|
|
|
|
|
|
| Author |
Message |
Young Coot
*nix forums beginner
Joined: 15 Mar 2005
Posts: 16
|
 Posted: Wed Jul 19, 2006 6:11 am Post subject:
Re: Mplayer-gtk-esound 0.99.8 cannot compile on FreeBSD 4.11
|
|
|
Earlier in the linear time track, on Tuesday 18 July 2006 09:37 pm, Conrad
J. Sabatier wrote:
| Quote: | Seriously, why this die-hard commitment to sticking with 4.11? What do
you see as the advantage(s) to this? What do you see as preventing you
from upgrading to a more recent version (say, 6.x)?
|
Okay, here's how it is: I like what I've seen of 6.1 so far, and when I
upgrade (not if, but when) it will more than likely be to the stable
version on that branch. But I hung on to 4-Stable during the trials and
tribulations of the 5.x branch, and I read the panicked posts about 5.x on
the FreeBSD mailing lists all during that time, and made up my mind that
I'd stick with 4 until the end of support. That decision worked for me.
I've had a good, stable FreeBSD machine for nearly five years, and have no
regrets about hanging on to the stable variant of 4. It's been great.
YC |
|
| Back to top |
|
 |
Conrad J. Sabatier
*nix forums Guru Wannabe
Joined: 07 Aug 2004
Posts: 142
|
| Posted: Wed Jul 19, 2006 4:37 am Post subject:
Re: Mplayer-gtk-esound 0.99.8 cannot compile on FreeBSD 4.11
|
|
|
Young Coot (Reply@Newsgroup.OnlyPlease) wrote:
[snip]
| Quote: | Interesting; thanks. Where would I look to find the syntax for such a
function, particularly the way "isinf" and "isnan" would be used?
|
Well, with an up-to-date version of FreeBSD, you would type "man isinf"
or "man isnan".
| Quote: | I'm thinking, in the future, there might be other ports that might
need one of the "round" functions to compile, particularly some vital
program that KDE depends on or something along those lines, that would
effectively freeze out the 4.x systems. Adding this "round"
functionality to the libm of 4.x would save the collective bacon of
all those (like myself) using 4.11 until the bitter end.
|
Seriously, why this die-hard commitment to sticking with 4.11? What do
you see as the advantage(s) to this? What do you see as preventing you
from upgrading to a more recent version (say, 6.x)?
--
Conrad J. Sabatier <conrads@cox.net> -- "In Unix veritas" |
|
| Back to top |
|
|
Young Coot
*nix forums beginner
Joined: 15 Mar 2005
Posts: 16
|
| Posted: Tue Jul 18, 2006 10:01 pm Post subject:
Re: Mplayer-gtk-esound 0.99.8 cannot compile on FreeBSD 4.11
|
|
|
Earlier in the linear time track, on Tuesday 18 July 2006 02:31 pm, Steven
G. Kargl wrote:
| Quote: | Replace
if (!isfinite(x))
return (x);
with
if (isinf(x) || isnan(x)
return (x);
|
Excellent; thank you.
| Quote: | a well-written application would check libm for a
functioning round().
|
Good point. That's probably what will enable the newer version of "mplayer"
to compile, is to check for that functionality and perform an alternate
calculation if it's not there.
| Quote: | I doubt that you'll find a committer
who will commit to the 4.11 branch unless it is a security
issue.
|
Which is why it would probably have to be added manually to the existing
library, either directly to "math.h" or some other user-created header file
which could be added manually as an "#include" somewhere along the chain.
YC
"fast and bulbous" |
|
| Back to top |
|
|
Steven G. Kargl
*nix forums Guru Wannabe
Joined: 02 Dec 2003
Posts: 146
|
| Posted: Tue Jul 18, 2006 9:31 pm Post subject:
Re: Mplayer-gtk-esound 0.99.8 cannot compile on FreeBSD 4.11
|
|
|
In article <12bqjuu5b4j0af9@corp.supernews.com>,
Young Coot <Reply@Newsgroup.OnlyPlease> writes:
| Quote: |
Interesting; thanks. Where would I look to find the syntax for such a
function, particularly the way "isinf" and "isnan" would be used.
|
man isinf
Replace
if (!isfinite(x))
return (x);
with
if (isinf(x) || isnan(x)
return (x);
| Quote: | I'm thinking, in the future, there might be other ports that might need one
of the "round" functions to compile, particularly some vital program that
KDE depends on or something along those lines, that would effectively
freeze out the 4.x systems. Adding this "round" functionality to the libm
of 4.x would save the collective bacon of all those (like myself) using
4.11 until the bitter end.
|
Well, a well-written application would check libm for a
functioning round(). I doubt that you'll find a committer
who will commit to the 4.11 branch unless it is a security
isssue.
--
Steve
http://troutmask.apl.washington.edu/~kargl/ |
|
| Back to top |
|
|
Young Coot
*nix forums beginner
Joined: 15 Mar 2005
Posts: 16
|
| Posted: Tue Jul 18, 2006 9:21 pm Post subject:
Re: Mplayer-gtk-esound 0.99.8 cannot compile on FreeBSD 4.11
|
|
|
Earlier in the linear time track, on Tuesday 18 July 2006 12:58 pm, Steven
G. Kargl wrote:
| Quote: | In article <12bqe0ms211s049@corp.supernews.com>,
Young Coot <Reply@Newsgroup.OnlyPlease> writes:
Earlier in the linear time track, on Tuesday 18 July 2006 12:23 pm,
Steven G. Kargl wrote:
In article <12bqcma51a1ut93@corp.supernews.com>,
Young Coot <Reply@Newsgroup.OnlyPlease> writes:
The biggest issue is the calling of the "roundf" function from "math.h"
during the later part of the compile process, a function that does not
apparently exist in the 4.x version of "math.h" but does exist in the
5.x and 6.x versions.
8==snipped==
As the person who wrote the round(3) family of functions,
you could add them to 4.11's libm but you'll need a
working version of isfinite(3). isfinite(3) was added in
5.1.
Thanks for that response and vital info: I was wondering if a new
supporting
library would be necessary. It would explain why the "round" family of
functions wasn't added to 4.x when it could have been otherwise.
That being the case, I'm also wondering what supporting structure would
be needed for "isfinite(3)" so it could reside on a 4.x system without
causing
a kernel panic during a compile. That could be messy.
I'd have to search the archives, but I believe that you would
need the fenv.h work done by David Schultz. This will be much
for difficult to backport.
You can replace the isfinite() call in round() by a suitable
combination of isinf() and isnan().
|
Interesting; thanks. Where would I look to find the syntax for such a
function, particularly the way "isinf" and "isnan" would be used?
I'm thinking, in the future, there might be other ports that might need one
of the "round" functions to compile, particularly some vital program that
KDE depends on or something along those lines, that would effectively
freeze out the 4.x systems. Adding this "round" functionality to the libm
of 4.x would save the collective bacon of all those (like myself) using
4.11 until the bitter end.
YC |
|
| Back to top |
|
|
Steven G. Kargl
*nix forums Guru Wannabe
Joined: 02 Dec 2003
Posts: 146
|
| Posted: Tue Jul 18, 2006 7:58 pm Post subject:
Re: Mplayer-gtk-esound 0.99.8 cannot compile on FreeBSD 4.11
|
|
|
In article <12bqe0ms211s049@corp.supernews.com>,
Young Coot <Reply@Newsgroup.OnlyPlease> writes:
| Quote: | Earlier in the linear time track, on Tuesday 18 July 2006 12:23 pm, Steven
G. Kargl wrote:
In article <12bqcma51a1ut93@corp.supernews.com>,
Young Coot <Reply@Newsgroup.OnlyPlease> writes:
The biggest issue is the calling of the "roundf" function from "math.h"
during the later part of the compile process, a function that does not
apparently exist in the 4.x version of "math.h" but does exist in the 5.x
and 6.x versions. It may be possible to add that definition safely
to "math.h" but I have not tried it. So, until that issue is addressed,
the compile process stops there on a 4.x system and can go no further.
As the person who wrote the round(3) family of functions,
you could add them to 4.11's libm but you'll need a
working version of isfinite(3). isfinite(3) was added in
5.1.
Thanks for that response and vital info: I was wondering if a new supporting
library would be necessary. It would explain why the "round" family of
functions wasn't added to 4.x when it could have been otherwise.
That being the case, I'm also wondering what supporting structure would be
needed for "isfinite(3)" so it could reside on a 4.x system without causing
a kernel panic during a compile. That could be messy.
|
I'd have to search the archives, but I believe that you would
need the fenv.h work done by David Schultz. This will be much
for difficult to backport.
You can replace the isfinite() call in round() by a suitable
combination of isinf() and isnan().
--
Steve
http://troutmask.apl.washington.edu/~kargl/ |
|
| Back to top |
|
|
Young Coot
*nix forums beginner
Joined: 15 Mar 2005
Posts: 16
|
| Posted: Tue Jul 18, 2006 7:41 pm Post subject:
Re: Mplayer-gtk-esound 0.99.8 cannot compile on FreeBSD 4.11
|
|
|
Earlier in the linear time track, on Tuesday 18 July 2006 12:23 pm, Steven
G. Kargl wrote:
| Quote: | In article <12bqcma51a1ut93@corp.supernews.com>,
Young Coot <Reply@Newsgroup.OnlyPlease> writes:
The biggest issue is the calling of the "roundf" function from "math.h"
during the later part of the compile process, a function that does not
apparently exist in the 4.x version of "math.h" but does exist in the 5.x
and 6.x versions. It may be possible to add that definition safely
to "math.h" but I have not tried it. So, until that issue is addressed,
the compile process stops there on a 4.x system and can go no further.
As the person who wrote the round(3) family of functions,
you could add them to 4.11's libm but you'll need a
working version of isfinite(3). isfinite(3) was added in
5.1.
|
Thanks for that response and vital info: I was wondering if a new supporting
library would be necessary. It would explain why the "round" family of
functions wasn't added to 4.x when it could have been otherwise.
That being the case, I'm also wondering what supporting structure would be
needed for "isfinite(3)" so it could reside on a 4.x system without causing
a kernel panic during a compile. That could be messy.
YC |
|
| Back to top |
|
|
Steven G. Kargl
*nix forums Guru Wannabe
Joined: 02 Dec 2003
Posts: 146
|
| Posted: Tue Jul 18, 2006 7:23 pm Post subject:
Re: Mplayer-gtk-esound 0.99.8 cannot compile on FreeBSD 4.11
|
|
|
In article <12bqcma51a1ut93@corp.supernews.com>,
Young Coot <Reply@Newsgroup.OnlyPlease> writes:
| Quote: |
The biggest issue is the calling of the "roundf" function from "math.h"
during the later part of the compile process, a function that does not
apparently exist in the 4.x version of "math.h" but does exist in the 5.x
and 6.x versions. It may be possible to add that definition safely
to "math.h" but I have not tried it. So, until that issue is addressed,
the compile process stops there on a 4.x system and can go no further.
|
As the person who wrote the round(3) family of functions,
you could add them to 4.11's libm but you'll need a
working version of isfinite(3). isfinite(3) was added in
5.1.
--
Steve
http://troutmask.apl.washington.edu/~kargl/ |
|
| Back to top |
|
|
Young Coot
*nix forums beginner
Joined: 15 Mar 2005
Posts: 16
|
| Posted: Tue Jul 18, 2006 7:21 pm Post subject:
Mplayer-gtk-esound 0.99.8 cannot compile on FreeBSD 4.11
|
|
|
Hello -
I'm posting this here for the benefit of those who might otherwise bang
their head against the wall fruitlessly trying to upgrade their 0.99.7_x
version of "mplayer-gtk-esound" to the newer 0.99.8 version on a FreeBSD
4.x box. Based on my experience, I'd say keep the older "mplayer".
to summarize briefly, I have FreeBSD 4.11 on an i386 system, last system
build was on 7/4/06, and have had a perfectly happy build of mplayer
0.99.7_15 on the system for a while. This new version in ports, in
addition to what others have written about on Usenet (particularly the
brand new "config" menu which "portupgrade" misses and drags in new strange
dependencies), apparently is built to compile on 5.x and 6.x systems but
fails pretty firmly on 4.11.
The biggest issue is the calling of the "roundf" function from "math.h"
during the later part of the compile process, a function that does not
apparently exist in the 4.x version of "math.h" but does exist in the 5.x
and 6.x versions. It may be possible to add that definition safely
to "math.h" but I have not tried it. So, until that issue is addressed,
the compile process stops there on a 4.x system and can go no further.
There were other annoyances, corrected by going into "config.h" in the
port's "work" directory and adding some definitions (such as for "PRId64"),
and I can list those if someone wants. The main point is that, for those
of us still happily using the 4.x versions of FreeBSD, this particular
version of "mplayer-gtk-esound" is not going to compile.
YC |
|
| Back to top |
|
|
Google
|
|
| Back to top |
|
|
|
|
The time now is Fri Nov 21, 2008 7:18 pm | All times are GMT
|
|
Free Credit Reports | Remortgaging | Loans | Loans | Books
|
|
Copyright © 2004-2005 DeniX Solutions SRL
|
|
|
|
Other DeniX Solutions sites:
Unix/Linux blog |
electronics forum |
medicine forum |
science forum |
|
|
Privacy Policy
|
Powered by phpBB © 2001, 2005 phpBB Group
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Forum index
»
*nix
»
BSD
»
FreeBSD
