[syslinux] OT:Question on gen-id.sh $id

[Michael D. Setzer II]

On 2 Apr 2011 at 10:31, Gene Cumm wrote:

Date sent:      	Sat, 2 Apr 2011 10:31:25 -0400
Subject:        	Re: [syslinux] OT:Question on gen-id.sh $id
From:           	Gene Cumm <gene.cumm at gmail.com>
To:             	For discussion of Syslinux and tftp-hpa 
<syslinux at zytor.com>
Copies to:      	"Michael D. Setzer II" 
<mikes at kuentos.guam.net>

> On Sat, Apr 2, 2011 at 05:11, Michael D. Setzer II
> <mikes at kuentos.guam.net> wrote:
> > Somewhat off-topic.
> > I've noticed for some time, when one recompiles the syslinux, that
> > it uses an id that is created by the gen-id.sh script that seems to
> > be a hex string by making the id variable equal to the $tim
> > variable.
> >
> > I had modified the script to append the the prexx of the directory
> > to the $tim variable.  Tested it with the pre15 and pre17 that came
> > out today. The first one would pull out the pre after the -, but the
> > second one which also works would also handle if the directory
> > was without a pre, but does assume that format is constant?
> >
> > This way the hdt reports the $tim with the pre version, rather than
> > the random $tim variable.
> >
> > The original line is id="$tim"
> >
> > if test -z "$id"; then
> >  id="$tim-`echo ${PWD} | rev | cut -f2 -d/ | rev | cut -f3 -d-`"
> >  id="$tim-`echo ${PWD} | rev | cut -f2 -d/ | rev | cut -b15-19`"
> > fi
> > echo "$id"
> >
> > Perhaps some other method would be better, but this is additional
> > info.
> 
> Please read doc/distrib.txt if you haven't already.
> 
> The reason you're seeing the hex string is that you didn't set other
> "things" first.  If you must compile from source (rather than just
> rebuilding the installer binaries with "make clean; make installer" ),
> I'd advise setting it to something unique (and doesn't match the
> official binaries).  I know Debian sets "things" so a string
> Debian-YEAR-MONTH-DAY is in the binaries to distinguish it from the
> official precompiled binaries in the official binary/source package.
> I believe this is the next best practice to just rebuilding the
> binaries but I don't know the official opinion.  I believe Debian
> policy is that everything needs to be built from source.
> 
> I believe a part of this comes from the fact that historically some
> common build systems have incorrectly built the binaries.  I'd
> consider distributions distributing these broken binaries to look like
> the official binaries which end users report problems to Syslinux
> (rather than their distro) similar to me signing your name for
> something and having a third party blaming you for my mistake.  By me
> signing my name, it would help you to know it came from me.
> 
> -- 
> -Gene

Thanks for the response. I had read that in the past, and that is 
why I didn't do something to change the method that is currently 
used, but added on the pre?? information.  This way it still shows 
the same hex string just adds the -pre17 to it on the latest.

I've made 65 updates on working on the next release of my g4l 
project, and have been using various pre versions as I build it and 
record the the version of syslinux/isolinux in a text file, but that 
requires booting the system, and then looking at the file. With this 
method the hdt that is part of the boot menu would be able to 
show not only that it is 4.04 with a hex string to note it was a user 
build version, but that it was a build using the -pre17 base. 

The only thing that I have changed, is I add a line to the cat.c file 
to let the user know to just press the enter key to return to the 
menu if they display one of the text files. I have noticed that a 
build changes a number of files, not jus the cat.c32 and it isn't just 
the hex number. I am guessing that this is due to compiler and 
library versions. 

I generally wait for a full released version of syslinux and other 
support programs before a full release, but in building alphas, I 
like to keep up on the changes others are working on.

Again, Thanks for the response and pointing me to information. I 
haven't done a make spotless, but will see if that does anything 
different. 

+----------------------------------------------------------+
  Michael D. Setzer II -  Computer Science Instructor      
  Guam Community College  Computer Center                  
  mailto:mikes at kuentos.guam.net                            
  mailto:msetzerii at gmail.com
  http://www.guam.net/home/mikes
  Guam - Where America's Day Begins                        
  G4L Disk Imaging Project maintainer 
  http://sourceforge.net/projects/g4l/
+----------------------------------------------------------+

http://setiathome.berkeley.edu (Original)
Number of Seti Units Returned:  19,471
Processing time:  32 years, 290 days, 12 hours, 58 minutes
(Total Hours: 287,489)

BOINC at HOME CREDITS
SETI        10512813.624229   |   EINSTEIN     5650097.270851
ROSETTA      2972600.997745   |   ABC          5123465.984288