[gentoo-user] [OT] confusing RE doesn't work in diff

[gentoo-user] [OT] confusing RE doesn't work in diff

[Zhang Weiwu]

My RE must be wrong but I can hardly successfully match a whole line
like the below example:

This works:
zhangweiwu@Bach:/tmp$ egrep "\b*\/[*].*[*]\/\b*" Calendar.php
        /* $Id: class.boalarm.inc.php,v 1.1.1.1 2005/03/18 09:17:36
dawnlinux Exp $ */
        /* $Id: class.boalarm.inc.php,v 1.11 2004/05/23 14:51:27
ralfbecker Exp $ */

This doesn't work:
zhangweiwu@Bach:/tmp$ egrep "^\b*\/[*].*[*]\/\b*$" Calendar.php

any hint?

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] [OT] confusing RE doesn't work in diff

[Christoph Gysin]

Zhang Weiwu wrote:
> This works:
> zhangweiwu@Bach:/tmp$ egrep "\b*\/[*].*[*]\/\b*" Calendar.php
>         /* $Id: class.boalarm.inc.php,v 1.1.1.1 2005/03/18 09:17:36
> dawnlinux Exp $ */
>         /* $Id: class.boalarm.inc.php,v 1.11 2004/05/23 14:51:27
> ralfbecker Exp $ */
> 
> This doesn't work:
> zhangweiwu@Bach:/tmp$ egrep "^\b*\/[*].*[*]\/\b*$" Calendar.php
> 
> any hint?

^ only matches at the beginning of a line. This isn't the case in your 
examples.

Christoph
-- 
echo mailto: NOSPAM !#$.'<*>'|sed 's. ..'|tr "<*> !#:2" org@fr33z3
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] [OT] confusing RE doesn't work in diff

[Mariusz Pękala]

[-- Attachment #1: Type: text/plain, Size: 852 bytes --]

On 2005-07-25 17:50:37 +0800 (Mon, Jul), Zhang Weiwu wrote:
> My RE must be wrong but I can hardly successfully match a whole line
> like the below example:
> 
> This works:
> zhangweiwu@Bach:/tmp$ egrep "\b*\/[*].*[*]\/\b*" Calendar.php
>         /* $Id: class.boalarm.inc.php,v 1.1.1.1 2005/03/18 09:17:36
> dawnlinux Exp $ */
>         /* $Id: class.boalarm.inc.php,v 1.11 2004/05/23 14:51:27
> ralfbecker Exp $ */
> 
> This doesn't work:
> zhangweiwu@Bach:/tmp$ egrep "^\b*\/[*].*[*]\/\b*$" Calendar.php
> 
> any hint?

man grep:
The symbol \b matches the empty string at the edge of a word

It seems that '/' is not considered to be a part of a word.

I suppose you may use "^[[:space:]]*/[*].*[*]/[[:space:]]*$" Calendar.php


-- 
No virus found in this outgoing message.
Checked by 'grep -i virus $MESSAGE'
Trust me.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-user] [OT] confusing RE doesn't work in diff

[Zhang Weiwu]

Mariusz Pękala wrote:
> On 2005-07-25 17:50:37 +0800 (Mon, Jul), Zhang Weiwu wrote:
> > > My RE must be wrong but I can hardly successfully match a whole line
> > > like the below example:
> > > 
> > > This works:
> > > zhangweiwu@Bach:/tmp$ egrep "\b*\/[*].*[*]\/\b*" Calendar.php
> > >         /* $Id: class.boalarm.inc.php,v 1.1.1.1 2005/03/18 09:17:36
> > > dawnlinux Exp $ */
> > >         /* $Id: class.boalarm.inc.php,v 1.11 2004/05/23 14:51:27
> > > ralfbecker Exp $ */
> > > 
> > > This doesn't work:
> > > zhangweiwu@Bach:/tmp$ egrep "^\b*\/[*].*[*]\/\b*$" Calendar.php
> > > 
> > > any hint?
> 
> man grep:
> The symbol \b matches the empty string at the edge of a word
> 
> It seems that '/' is not considered to be a part of a word.
> 
> I suppose you may use "^[[:space:]]*/[*].*[*]/[[:space:]]*$" Calendar.php

Thank you! Your RE suggested does work and I discovered this [[:space:]] pretty useful. I missed it in man grep.

However there is another strange behavor. This works:

zhangweiwu@Bach:/tmp$ cat header.inc.php
<?php
        /* $Id: header.inc.php.template,v 1.58 2005/07/08 05:00:14 milosch Exp $ */
zhangweiwu@Bach:/tmp$ grep "^[[:space:]]*/[*].*[*]/[[:space:]]*$" header.inc.php
        /* $Id: header.inc.php.template,v 1.58 2005/07/08 05:00:14 milosch Exp $ */

And this doesn't work:

zhangweiwu@Bach:/tmp$ :>empty
zhangweiwu@Bach:/tmp$ diff -I "^[[:space:]]*/[*].*[*]/[[:space:]]*$" header.inc.php empty
1,2d0
< <?php
<       /* $Id: header.inc.php.template,v 1.58 2005/07/08 05:00:14 milosch Exp $ */

in man it's being said, 
       -I RE  --ignore-matching-lines=RE
              Ignore changes whose lines all match RE.

I read man grep and it seems there are two sets of RE, extended and basic. But it's not mensioned in diff(1) wheather or not diff(1) takes extended or basic RE. I am using Debian so it should be gnu diff.

Any thing more to teach me?

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] [OT] confusing RE doesn't work in diff

[Christoph Gysin]

Zhang Weiwu wrote:
> zhangweiwu@Bach:/tmp$ diff -I "^[[:space:]]*/[*].*[*]/[[:space:]]*$" header.inc.php empty
> 1,2d0
> < <?php
> <       /* $Id: header.inc.php.template,v 1.58 2005/07/08 05:00:14 milosch Exp $ */
> 
> in man it's being said, 
>        -I RE  --ignore-matching-lines=RE
>               Ignore changes whose lines all match RE.
> 
> I read man grep and it seems there are two sets of RE, extended and basic. But it's not mensioned in diff(1) wheather or not diff(1) takes extended or basic RE. I am using Debian so it should be gnu diff.

If it's not mentioned, it possibly uses basic regular expressions.

try:

$ diff -I '/\* \$Id.*\$ \*/' header.inc.php empty

Christoph
-- 
echo mailto: NOSPAM !#$.'<*>'|sed 's. ..'|tr "<*> !#:2" org@fr33z3
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] [OT] confusing RE doesn't work in diff

[Mariusz Pękala]

[-- Attachment #1: Type: text/plain, Size: 2208 bytes --]

On 2005-07-27 09:43:25 +0200 (Wed, Jul), Christoph Gysin wrote:
> Zhang Weiwu wrote:
> >zhangweiwu@Bach:/tmp$ diff -I "^[[:space:]]*/[*].*[*]/[[:space:]]*$" 
> >header.inc.php empty
> >1,2d0
> >< <?php
> ><       /* $Id: header.inc.php.template,v 1.58 2005/07/08 05:00:14 milosch 
> >Exp $ */
> >
> >in man it's being said, 
> >       -I RE  --ignore-matching-lines=RE
> >              Ignore changes whose lines all match RE.
> >
> >I read man grep and it seems there are two sets of RE, extended and basic. 
> >But it's not mensioned in diff(1) wheather or not diff(1) takes extended 
> >or basic RE. I am using Debian so it should be gnu diff.
> 
> If it's not mentioned, it possibly uses basic regular expressions.
> 
> try:
> 
> $ diff -I '/\* \$Id.*\$ \*/' header.inc.php empty
> 
> Christoph

Indeed strange. BUT the clue lies in (after many strange results) in
'man grep' where, on my host, it says exactly:

-I regexp
   Ignore changes that just insert or delete lines that match regexp.

Notice the word: 'changes'.
It seems that it means: ignore if WHOLE GROUP of DELETED OR INSERTED
lines matches (every added or deleted line) the provided regexp.

In your case the first line of header.inc.php is included in change-group
but does not matches RE.

Try (to be very simple):
$ diff -I '^.*/.*$' header empty
1,2d0
< <?php
<         /* $Id: header.inc.php.template,v 1.58 2005/07/08 05:00:14 milosch Exp $ */

$ echo '<?php' > empty
$ diff -I '^.*/.*$' header empty
(nothing)
$ :>empty
$ diff -I '^.*php.*$' header empty
(nothing - because BOTH lines, considered CHANGED OR DELETED have the
word 'php' inside)

Now the difficult part: how to solve the real problem. :-)

* You may compare grepped files (with grepped out /* $Id:$ */)
* You may try to filter diff output... but I don't like this (by
  intuition - I may be wrong)
* experiment with various options of diff. If only output (not retcode)
  matters then good candidate maybe may be --changed-group-format, which is
  weel described in 'info diff'
* or something other...

HTH

-- 
No virus found in this outgoing message.
Checked by 'grep -i virus $MESSAGE'
Trust me.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]