Re: jigdo 0.6.7(solaris8)/woody template: this is not a template
----- Forwarded message from "Heine, Dieter" <Dieter.Heine@astrium-space.com> -----
Date: Thu, 27 Jun 2002 10:38:14 +0200 (MEST)
From: "Heine, Dieter" <Dieter.Heine@astrium-space.com>
Reply-To: "Heine, Dieter" <Dieter.Heine@astrium-space.com>
Subject: Re: jigdo 0.6.7(solaris8)/woody template: this is not a template
Hello Richard,
Finally again in english for the list here comes my summary for the thread:
(Bitte prüfen und evtl in FAQ übernehmen - Kommentare wieder in deutsch :-)
jigdo 0.6.7(solaris8)/woody template: this is not a template
( combination with endless looping of jigdo-file during md5sum call)
I. Symptom and Resolution
Problem: "Successfull compilation" doesnot work for translation between
big/little endian (byte swap). Requirements for compiling on different
platforms are missing. Confusion between the choice of jigdo-port and
jigdo-file.
Solution: After Author mentioned gcc 3.0 beeing used his solaris 8 binaries were
uploaded for release 0.6.7. Since these binaries solved every problem
my compiler now still waits for an update.
II. Detailed description:
jigdo-port was recommended to be easier to port to solaris (jigdo-easy 2.1).
This may be true so I used it before trying to compile jigdo-file 0.6.7. But
then some failed downloads for iso-images with jigdo-easy ended with messages
that did not really have the character of error messages (see III below).
Since jigdo-file has the advantage of beeing a complete command line tool it
showed a more or less reasonable error message (better then jigdo-port but still
not helpfull)
jigdo-file make-image -j (URL Debian pre-release).jigdo
-t woody-i386-1.iso.template
jigdo-file make-image: `woody-i386-1.iso.template' is not a template file
Though both compilations with gcc 2.95.2 pretended to end successfull the
missing requirements and endian problems turned both binaries successfully into
harmfull junk tools. Their traps are not recognised by both scripts jigdo-easy
and jigdo-lite. Especially jigdo-easy surprisingly starts longer downloads while
jigdo-file breaks immediately with the template error message. Jigdo-lite ends
in an endless loop during md5sum checking.
III. Complete Problem description
and steps until getting solution by the author.
Platform : Ultrasparc II /Sparcation 10 (SUNW,Ultra-5_10
Generic_108528-05)
Environment: Solaris 8 04/2001+gcc 2.95.2 (Companion CD)
Problem: jigdo-easy+jigdo-port (2.1) reports after hours of download:
.... missing files ... please try other servers for completeness ...
( prompting with '>' for restart with menue for server selection.
(Retries ended with same results while the iso-tmp file was
missing. many downloaded files in vain without iso.tmp file)
Step 1: Compiling jigdo-file from Sources of the author.
README for compiling jigdo-file recommends no gui and no caching. However
several libs were missing and needed from the companion CD of solaris 8. They
may be also found on http://sunfreeware.com. Since requirements are missing the
gcc version should be 3 due to hints from the author.
Step 2: (Final solution) loading the precompiled binaries from the author
Finally another successfull compilation with gcc 2.95.2 turned out not to work
in practise. This was the moment when pre compiled binaries from the author
showed the md5sum function working for the first time. They are available now on
the jigdo web page.
jigdo-file md5sum woody-i386-1.iso.template
(100% showing progress ...)
Like every happy end the 'lite' tools at the end of the tunnel came quickly near
and enlightened a complete jigdo-lite download ;-) . Jigdo-easy/jigdo-port may
be depricated from my point of view because jigdo-lite enables server switching
in case of missing files.
IV. Summary, Views and hints
platform independend porting of free Software depends on a working GNU compiler
with an exact release. It's fatal when the environment and requirements for
different platforms differ in vital conditions for a successfull port (endian).
It's easier to search for missing libraries after the requirements are valid.
Error conditions during runtime may get really fatal. For Solaris the split off
for supporting more systems (jigdo-port) may not be the first choice in
comparison to jigdo-file because in my case it relaxed some error conditions .
If the Solaris binaries for solaris 8 of jigdo are not suitable for all Solaris
Releases there are still left some pitfalls in den sources. However maybe some
release descriptions need to be more precise and descriptive .
Kind regards Dieter Heine
Systemadministration Unix
Astrium GmbH Raumfahrt Bremen
EMail: Dieter.Heine@astrium-space.com
Fax: 49 421 539 5698
Phone: 49 421 539 4989 User Help Desk Astrium Friedrichshafen
Phone (office Bremen): on request
----- End forwarded message -----
--
__ _
|_) /| Richard Atterer
| \/¯| http://atterer.net
¯ '` ¯
--
To UNSUBSCRIBE, email to debian-cd-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: