bill@persci.UUCP (09/14/85)
*** FOLLOW THIS LINE WITH YOUR MESSAGE ***
I see that yet another "shar" script has been posted to net.sources. I would
like to see if the net could agree on a standard format for shar files,
especially for the first line. I use a simple script to extract shar files
from within postings, but it looks for a line that starts with:
#!/bin/sh
or something like that. It does *NOT* work with shar files that start with
a colon in column 1 of the first line! I'm seeing a few of these coming over
the net now, and although it doesn't take a tremendous effort to extract these,
it is a little annoying after having received and used the "unshar" script.
--
William Swan {ihnp4,decvax,allegra,...}!uw-beaver!tikal!persci!billjjg@security.UUCP (Jeff Glass) (09/16/85)
In article <396@persci.UUCP> bill@persci.UUCP writes: > I see that yet another "shar" script has been posted to net.sources. I would > like to see if the net could agree on a standard format for shar files, > especially for the first line. I use a simple script to extract shar files > from within postings, but it looks for a line that starts with: > #!/bin/sh If you replace the sed command in your version of unshar with this sed command, it will be able to recognize shar scripts which begin with a comment line ( "#" or ":" ) or with a "cut here" line. It also recognizes a line consisting of "-- " as the beginning of a signature, and stops the unshar. sed is an amazing program, even if it does remind me of assembly language... sed -n -e "/^#/b copy" -e "/^:/b copy" \ -e "/^--*[ ]*[Cc][Uu][Tt]/b comment" \ -e d \ -e :comment -e "s/^./#/" \ -e :copy -e '/^-- $/q' -e p -e n -e "b copy" \ ${1-} | /bin/sh -- security!jjg@mitre-bedford.ARPA (MIL) {allegra,ihnp4,utzoo,philabs,uw-beaver}!linus!security!jjg (UUCP)
jimb@amdcad.UUCP (Jim Budler) (09/16/85)
In article <396@persci.UUCP> bill@persci.UUCP writes: >... >from within postings, but it looks for a line that starts with: > #!/bin/sh >or something like that. It does *NOT* work with shar files that start with >a colon in column 1 of the first line! I'm seeing a few of these coming over And what are the poor folk with v7 supposed to do? v7 sh recognizes a ':' as a comment and '#: program not found' for your solution. -- Jim Budler Advanced Micro Devices, Inc. (408) 749-5806 UUCPnet: {ucbvax,decwrl,ihnp4,allegra,intelca}!amdcad!jimb Compuserve: 72415,1200 "... Don't sue me, I'm just the piano player!...."
campbell@maynard.UUCP (Larry Campbell) (09/17/85)
> >... > >from within postings, but it looks for a line that starts with: > > #!/bin/sh > >or something like that. It does *NOT* work with shar files that start with > >a colon in column 1 of the first line! I'm seeing a few of these coming over > > And what are the poor folk with v7 supposed to do? v7 sh recognizes > a ':' as a comment and '#: program not found' for your solution. > -- > Jim Budler > UUCPnet: {ucbvax,decwrl,ihnp4,allegra,intelca}!amdcad!jimb It doesn't handle all the cases (like metachars after the #) but I wrote the smallest possible C program: "main(){}", called it #, and put it in /usr/local/bin. -- Larry Campbell decvax!genrad The Boston Software Works, Inc. \ 120 Fulton St. seismo!harvard!wjh12!maynard!campbell Boston MA 02109 / / ihnp4 cbosgd ARPA: campbell%maynard.uucp@harvard.arpa
bill@persci.UUCP (09/17/85)
In article <4048@amdcad.UUCP> jimb@amdcad.UUCP (Jim Budler) writes: >>... >>from within postings, but it looks for a line that starts with: >> #!/bin/sh >>or something like that. It does *NOT* work with shar files that start with >>a colon in column 1 of the first line! I'm seeing a few of these coming over >And what are the poor folk with v7 supposed to do? v7 sh recognizes >a ':' as a comment and '#: program not found' for your solution. Jim, if you re-read my previous posting, you may see that I'm not so concerned about *which* format is used as I am about a *standard* format. The problem (and it *is* a small one) is that an "unshar" script must be able to handle all of the formats that people are using. Whether I type "csh foo", "sh foo", "foo", or "unshar foo", it doesn't matter to me (forgive grammer, plz). I would like to be able to treat (unpack) shar files one way, all the time. Preferably not by editing them first to remove all the stuff at the top (I am lazy and impatient, and vi and emacs are incredibly slow to come up), hence the "unshar" script. 'Nuff said? -- William Swan {ihnp4,decvax,allegra,...}!uw-beaver!tikal!persci!bill
ken@rochester.UUCP (Blittermuis) (09/17/85)
In article <4048@amdcad.UUCP> jimb@amdcad.UUCP (Jim Budler) writes: >And what are the poor folk with v7 supposed to do? v7 sh recognizes >a ':' as a comment and '#: program not found' for your solution. Modify the unshar script to filter out lines starting with # or use the C unshar. Ken -- UUCP: ..!{allegra,decvax,seismo}!rochester!ken ARPA: ken@rochester.arpa USnail: Dept. of Comp. Sci., U. of Rochester, NY 14627. Voice: Ken!
ahby@meccts.UUCP (Shane P. McCarron) (09/18/85)
In article <396@persci.UUCP> bill@persci.UUCP writes: >it looks for a line that starts with: > #!/bin/sh >or something like that. It does *NOT* work with shar files that start with >a colon in column 1 of the first line! On our Sys III Plexus, csh will attempt to execute any shell that has a '#' at the beginning of the first line. I assume that some other systems have a similar anomaly. However, starting the first line with a ':' will eliminate this problem. That is why some scripts generate a colon in the first line. -- Shane P. McCarron Minnesota Educational Computing Corporation - Technical Services UUCP ihnp4!dicomed!meccts!ahby
mason@utcsri.UUCP (Dave Mason) (09/19/85)
In article <153@maynard.UUCP> campbell@maynard.UUCP (Larry Campbell) writes: >> >... >> >from within postings, but it looks for a line that starts with: >> > #!/bin/sh >> >... >> And what are the poor folk with v7 supposed to do? v7 sh recognizes >> a ':' as a comment and '#: program not found' for your solution. >> UUCPnet: {ucbvax,decwrl,ihnp4,allegra,intelca}!amdcad!jimb (Jim Budler) > >It doesn't handle all the cases (like metachars after the #) but I wrote >the smallest possible C program: "main(){}", called it #, and put it >in /usr/local/bin. true(1) already does this & is 0 bytes long. Just link to it, or cat /dev/null ># ; chmod +x # -- Usenet: {dalcs dciem garfield musocs qucis sask titan trigraph ubc-vision utzoo watmath allegra cornell decvax decwrl ihnp4 uw-beaver} !utcsri!mason Dave Mason, U. Toronto CSRI CSNET: mason@Toronto ARPA: mason%Toronto@CSNet-Relay
peter@graffiti.UUCP (Peter da Silva) (09/26/85)
> It doesn't handle all the cases (like metachars after the #) but I wrote > the smallest possible C program: "main(){}", called it #, and put it > in /usr/local/bin. That's not a solution, since the #: messages only annoy. Metacharacters are the things that cause most of the problems. Shar files should start with : '...'.
dca@edison.UUCP (David C. Albrecht) (11/08/85)
> In article <396@persci.UUCP> bill@persci.UUCP writes: > >... > >from within postings, but it looks for a line that starts with: > > #!/bin/sh > >or something like that. It does *NOT* work with shar files that start with > >a colon in column 1 of the first line! I'm seeing a few of these coming over > > And what are the poor folk with v7 supposed to do? v7 sh recognizes > a ':' as a comment and '#: program not found' for your solution. > -- Write a friggin sed script for god sake, we are not supposed to be helpless here this is .sources after all. David Albrecht
jimb@amdcad.UUCP (Jim Budler) (11/15/85)
Boy do you get your news late!! I wrote that a long time ago.
BUT:
1) Why is it that the computer that does not recognize #! /bin/sh
has to write a sed script to allow for the original poster
writing an inadequate script?
2) Computers that recognize '#! /bin/sh' to direct a script to sh
also recognize ':' for the same purpose. Computers that do not
recognize '#! /bin/sh' also recognize ':'.
In the quest for a standard shar format, shouldn't a potential developer
work for the lowest common denominator?
Therefore:
a) Shar scripts should be 'sh' scripts, because some unlucky
people don't have 'csh', but everyone has 'sh'.
b) Shar scripts, since they follow item 'a)' above should use ':'
for the reasons mentioned in '2)', above.
By the way, I get by quite well. I have BSD4.2, csh, unshar, and sed. But I
believe in the practices of Columbia Univ. i.e. try to make your
product compatible for the most people possible, even if that requires
more effort on the developer's part.
--
Jim Budler
Advanced Micro Devices, Inc.
(408) 749-5806
Usenet: {ucbvax,decwrl,ihnp4,allegra,intelca}!amdcad!jimb
Compuserve: 72415,1200
Bogus newsgroup: net.news: Move to end of .newsrc[yn^L]?
Don't be dictators, use thought.tony@xios.UUCP (Keeper Of News) (11/22/85)
As an addendum to the request for a standard shar format, could we eliminate the comments after the 'fi's that some shar programs generate. Some primative sh's (ie. the one I use) croak on this. And I'm sure others do too. -- ------------------------------------------------------------------------------- Tony Lill Keeper of News @ Xios Systems Corporation 1024 Morrison Drive, Ottawa, Ontario Canada, K2H 9E5 (613) 726-8850 utzoo!dciem!nrcaero!xios!tony ------------------------------------------------------------------------------- Not the edge of the world, but we can see it from here.
dv@well.UUCP (David W. Vezie) (12/03/85)
echo #define notdef >> /usr/include/stdio.h
Could I propose a "standard" for articles posted to all.sources?
(I realize that by asking, I'm expecting a negative answer... :-). It
is actually simple:
1, entire article is a shar file (meaning, all you should have
to do is strip off the header and feed to sh).
2, the order of the files in the shar file are as such:
a, README - description of what the package contains
(this is usually found ABOVE the shar package)
b, man pages
c, Makefiles
d, Header files
e, Source files
3, even if it is one file, package it as a shar file with a
README. That way I don't have to use an editor with it,
I can just run it through unshar. It also means that I
don't have to save the article itself to have all the
information about the software.
4, use a shell archiver that puts an "exit 0" at the end of
the article (before the .signature).
Comments?
---
David W. Vezie
{dual|hplabs}!well!dv - Whole Earth 'Lectronics Link, Sausalito, CA
(4 lines, 114 chars)avolio@decuac.UUCP (Frederick M. Avolio) (12/03/85)
In article <314@well.UUCP>, dv@well.UUCP (David W. Vezie) writes: > Could I propose a "standard" for articles posted to all.sources? Not only can you ... you did! > 1, entire article is a shar file Good idea -- has to be done by poster, not in software > 2, the order of the files in the shar file are as such: ... Why? I don't see why it matters as long as they are in there. Why order the files? I say "shar * > ~/sharfile" > 3, even if it is one file, package it as a shar file with a > README. ... Very good idea -- again, has to be done by poster, not in software > 4, use a shell archiver that puts an "exit 0" at the end of > the article (before the .signature). Gary Perlman's does this as well as puts time stamp, user name, etc. in a shell comment and allows one to create a shar file that checks character counts, etc. at unshar time. Remember, Gary's is one that people scoffed at because it is a "complicated" C program instead of the simple 8 liner distributed with news software: for i do echo "echo x - $i" echo "sed 's/^X//' >$i <<'*-*-END-of-$i-*-*'" sed 's/^/X/' $i echo "*-*-END-of-$i-*-*" done echo exit The above can easily be changed to add an exit 0 line. But Gary's is fast and has some nice features. And that is why I like it. That and, as Gary said in a recent magazine article, it's nice to be mentioned on the manual page :-) ! -- Fred @ DEC Ultrix Applications Center {decvax,seismo,cbosgd}!decuac!avolio
dave@lsuc.UUCP (David Sherman) (12/06/85)
For the benefits of those of us who sometimes read news at low
speeds, please put the man page first in the shar file. That way
it's easy to find and read. Maybe a README can go before, but not
the C source. Thanks.
Dave Sherman
--
{ ihnp4!utzoo pesnta utcs hcr decvax!utcsri } !lsuc!dave