Pod​::Parser well enough to find bugs quickly).

It just doesn't look like valid syntax to me.

The X\<> stuff can nest. Once upon a time if section/item was specified with F\<> (say) you _had_ to have F\<> in L\<> too or it did not match. The parser should be handling that these days. And I seem to recall that for cross linking purposes we were going to flatten the markup.

Which is not to say POD authors cannot re-write their POD to avoid the issue.

Ronald -- Nick Ing-Simmons http​://www.ni-s.u-net.com/

p5pRT commented 22 years ago

From @gsar

On Tue\, 09 Jul 2002 13​:18​:45 PDT\, Gurusamy Sarathy wrote​:

On Tue\, 09 Jul 2002 21​:56​:01 +0200\, Rafael Garcia-Suarez wrote​:

(Sarathy\, will 5.6.2 use the same hashing algorithm as 5.6.1 ?)

Highly unlikely. I wouldn't want to "break" code that accidentally

s/un//\, I mean!

depends on the hashing order in a maintenance release.

Sarathy gsar@​ActiveState.com

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

On Tue\, 9 Jul 2002\, Ilya Zakharevich wrote​:

On Tue\, Jul 09\, 2002 at 08​:44​:04PM +0100\, Matt Sergeant wrote​:

On Tue\, 9 Jul 2002\, Ilya Zakharevich wrote​:

On Tue\, Jul 09\, 2002 at 10​:55​:36AM -0700\, David Dyck wrote​:

What exactly is L\<gnuplot F\<term/README>> supposed to mean?

Be a crosslink to a section named as given...

If that's the case then perlpodspec needs updating. But it doesn't look right to me.

*What* does not look right? That one may need to refer to an arbitrary section name?

That should be L\<"gnuplot "|term/README> or L\<"gnuplot "|term​::README>\, but POD doesn't allow for links like the one above\, according to perlpodspec. And since it was agreed nearly a year ago now that perlpodspec was the last word in POD\, it needs to be fixed if you want the above to work.

-- \<!-- Matt --> \<​:->Get a smart net\</​:->

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

Matt Sergeant \matt@&#8203;sergeant\.org writes​:

On Tue\, 9 Jul 2002\, Ilya Zakharevich wrote​:

On Tue\, Jul 09\, 2002 at 08​:44​:04PM +0100\, Matt Sergeant wrote​:

On Tue\, 9 Jul 2002\, Ilya Zakharevich wrote​:

On Tue\, Jul 09\, 2002 at 10​:55​:36AM -0700\, David Dyck wrote​:

What exactly is L\<gnuplot F\<term/README>> supposed to mean?

A link to (say)​:

=head2 gnuplot F\<term/README>

and not a link to

=head2 gnuplot F\<another/path>

I believe that L\<gnuplot term/README> is supposed to work someday.

That should be L\<"gnuplot "|term/README> or L\<"gnuplot "|term​::README>\,

No\, both of those are changing the text displayed\, rather than target a section name.

but POD doesn't allow for links like the one above\, according to perlpodspec. And since it was agreed nearly a year ago now that perlpodspec was the last word in POD\, it needs to be fixed if you want the above to work. -- Nick Ing-Simmons http​://www.ni-s.u-net.com/

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

On Tue\, 9 Jul 2002\, Nick Ing-Simmons wrote​:

Matt Sergeant \matt@&#8203;sergeant\.org writes​:

On Tue\, 9 Jul 2002\, Ilya Zakharevich wrote​:

On Tue\, Jul 09\, 2002 at 08​:44​:04PM +0100\, Matt Sergeant wrote​:

On Tue\, 9 Jul 2002\, Ilya Zakharevich wrote​:

On Tue\, Jul 09\, 2002 at 10​:55​:36AM -0700\, David Dyck wrote​:

What exactly is L\<gnuplot F\<term/README>> supposed to mean?

A link to (say)​:

=head2 gnuplot F\<term/README>

and not a link to

=head2 gnuplot F\<another/path>

I believe that L\<gnuplot term/README> is supposed to work someday.

OK\, I'll look into why perlpodspec is confusing in this regard. It may just be something missing\, rather than a broken spec.

-- \<!-- Matt --> \<​:->Get a smart net\</​:->

p5pRT commented 22 years ago

From @tamias

On Tue\, Jul 09\, 2002 at 03​:47​:26PM -0400\, Ilya Zakharevich wrote​:

On Tue\, Jul 09\, 2002 at 08​:44​:04PM +0100\, Matt Sergeant wrote​:

On Tue\, 9 Jul 2002\, Ilya Zakharevich wrote​:

On Tue\, Jul 09\, 2002 at 10​:55​:36AM -0700\, David Dyck wrote​:

What exactly is L\<gnuplot F\<term/README>> supposed to mean?

Be a crosslink to a section named as given...

If that's the case then perlpodspec needs updating. But it doesn't look right to me.

*What* does not look right? That one may need to refer to an arbitrary section name?

Is that what it's trying to do? I believe the right way to do that is with E\<>. This would probably work​:

L\<gnuplot F\<termE\README>>

Ronald

p5pRT commented 22 years ago

From @nwc10

On Tue\, Jul 09\, 2002 at 09​:56​:01PM +0200\, Rafael Garcia-Suarez wrote​:

On 2002.07.09 20​:36 Ilya Martynov wrote​:

I'm trying to backport Data​::Dumper which comes with latest Perls (I'm using 5.8.0-RC1 - I guess Data​::Dumper in RC2 doesn't differs too much).

My first attempt is trying to make it work on 5.6.1.

With minor patching I could make it compile and even pass 70% of tests. The other 30% cause severe headache for me.

One problem is that Data​::Dumper seems to traverse data structures with circular references in different order in 5.8.0 and 5.6.1.

That's documented in perldelta :

: Although "you shouldn't do that"\, it was possible to write code that : depends on Perl's hashed key order (Data​::Dumper does this). The new : algorithm "One-at-a-Time" produces a different hashed key order.

I think that MLDBM has a similar problem.

Now I have a question​: what to do with these tests? Just skip them in old Perls? Or should I add output samples for 5.6.1 (and probably watch them to fail in 5.6.0 and 5.005 :( )? Any other briliant ideas?

I think that you should check whether $] \<= 5.006 to know what output to expect.

It's more fun than that. The hashing algorithm changed from 5.005 to 5.6\, and again from 5.6 to 5.8 The hash order also differs between EBCDIC and ASCII-based machines. I suspect it may also differ subtly on Crays for some data\, as U32 is 64 bits long there\, and U32 is what is used for hashing

Currently there is a horrible yucky nasty bletcherous bodge (oh\, by the way did I say I don't like it?) in the test - it sorts the output and the expected on EBCDIC machines before comparing. It all went wrong (still unresolved) when I added a test for something failing (that I fixed) to do with storing numbers ending in newline\, such as "12\n"\, which happen to match /^\d+$/

IIRC I took perl 5.7.3ish\, made 2 more copies\, and replaced PERL_HASH in hv.h with the hash algorithms from 5.005 and 5.6. I was trying to re-write Data​::Dumper's tests so that they were hash key order independent. However\, it started to become a horrible mess of a test\, as one you start testing self-referential data structures\, what Data​::Dumper returns changes markedly based on which things it meets first. So it's not pretty\, and Jarkko didn't like it when I sent a first cut to him. So I'm not sure of a good way to do it. I think that the experimental version survives\, but it's on my FreeBSD machine which is currently being rebuilt\, so I'm not likely to find it until I finish the rebuild tomorrow night.

IIRC there are some systematic obscure classes of things Data​::Dumper fails on - Randall Schwartz reported some\, again IIRC. But that's not a backporting issue.

Nicholas Clark -- Even better than the real thing​: http​://nms-cgi.sourceforge.net/

p5pRT commented 22 years ago

From @tamias

On Tue\, Jul 09\, 2002 at 10​:25​:54PM +0100\, Matt Sergeant wrote​:

On Tue\, 9 Jul 2002\, Nick Ing-Simmons wrote​:

What exactly is L\<gnuplot F\<term/README>> supposed to mean?

A link to (say)​:

=head2 gnuplot F\<term/README>

and not a link to

=head2 gnuplot F\<another/path>

I believe that L\<gnuplot term/README> is supposed to work someday.

OK\, I'll look into why perlpodspec is confusing in this regard. It may just be something missing\, rather than a broken spec.

The current spec for L\<> is inherently broken. :)

Ronald

p5pRT commented 22 years ago

From @gsar

On 10 Jul 2002 00​:09​:42 +0400\, Ilya Martynov wrote​:

GS> I'm ok with someone backporting and maintaining it outside perl as GS> long as they take over its maintenance too. Let me know if you wish GS> to be the maintainer and I can assign rights to the namespace to you GS> in PAUSE.

I don't mind maintaining backport. My CPAN ID is ILYAM.

I've marked you now as a co-maintainer so you should be able to upload new versions &c to CPAN.

Thanks.

Sarathy gsar@​ActiveState.com

p5pRT commented 22 years ago

From @gsar

On Tue\, 09 Jul 2002 23​:08​:14 BST\, Nicholas Clark wrote​:

IIRC I took perl 5.7.3ish\, made 2 more copies\, and replaced PERL_HASH in hv.h with the hash algorithms from 5.005 and 5.6. I was trying to re-write Data​::Dumper's tests so that they were hash key order independent. However\, it started to become a horrible mess of a test\, as one you start testing self-referential data structures\, what Data​::Dumper returns changes markedly based on which things it meets first. So it's not pretty\, and Jarkko didn't like it when I sent a first cut to him. So I'm not sure of a good way to do it.

You could enable the hash key sorting option always.

Sarathy gsar@​ActiveState.com

p5pRT commented 22 years ago

From @schwern

On Tue\, Jul 09\, 2002 at 10​:36​:44PM +0400\, Ilya Martynov wrote​:

These outputs are actually equivalent but Data​::Dumper's test compares output as string with samples inlined in test script so corresponding test fails in 5.6.1.

Now I have a question​: what to do with these tests? Just skip them in old Perls? Or should I add output samples for 5.6.1 (and probably watch them to fail in 5.6.0 and 5.005 :( )? Any other briliant ideas?

One way out is rather than comparing the output of Data​::Dumper against a static string\, which then has a dependency on hash ordering\, you can simply take a known data structure\, dump it\, eval it back in and then compare it against the original data structure.

The problem is comparing two complex data structures with circular references. Test​::More's is_deeply() doesn't like them (patches welcome).

-- This sig file temporarily out of order.

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

Ronald J Kimball \rjk@&#8203;linguist\.Thayer\.dartmouth\.edu writes​:

Is that what it's trying to do? I believe the right way to do that is with E\<>. This would probably work​:

L\<gnuplot F\<termE\README>>

Ah\, yes. The original *is* invalid syntax. You can't have an unescaped literal / in the middle of L\<> unless you mean it\, and in this case the author didn't mean it.

-- Russ Allbery (rra@​stanford.edu) \<http​://www.eyrie.org/~eagle/>

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

On Tue\, 9 Jul 2002\, Ronald J Kimball wrote​:

On Tue\, Jul 09\, 2002 at 10​:25​:54PM +0100\, Matt Sergeant wrote​:

On Tue\, 9 Jul 2002\, Nick Ing-Simmons wrote​:

What exactly is L\<gnuplot F\<term/README>> supposed to mean?

A link to (say)​:

=head2 gnuplot F\<term/README>

and not a link to

=head2 gnuplot F\<another/path>

I believe that L\<gnuplot term/README> is supposed to work someday.

OK\, I'll look into why perlpodspec is confusing in this regard. It may just be something missing\, rather than a broken spec.

The current spec for L\<> is inherently broken. :)

Well then by inheritance\, so is the spec for POD. It doesn't have to be this way. Sean Burke put a lot of effort into making it not so (and did a damn good job against all odds).

-- \<!-- Matt --> \<​:->Get a smart net\</​:->

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

On Tue\, Jul 09\, 2002 at 06​:11​:16PM -0400\, Ronald J Kimball wrote​:

*What* does not look right? That one may need to refer to an arbitrary section name?

Is that what it's trying to do? I believe the right way to do that is with E\<>. This would probably work​:

L\<gnuplot F\<termE\README>>

Even if the section name does not contain E\? Looks fishy to me...

Ilya

p5pRT commented 22 years ago

From @tamias

On Wed\, Jul 10\, 2002 at 06​:52​:39AM +0100\, Matt Sergeant wrote​:

On Tue\, 9 Jul 2002\, Ronald J Kimball wrote​:

On Tue\, Jul 09\, 2002 at 10​:25​:54PM +0100\, Matt Sergeant wrote​:

OK\, I'll look into why perlpodspec is confusing in this regard. It may just be something missing\, rather than a broken spec.

The current spec for L\<> is inherently broken. :)

Well then by inheritance\, so is the spec for POD. It doesn't have to be this way. Sean Burke put a lot of effort into making it not so (and did a damn good job against all odds).

That's true\, and the POD spec is much more readable and consistent now. However\, the basic syntax for L\<> has not changed; the spec mentions how tricky it is to parse and how complex it is.

Ronald

p5pRT commented 22 years ago

From @tamias

On Wed\, Jul 10\, 2002 at 08​:18​:59AM -0400\, Ilya Zakharevich wrote​:

On Tue\, Jul 09\, 2002 at 06​:11​:16PM -0400\, Ronald J Kimball wrote​:

*What* does not look right? That one may need to refer to an arbitrary section name?

Is that what it's trying to do? I believe the right way to do that is with E\<>. This would probably work​:

L\<gnuplot F\<termE\README>>

Even if the section name does not contain E\? Looks fishy to me...

Is it any fishier than a URL like http​://www.example.com/file%20name.txt when the actual file name doesn't contain %20?

Ronald

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

On Wed\, Jul 10\, 2002 at 09​:28​:32AM -0400\, Ronald J Kimball wrote​:

Is that what it's trying to do? I believe the right way to do that is with E\<>. This would probably work​:

L\<gnuplot F\<termE\README>>

Even if the section name does not contain E\? Looks fishy to me...

Is it any fishier than a URL like http​://www.example.com/file%20name.txt when the actual file name doesn't contain %20?

Yes\, of course. Since a URL can have no embedded URLs\, but a header might have a L\<> inside...

Ilya

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

On Wed\, 10 Jul 2002 at 08​:18 -0400\, Ilya Zakharevich \<ilya@​math.ohio-state....​:

On Tue\, Jul 09\, 2002 at 06​:11​:16PM -0400\, Ronald J Kimball wrote​:

*What* does not look right? That one may need to refer to an arbitrary section name?

Is that what it's trying to do? I believe the right way to do that is with E\<>. This would probably work​:

L\<gnuplot F\<termE\README>>

Even if the section name does not contain E\? Looks fishy to me...

While the above does display something 'reasonable' when passed through pod2text and pod2man\, both lynx and netscape display something like   gnuplot term/README with the output from pod2html where the html looks like \gnuplot \term/README\\

I'm sure the link is not where the original author intended it to go.

  David

p5pRT commented 22 years ago

From [Unknown Contact. See original ticket]

David Dyck \dcd@&#8203;tc\.fluke\.com writes​:

Ilya Zakharevich \<ilya@​math.ohio-state....​:

On Tue\, Jul 09\, 2002 at 06​:11​:16PM -0400\, Ronald J Kimball wrote​:

E\<>. This would probably work​: L\<gnuplot F\<termE\README>>

Even if the section name does not contain E\? Looks fishy to me...

/ is a special character inside L\<>. It separates the page name from the section within that page. If you mean a literal slash\, you have to escape it. That doesn't look fishy to me at all; it seems quite logical and straightforward.

While the above does display something 'reasonable' when passed through pod2text and pod2man\, both lynx and netscape display something like gnuplot term/README with the output from pod2html where the html looks like \gnuplot \term/README\\

I'm sure the link is not where the original author intended it to go.

pod2html is broken. E\ should not be turned into an HTML entity\, and it should be turned into a / for the href tag (and something should also be done with the F\<> markup).

-- Russ Allbery (rra@​stanford.edu) \<http​://www.eyrie.org/~eagle/>

p5pRT commented 20 years ago

From @smpeters

Looking through this thread\, it appears that L\<gnuplot F\<term/README>> is invalid since the L\<> will treat the "/" as a literal within its scope. The correct syntax\, which works\, is​:

L\<gnuplot F\<termE\README>>

p5pRT commented 20 years ago

@smpeters - Status changed from 'open' to 'resolved'