Bug #101866 for POSIX-1003: No bits/signum.h on FreeBSD systems

Mon Feb 02 01:47:06 2015 SREZIC [...] cpan.org - Ticket created

Subject:

No bits/signum.h on FreeBSD systems

bits/signum.h does not exist on FreeBSD systems (and probably on no non-Linux system) and prevents compilation of 1003.c. I think also that including it is not necessary on Linux systems, too.

Mon Feb 02 01:51:56 2015 SREZIC [...] cpan.org - Correspondence added

On 2015-02-02 01:47:06, SREZIC wrote: Show quoted text

> bits/signum.h does not exist on FreeBSD systems (and probably on no > non-Linux system) and prevents compilation of 1003.c. I think also > that including it is not necessary on Linux systems, too.

Removing the include seems to work for FreeBSD and Linux. See the attached patch. Regards, Slaven

Subject:

POSIX-1003-0.99_01-RT101866.patch

From 29471dd4582f6dfddae91932fdbf9f663e7d652b Mon Sep 17 00:00:00 2001 From: Slaven Rezic <slaven@rezic.de> Date: Mon, 2 Feb 2015 07:47:35 +0100 Subject: [PATCH] no bits/signum.h on FreeBSD (RT #101866) --- 1003.xs | 1 - 1 file changed, 1 deletion(-) diff --git a/1003.xs b/1003.xs index 77fd714..1549dd3 100644 --- a/1003.xs +++ b/1003.xs @@ -175,7 +175,6 @@ #define I_SIGNAL #ifdef I_SIGNAL #include <signal.h> -#include <bits/signum.h> #endif /* -- 2.1.2

Mon Feb 02 02:47:49 2015 Mark [...] Overmeer.net - Correspondence added

Subject:	Re: [rt.cpan.org #101866] No bits/signum.h on FreeBSD systems
Date:	Mon, 2 Feb 2015 08:47:35 +0100
To:	Slaven_Rezic via RT <bug-POSIX-1003 [...] rt.cpan.org>
From:	Mark Overmeer <mark [...] overmeer.net>

* Slaven_Rezic via RT (bug-POSIX-1003@rt.cpan.org) [150202 06:47]: Show quoted text

> Mon Feb 02 01:47:06 2015: Request 101866 was acted upon. > Transaction: Ticket created by SREZIC > Queue: POSIX-1003 > Subject: No bits/signum.h on FreeBSD systems > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=101866 > > > bits/signum.h does not exist on FreeBSD systems (and probably on no > non-Linux system) and prevents compilation of 1003.c. I think also that > including it is not necessary on Linux systems, too.

Hi Slaven, It's a left-over from the development work: didn't know why I couldn't get it to work. I have moved to development releases to flea some of these problems out. Maybe you have a solution for my issue (the local smoker guru's are silent) I run a script "manuals.PL" which adds tables with available constants to manual pages. This script produces a line like "module fcntl ... " when it runs. When the flag "AUTOMATED_TESTING" is on, it also dumps all those constants. I have written a script which will collect all test results and process that output. Then, it will construct huge pages showing for each constant what the value is for each of the flatforms. I have not found such a overview anywhere on internet. However... report http://www.cpantesters.org/cpan/report/61e7ecec-a925-11e4-997c-1daedfbfc7aa (which originated from you) shows module confstr has 65 constants, 0 are undefined module errno has 134 constants, 0 are undefined module poll has 11 constants, 0 are undefined module fsys_stat has 20 constants, 0 are undefined but has "AUTOMATED_TESTING" not set, which probably explains the lack of additional lines. On the other hand, report http://www.cpantesters.org/cpan/report/dbc21c4a-a8da-11e4-b6a0-e0af8971dd2f does have "AUTOMATED_TESTING" set, but does not show any line. Any idea what's happening here? -- Regards, MarkOv ------------------------------------------------------------------------ Mark Overmeer MSc MARKOV Solutions Mark@Overmeer.net solutions@overmeer.net http://Mark.Overmeer.net http://solutions.overmeer.net

Mon Feb 02 02:47:49 2015 The RT System itself - Status changed from 'new' to 'open'

Tue Feb 03 02:29:23 2015 SREZIC [...] cpan.org - Correspondence added

On 2015-02-02 02:47:49, Mark@Overmeer.net wrote: Show quoted text

> * Slaven_Rezic via RT (bug-POSIX-1003@rt.cpan.org) [150202 06:47]:

> > Mon Feb 02 01:47:06 2015: Request 101866 was acted upon. > > Transaction: Ticket created by SREZIC > > Queue: POSIX-1003 > > Subject: No bits/signum.h on FreeBSD systems > > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=101866 > > > > > bits/signum.h does not exist on FreeBSD systems (and probably on no > > non-Linux system) and prevents compilation of 1003.c. I think also that > > including it is not necessary on Linux systems, too.

> > Hi Slaven, > > It's a left-over from the development work: didn't know why I couldn't > get it to work. I have moved to development releases to flea some of > these problems out. > > Maybe you have a solution for my issue (the local smoker guru's are > silent) I run a script "manuals.PL" which adds tables with available > constants to manual pages. This script produces a line like > "module fcntl ... " > when it runs. > > When the flag "AUTOMATED_TESTING" is on, it also dumps all those > constants. I have written a script which will collect all test results > and process that output. Then, it will construct huge pages showing > for each constant what the value is for each of the flatforms. I have > not found such a overview anywhere on internet. > > However... report > http://www.cpantesters.org/cpan/report/61e7ecec-a925-11e4-997c-1daedfbfc7aa > (which originated from you) shows > module confstr has 65 constants, 0 are undefined > module errno has 134 constants, 0 are undefined > module poll has 11 constants, 0 are undefined > module fsys_stat has 20 constants, 0 are undefined > but has "AUTOMATED_TESTING" not set, which probably explains the lack > of additional lines. > > On the other hand, report > http://www.cpantesters.org/cpan/report/dbc21c4a-a8da-11e4-b6a0-e0af8971dd2f > does have "AUTOMATED_TESTING" set, but does not show any line. > > Any idea what's happening here?

CPAN::Reporter reports include only the output of the _last_ one of the configure/build/test phases. However, the *.PL handling happens in the build phase and the output would only show up if the build phase failed. Now why you see the *.PL output in my report? It seems that the makefile dependencies are somehow problematic, and parts of the build phase are re-executed when running "make test". I don't know why this is only happening on my smokers, but it is reproducible (debian/wheezy, debian/jessie). AFAIK Andreas is using debian/sid on his smokers. Maybe it also depends on the particular ExtUtils::MakeMaker version, or even on the date/time. Regards, Slaven

Tue Feb 03 04:29:00 2015 Mark [...] Overmeer.net - Correspondence added

Subject:	Re: [rt.cpan.org #101866] No bits/signum.h on FreeBSD systems
Date:	Tue, 3 Feb 2015 10:28:41 +0100
To:	Slaven_Rezic via RT <bug-POSIX-1003 [...] rt.cpan.org>
From:	Mark Overmeer <mark [...] overmeer.net>

* Slaven_Rezic via RT (bug-POSIX-1003@rt.cpan.org) [150203 07:29]: Show quoted text

> CPAN::Reporter reports include only the output of the _last_ one of the > configure/build/test phases. However, the *.PL handling happens in the > build phase and the output would only show up if the build phase failed.

That explains it. Thanks! Show quoted text

> Now why you see the *.PL output in my report?

Why is 'AUTOMATED_TESTING' not defined in your setup? Are you using very old smokers code? -- Regards, MarkOv ------------------------------------------------------------------------ Mark Overmeer MSc MARKOV Solutions Mark@Overmeer.net solutions@overmeer.net http://Mark.Overmeer.net http://solutions.overmeer.net

Tue Feb 03 15:46:19 2015 SREZIC [...] cpan.org - Correspondence added

On 2015-02-03 04:29:00, Mark@Overmeer.net wrote: Show quoted text

> * Slaven_Rezic via RT (bug-POSIX-1003@rt.cpan.org) [150203 07:29]:

> > CPAN::Reporter reports include only the output of the _last_ one of the > > configure/build/test phases. However, the *.PL handling happens in the > > build phase and the output would only show up if the build phase failed.

> > That explains it. Thanks! >

> > Now why you see the *.PL output in my report?

> > Why is 'AUTOMATED_TESTING' not defined in your setup? Are you using > very old smokers code?

I am not using a ready-made smoker at all, but just a smart wrapper around CPAN.pm. And actually you can't tell whether I am smoking, or really installing stuff for later use. And finally the semantics of 'AUTOMATED_TESTING' are not well defined, different people misuse this variable for different things. I'll include setting of this variable for your distribution in a distroprefs file, so you'll get more information in the future. BTW, 0.99_02 looks fine on FreeBSD (but not with older perl versions), so I am resolving this ticket. Regards, Slaven

Tue Feb 03 15:46:20 2015 SREZIC [...] cpan.org - Status changed from 'open' to 'resolved'

Tue Feb 03 15:46:21 2015 SREZIC [...] cpan.org - Fixed in 0.99_02 added