LinuxQuestions.org
Visit Jeremy's Blog.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware
User Name
Password
Slackware This Forum is for the discussion of Slackware Linux.

Notices


Reply
  Search this Thread
Old 06-13-2018, 12:32 PM   #1411
GazL
Senior Member
 
Registered: May 2008
Posts: 4,754
Blog Entries: 14

Rep: Reputation: Disabled

Quote:
Originally Posted by GazL View Post
I won't go as far as to request rng-tools, but please can we have support for rngd startup in rc.S for those of us who prefer to leverage hardware rngd rather than haveged's software based approach.

Something like:
Code:
# Start the entropy daemon:
if [ -x /etc/rc.d/rc.rngd ]; then
  # Start the rngd entropy daemon:
  /etc/rc.d/rc.rngd start
elif [ -x /etc/rc.d/rc.haveged ]; then
  # Start the haveged entropy daemon:
  /etc/rc.d/rc.haveged start
fi
One probably doesn't want to be running both as the interaction between the two isn't fully understood, hence the 'elif'.
Just quoting myself to add some afterthoughts:

rngd needs to start after udev, otherwise /dev/hwrandom doesn't exist. But I noticed that haveged is currently being started before /dev/random has been seeded at the end of rc.S. I wonder whether in might not be better to start haveged in rc.M? After udev, but before rc.inet1 gets called seems like a sensible place for it.

Last edited by GazL; 06-13-2018 at 12:40 PM.
 
1 members found this post helpful.
Old 06-13-2018, 02:11 PM   #1412
upnort
Member
 
Registered: Oct 2014
Distribution: Slackware, Proxmox, Debian, CentOS, Ubuntu MATE
Posts: 643

Rep: Reputation: Disabled
Bug report: rc.atd

Pat,

Thank you for adding rc.atd and rc.crond.

In rc.atd:

As Is
/usr/sbin/atd $CROND_OPTS

Change To
/usr/sbin/atd $ATD_OPTS
 
2 members found this post helpful.
Old 06-13-2018, 05:13 PM   #1413
qunying
Member
 
Registered: Jun 2002
Distribution: Slackware
Posts: 140

Rep: Reputation: 77
If possible, I would like to see GCC 8.1 gets into current, as its gnat compiler will be more or less on par the GNAT community 2018 release.
 
1 members found this post helpful.
Old 06-13-2018, 05:22 PM   #1414
Darth Vader
Senior Member
 
Registered: May 2008
Location: Romania
Distribution: DARKSTAR Linux 2008.1
Posts: 2,567

Rep: Reputation: 1087Reputation: 1087Reputation: 1087Reputation: 1087Reputation: 1087Reputation: 1087Reputation: 1087Reputation: 1087
Quote:
Originally Posted by qunying View Post
If possible, I would like to see GCC 8.1 gets into current, as its gnat compiler will be more or less on par the GNAT community 2018 release.
Our BDFL said somewhere that this thing is out of discussion, because the GCC 8.1 is unable to compile the packages included in Slackware without heavy patches everywhere.

So, GCC 8.1 is so modern, that the developers around of World still has to adapt their software to it.

OK, maybe it is not the GNAT community case, but not many (if any) programs from Slackware are written in this programming language.
 
Old 06-13-2018, 07:59 PM   #1415
qunying
Member
 
Registered: Jun 2002
Distribution: Slackware
Posts: 140

Rep: Reputation: 77
Sorry, I miss the thread that have the discussion on GCC 8.1.
 
Old 06-13-2018, 09:37 PM   #1416
upnort
Member
 
Registered: Oct 2014
Distribution: Slackware, Proxmox, Debian, CentOS, Ubuntu MATE
Posts: 643

Rep: Reputation: Disabled
Pat,

I would like to propose that certain rc.d scripts using command line parameters use a file in /etc/default/. I am including some files below that I have used for some years. For these snippets I restored the parameters to the stock defaults.

Code:
# /etc/default/syslogd

# Parameters to support /etc/rc.d/rc.syslogd.
# Use this file to override defaults.

# -m 0 disables the -- MARK -- timestamp.
# -c disables suppressing repeated messages

#SYSLOGD_OPTIONS="-m 0 -c"

# -c 3 displays level 'error' or higher messages on console
# -x disable broken EIP translation

KLOGD_OPTIONS="-c 3 -x"
Code:
# /etc/default/sendmail

# Parameters to support /etc/rc.d/rc.sendmail.
# Use this file to override defaults.

QUEUE_TIME="25m"
Code:
# /etc/default/smartmontools

# Parameters to support /etc/rc.d/rc.smartd.
# Use this file to override defaults.

# Refer to the man page for configuring a separate log file.
#SMARTD_OPTIONS="-l local3"
Code:
# /etc/default/sysctl

# Parameters to support /etc/rc.d/rc.S.
# Use this file to override defaults.

# -q Use this option to not display the values to stdout.
# -e Use this option to ignore errors about unknown keys.
# --system Load settings from all system configuration files.

SYSCTL_OPTIONS="-e --system"
Proposed rc.S patch: (perhaps move the entire sysctl snippet into a new rc.sysctl script?)

Code:
 # Configure kernel parameters:
+SYSCTL_OPTIONS="-e --system"
+# Source user defined parameters.
+if [ -r /etc/default/sysctl ]; then
+  . /etc/default/sysctl
+fi
 if [ -x /sbin/sysctl -a -r /etc/sysctl.conf ]; then
-  echo "Configuring kernel parameters:  /sbin/sysctl -e --system"
-  /sbin/sysctl -e --system
+  echo "Configuring kernel parameters:  /sbin/sysctl $SYSCTL_OPTIONS"
+  /sbin/sysctl $SYSCTL_OPTIONS
 elif [ -x /sbin/sysctl  ]; then
-  echo "Configuring kernel parameters:  /sbin/sysctl -e --system"
+  echo "Configuring kernel parameters:  /sbin/sysctl $SYSCTL_OPTIONS"
   # Don't say "Applying /etc/sysctl.conf" or complain if the file doesn't exist
-  /sbin/sysctl -e --system 2> /dev/null | grep -v "Applying /etc/sysctl.conf"
+  /sbin/sysctl $SYSCTL_OPTIONS 2> /dev/null | grep -v "Applying /etc/sysctl.conf"
 fi
Please let me know of other scripts if this idea is palatable.
 
1 members found this post helpful.
Old 06-14-2018, 04:03 AM   #1417
Petri Kaukasoina
Member
 
Registered: Mar 2007
Posts: 336

Rep: Reputation: 183Reputation: 183
Quote:
Originally Posted by upnort View Post
Pat,

Thank you for adding rc.atd and rc.crond.
If a user happens to run some other progam named with the same name 'atd', it will break the logic. '/etc/rc.d/rc.atd start' won't start the real atd service, and '/etc/rc.d/rc.atd stop' will kill the other innocent atd, too. Maybe:

Code:
-  if ! /usr/bin/pgrep --ns $$ --exact atd 1> /dev/null 2> /dev/null ; then
+  if ! /usr/bin/pgrep --ns $$ --exact --euid daemon atd 1> /dev/null 2> /dev/null ; then
and

Code:
-  /usr/bin/pkill --ns $$ --exact atd 2> /dev/null
+  /usr/bin/pkill --ns $$ --exact --euid daemon atd 2> /dev/null
to only consider the atd run as 'daemon'.

And the same with rc.crond, except with '--euid root'.
 
1 members found this post helpful.
Old 06-14-2018, 05:15 AM   #1418
GazL
Senior Member
 
Registered: May 2008
Posts: 4,754
Blog Entries: 14

Rep: Reputation: Disabled
Pat, thank you for the rng changes. much appreciated.

I've made a number of requests of late and you've been very accommodating with them all, so I was in two minds whether to mentions this one or not, but at the risk of being overly demanding/annoying I'll throw another suggestion your way. If you want to ignore it then that's fine.


There's a number of things done in rc.M that could be classed as system maintenance: font cache, icon cache, gtk modules and the like that may require running after software or font changes on the system. For convenience I've been separating them out into a rc.maintenance script which makes it easier to re-run them if required.

I invoke them from inittab as follows:
Code:
# Script to run when going multi user.
rc:2345:wait:/etc/rc.d/rc.M

# System maintenance.
m1:2345:once:/etc/rc.d/rc.maintenance
.. but they could equally be done by the usual if [ -x ....] construct in rc.M itself.


There are some minor differences/enhancements between what's currently in rc.M and what I've written in rc.maintenance, but it pretty much does the same thing. The HOME=/root on update-gdk-pixbuf-loaders is to stop /.cache being created (which some people were complaining about on this forum recently).

Anyway, my patch attached for your consideration. I've been using it a couple of years now.

Note: because I invoke rc.maintenance from inittab there's no if [ -x ... ] construct included in the rc.M patch, so if you like the idea of separating it out, but not the idea of running it from inittab then that would need to be added.


And with that, I'll leave you in peace for a while...
Unless I spot anything broken.

Last edited by GazL; 07-19-2018 at 07:26 AM.
 
Old 06-14-2018, 06:34 AM   #1419
ReaperX7
LQ Guru
 
Registered: Jul 2011
Location: California
Distribution: Slackware64-Current
Posts: 6,437
Blog Entries: 15

Rep: Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000
Quote:
Originally Posted by Darth Vader View Post
Our BDFL said somewhere that this thing is out of discussion, because the GCC 8.1 is unable to compile the packages included in Slackware without heavy patches everywhere.

So, GCC 8.1 is so modern, that the developers around of World still has to adapt their software to it.

OK, maybe it is not the GNAT community case, but not many (if any) programs from Slackware are written in this programming language.
GCC has a well known issue when new versions of it arrive that break various packages and builds. I am curious if this continues as a trend if it might be wise to switch compiling packages to LLVM for clang and clang++ as LLVM, unless it too has version release build issues as well.
 
Old 06-14-2018, 07:01 AM   #1420
a4z
Senior Member
 
Registered: Feb 2009
Posts: 1,683

Rep: Reputation: 706Reputation: 706Reputation: 706Reputation: 706Reputation: 706Reputation: 706Reputation: 706
Quote:
Originally Posted by ReaperX7 View Post
GCC has a well known issue when new versions of it arrive that break various packages and builds. I am curious if this continues as a trend if it might be wise to switch compiling packages to LLVM for clang and clang++ as LLVM, unless it too has version release build issues as well.
I would expect less compatibility for CLang/LLVM than with gcc. CLang/LLVM cares less if Linux and the Linux landscape compiles, they care what google and apple need.
also, I do not know what actual problem Pat meant Darth Vader referred to, often it's not allone gcc but software where smart people thought they use some non-standard but in one version existing behavior.
but in general gcc has less issues as some GNU haters spread in their propaganda . Of courser, therefore there is no AppleGCC, because it is AppleClang ;-)
 
Old 06-14-2018, 09:23 AM   #1421
qunying
Member
 
Registered: Jun 2002
Distribution: Slackware
Posts: 140

Rep: Reputation: 77
For the GCC 8.1 issues, please read the thread:
https://www.linuxquestions.org/quest...02/page20.html
 
1 members found this post helpful.
Old 06-14-2018, 10:04 AM   #1422
mlpa
Member
 
Registered: May 2008
Location: Aveiro
Distribution: Slackware
Posts: 524

Rep: Reputation: 45
Did not see this request here yet.
It is rather simple, the gvim application is in "Other" category in my menu.
It can be fixed by Office in the Categories part of the gvim.desktop file.
 
Old 06-14-2018, 10:44 AM   #1423
a4z
Senior Member
 
Registered: Feb 2009
Posts: 1,683

Rep: Reputation: 706Reputation: 706Reputation: 706Reputation: 706Reputation: 706Reputation: 706Reputation: 706
Quote:
Originally Posted by qunying View Post
For the GCC 8.1 issues, please read the thread:
https://www.linuxquestions.org/quest...02/page20.html
thanks, there is nothing that serious

ad Pats message, staying on 7.3.1 is for the moment OK,
it's the latest compiler in the devtoolset from RHEL and the compiler on SUSE/openSUSE leap 15, so it's on enterprise level
and there is indeed no huge advantage in going to 8.1 at the moment, but 8.2 should not be that far away either ;-)
 
Old 06-14-2018, 10:08 PM   #1424
ReaperX7
LQ Guru
 
Registered: Jul 2011
Location: California
Distribution: Slackware64-Current
Posts: 6,437
Blog Entries: 15

Rep: Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000Reputation: 2000
Quote:
Originally Posted by a4z View Post
thanks, there is nothing that serious

ad Pats message, staying on 7.3.1 is for the moment OK,
it's the latest compiler in the devtoolset from RHEL and the compiler on SUSE/openSUSE leap 15, so it's on enterprise level
and there is indeed no huge advantage in going to 8.1 at the moment, but 8.2 should not be that far away either ;-)
From what I know about LLVM and Clang, and have read, it is a fairly agnostic C compiler with ports on just about every system out there including BSD, Linux, Windows, Darwin, etc. The myth that it is tailored to Apple and Google only is a myth, and only a myth. Yes, they sponsored a lot of the development, but then again, often at times big projects often get sponsors to help pay for things. As far as compatibility, I have heard VERY little about if anything compiled with clang and clang++ does anything out of the ordinary. Both have to conform to the ISO/IEC C and C++ standards as it is, so there will probably no issues with how compatible something is or isn't. The only major difference I've seen is LLVM tends to be more friendly with debugging and output logs for developers to trace back problems in the code more thoroughly.

Often at times when newer GCC versions are released, patches will need to be applied to correct variations in how the language used in the code varies from generation to generation. The aspect 8.1 breaks things will probably mean if 8.2 was to come out soon, you'd still need patches to build with 8.2 unless the developers of each package update to build with that new versions which may or may not happen. Even then you may need patches from 8.1 to 8.2 anyway.

The only issue ever raised was by Richard Stallman who argued in favor of GCC strictly because it was licensed under GPL and GNU licenses and was completely free software, which is moot because LLVM is also completely free software under the University of Illinois - NCSA Open Source license.

Last edited by ReaperX7; 06-14-2018 at 10:12 PM.
 
Old 06-15-2018, 01:09 AM   #1425
gmgf
Senior Member
 
Registered: Jun 2012
Location: Bergerac, France
Distribution: Slackware
Posts: 1,573

Rep: Reputation: 614Reputation: 614Reputation: 614Reputation: 614Reputation: 614Reputation: 614
cmake-3.11.4:

https://blog.kitware.com/cmake-3-11-...-for-download/
https://cmake.org/files/v3.11/cmake-3.11.4.tar.gz
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
[SOLVED] Requests for -current (20151216) rworkman Slackware 3441 12-28-2017 03:50 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware

All times are GMT -5. The time now is 09:18 PM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration