[PATCH 1/2] --cvs-exclude for mkfs.jffs2

[PATCH 1/2] --cvs-exclude for mkfs.jffs2

[Michael Richardson]

[-- Attachment #1: Type: text/plain, Size: 3218 bytes --]



This adds an option to ignore CVS directories.
We keep bunch of our system that are hard to recreated easily from source
in CVS, and we don't really want to waste space with CVS/Entries, etc.

Man page in following patch.

Signoff: Michael Richardson <mcr@sandelman.ca>

---

 util/mkfs.jffs2.c |   28 ++++++++++++++++++++++++++++
 1 files changed, 28 insertions(+), 0 deletions(-)

applies-to: 078d53047fee81aaab9205a659805a07715076e5
092cb3518a5faa486e62a83796a280f3d45e1229
diff --git a/util/mkfs.jffs2.c b/util/mkfs.jffs2.c
index 7e60a3f..fe0cc63 100644
--- a/util/mkfs.jffs2.c
+++ b/util/mkfs.jffs2.c
@@ -69,6 +69,14 @@
 #undef crc32
 #include "crc32.h"
 
+#ifndef TRUE
+#define TRUE 1
+#endif
+#ifndef FALSE
+#define FALSE 0
+#endif
+
+
 /* Do not use the wierd XPG version of basename */
 #undef basename
 
@@ -101,6 +109,7 @@ static int verbose = 0;
 static int squash_uids = 0;
 static int squash_perms = 0;
 static int fake_times = 0;
+static int CVSexclude = FALSE;
 int target_endian = __BYTE_ORDER;
 static const char *const app_name = "mkfs.jffs2";
 static const char *const memory_exhausted = "memory exhausted";
@@ -391,6 +400,15 @@ static struct filesystem_entry *recursiv
 			continue;
 		}
 
+		/* ignore the CVS directories */
+		if(CVSexclude &&
+		   (strcmp(dp->d_name, "CVS")==0
+		    || strcmp(dp->d_name, ".KEEP_ME")==0
+		    )) {
+		  free(dp);
+		  continue;
+		}
+
 		asprintf(&hpath, "%s/%s", hostpath, dp->d_name);
 		if (lstat(hpath, &sb)) {
 			perror_msg_and_die("%s", hpath);
@@ -1171,6 +1189,7 @@ static struct option long_options[] = {
 	{"disable-compressor", 1, NULL, 'x'},
 	{"test-compression", 0, NULL, 't'},
 	{"compressor-priority", 1, NULL, 'y'},
+	{"cvs-exclude", 0, NULL, 'C'},
 	{"incremental", 1, NULL, 'i'},
 	{NULL, 0, NULL, 0}
 };
@@ -1183,6 +1202,7 @@ static char *helptext =
 	"                          not specified, the output is padded to the end of\n"
 	"                          the final erase block\n"
 	"  -r, -d, --root=DIR      Build file system from directory DIR (default: cwd)\n"
+        "  --cvs-exclude           Ignore CVS directories, and .KEEP_ME files\n"
 	"  -s, --pagesize=SIZE     Use page size (max data node size) SIZE (default: 4KiB)\n"
 	"  -e, --eraseblock=SIZE   Use erase block size SIZE (default: 64KiB)\n"
 	"  -c, --eraseblock-header=SIZE  Size of eraseblock header (default 28)\n"
@@ -1388,6 +1408,10 @@ int main(int argc, char **argv)
 				rootdir = xstrdup(optarg);
 				break;
 
+ 		        case 'C':
+			        CVSexclude = TRUE;
+			        break;
+
 			case 's':
 				page_size = strtol(optarg, NULL, 0);
 				break;
@@ -1548,6 +1572,10 @@ int main(int argc, char **argv)
 	if (!(cwd = getcwd(0, GETCWD_SIZE)))
 		perror_msg_and_die("getcwd failed");
 
+	if(verbose) {
+	  printf("Initializing image from %s (%s)\n", rootdir, cwd);
+	}
+
 	if(in_fd != -1)
 		parse_image();
 
---
0.99.9.GIT


-- 
]       ON HUMILITY: to err is human. To moo, bovine.           |  firewalls  [
]   Michael Richardson,    Xelerance Corporation, Ottawa, ON    |net architect[
] mcr@xelerance.com      http://www.sandelman.ottawa.on.ca/mcr/ |device driver[
] panic("Just another Debian GNU/Linux using, kernel hacking, security guy"); [

[-- Attachment #2: Type: application/pgp-signature, Size: 480 bytes --]

Re: [PATCH 1/2] --cvs-exclude for mkfs.jffs2

[Wolfgang Denk]

In message <v0mzjofaup.fsf@marajade.sandelman.ca> you wrote:
> 
> This adds an option to ignore CVS directories.

This doesn't make sense to me. You can and should use "cvs export" to
create the directory trees you want to package.

You use  CVS,  others  use  GIT,  or  Subversion,  or  ClearCase,  or
Monotone,  or darcs, or ... - it would not be really practical to add
special rules for everybody's favorite tool.

Best regards,

Wolfgang Denk

-- 
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de
You can observe a lot just by watchin'.                  - Yogi Berra

Re: [PATCH 1/2] --cvs-exclude for mkfs.jffs2

[Bernd Petrovitsch]

On Tue, 2005-11-29 at 11:09 +0100, Wolfgang Denk wrote:
> In message <v0mzjofaup.fsf@marajade.sandelman.ca> you wrote:
[...]
> > This adds an option to ignore CVS directories.
> 
> This doesn't make sense to me. You can and should use "cvs export" to
> create the directory trees you want to package.

Or do a `find \( -name CVS -o -name .svn \) -print0 | xargs -0r rm -rf`
before.
Especially the name `.KEEP_ME` is not even standardized (more .keep
IIRC) - not even in the CVS world AFAIK.

> You use  CVS,  others  use  GIT,  or  Subversion,  or  ClearCase,  or
> Monotone,  or darcs, or ... - it would not be really practical to add
> special rules for everybody's favorite tool.

Ar at least name and implement the option as '--exclude' similar to
`tar` and `diff`.

	Bernd
-- 
Firmix Software GmbH                   http://www.firmix.at/
mobil: +43 664 4416156                 fax: +43 1 7890849-55
          Embedded Linux Development and Services

Re: [PATCH 1/2] --cvs-exclude for mkfs.jffs2

[Jörn Engel]

On Tue, 29 November 2005 11:09:37 +0100, Wolfgang Denk wrote:
> In message <v0mzjofaup.fsf@marajade.sandelman.ca> you wrote:
> > 
> > This adds an option to ignore CVS directories.
> 
> This doesn't make sense to me. You can and should use "cvs export" to
> create the directory trees you want to package.
> 
> You use  CVS,  others  use  GIT,  or  Subversion,  or  ClearCase,  or
> Monotone,  or darcs, or ... - it would not be really practical to add
> special rules for everybody's favorite tool.

More importantly.  People may actually want to _keep_ the magical
subdirectories.

Write one tool to do one thing and do it well.  Applies here as well.

Jörn

-- 
I've never met a human being who would want to read 17,000 pages of
documentation, and if there was, I'd kill him to get him out of the
gene pool.
-- Joseph Costello

Re: [PATCH 1/2] --cvs-exclude for mkfs.jffs2

[Michael Richardson]

[-- Attachment #1: Type: text/plain, Size: 1399 bytes --]


>>>>> "Wolfgang" == Wolfgang Denk <wd@denx.de> writes:
    Wolfgang> In message <v0mzjofaup.fsf@marajade.sandelman.ca> you wrote:
    >> This adds an option to ignore CVS directories.

    Wolfgang> This doesn't make sense to me. You can and should use "cvs
    Wolfgang> export" to create the directory trees you want to package.

  yes, one could do that. I chose not to.
  It isn't like I forced this option to be on.

  cvs export requires that I actually check my changes in before I can test
them, which causes a lot of needless churn.  
  It may also mean that developers in my team will check in debug statements
and forget to remove them.  
  
    Wolfgang> You use CVS, others use GIT, or Subversion, or ClearCase, or
    Wolfgang> Monotone, or darcs, or ... - it would not be really practical
    Wolfgang> to add special rules for everybody's favorite tool.

  It would very much be practical to do this.
  I'd rather have a more general --exclude.

  Of course, you argue for removing equivalent options from rsync and other
things.  Accept the patch or not. 

-- 
]       ON HUMILITY: to err is human. To moo, bovine.           |  firewalls  [
]   Michael Richardson,    Xelerance Corporation, Ottawa, ON    |net architect[
] mcr@xelerance.com      http://www.sandelman.ottawa.on.ca/mcr/ |device driver[
] panic("Just another Debian GNU/Linux using, kernel hacking, security guy"); [

[-- Attachment #2: Type: application/pgp-signature, Size: 480 bytes --]

Re: [PATCH 1/2] --cvs-exclude for mkfs.jffs2

[Josh Boyer]

On 11/29/05, Michael Richardson <mcr@sandelman.ottawa.on.ca> wrote:
>
> >>>>> "Wolfgang" == Wolfgang Denk <wd@denx.de> writes:
>     Wolfgang> In message <v0mzjofaup.fsf@marajade.sandelman.ca> you wrote:
>     >> This adds an option to ignore CVS directories.
>
>     Wolfgang> This doesn't make sense to me. You can and should use "cvs
>     Wolfgang> export" to create the directory trees you want to package.
>
>   yes, one could do that. I chose not to.
>   It isn't like I forced this option to be on.
>
>   cvs export requires that I actually check my changes in before I can test
> them, which causes a lot of needless churn.
>   It may also mean that developers in my team will check in debug statements
> and forget to remove them.
>
>     Wolfgang> You use CVS, others use GIT, or Subversion, or ClearCase, or
>     Wolfgang> Monotone, or darcs, or ... - it would not be really practical
>     Wolfgang> to add special rules for everybody's favorite tool.
>
>   It would very much be practical to do this.
>   I'd rather have a more general --exclude.
>
>   Of course, you argue for removing equivalent options from rsync and other
> things.  Accept the patch or not.

I believe the answer is "not".

josh

Re: [PATCH 1/2] --cvs-exclude for mkfs.jffs2

[Josh Boyer]

On 11/29/05, Josh Boyer <jwboyer@gmail.com> wrote:
> On 11/29/05, Michael Richardson <mcr@sandelman.ottawa.on.ca> wrote:
> >
> > >>>>> "Wolfgang" == Wolfgang Denk <wd@denx.de> writes:
> >     Wolfgang> In message <v0mzjofaup.fsf@marajade.sandelman.ca> you wrote:
> >     >> This adds an option to ignore CVS directories.
> >
> >     Wolfgang> This doesn't make sense to me. You can and should use "cvs
> >     Wolfgang> export" to create the directory trees you want to package.
> >
> >   yes, one could do that. I chose not to.
> >   It isn't like I forced this option to be on.
> >
> >   cvs export requires that I actually check my changes in before I can test
> > them, which causes a lot of needless churn.
> >   It may also mean that developers in my team will check in debug statements
> > and forget to remove them.
> >
> >     Wolfgang> You use CVS, others use GIT, or Subversion, or ClearCase, or
> >     Wolfgang> Monotone, or darcs, or ... - it would not be really practical
> >     Wolfgang> to add special rules for everybody's favorite tool.
> >
> >   It would very much be practical to do this.
> >   I'd rather have a more general --exclude.
> >
> >   Of course, you argue for removing equivalent options from rsync and other
> > things.  Accept the patch or not.
>
> I believe the answer is "not".

To clarify, the answer is no because you've hard coded this exclude
stuff in the tool itself.  Making a more generic --exclude which reads
what to exclude from a file might be an option that is acceptable. 
Then for your specific case you could simply have "CVS" and ".KEEP_ME"
listed there.

josh

Re: [PATCH 1/2] --cvs-exclude for mkfs.jffs2

[Michael Richardson]

[-- Attachment #1: Type: text/plain, Size: 1303 bytes --]


>>>>> "Josh" == Josh Boyer <jwboyer@gmail.com> writes:
    >> I believe the answer is "not".

    Josh> To clarify, the answer is no because you've hard coded this exclude
    Josh> stuff in the tool itself.  Making a more generic --exclude which
    Josh> reads what to exclude from a file might be an option that is
    Josh> acceptable. Then for your specific case you could simply have "CVS"
    Josh> and ".KEEP_ME" listed there.

  Thank you --- I appreciate constructive feedback! 

  I'm happy to do a proper --exclude. I would have rather done it that way to
begin with, but it was hard to hack up in a moment. I posted because I wanted
to clear my queue of patches.
  I would propose, like rsync:
        --exclude=glob
        --exclude-from=FILE
        --scm-exclude=FILE                         

  Last one could be set to, for instance:
           --scm-exclude=.cvsignore
 
  (and could occur multiple times)

  [did my patch to create jffs2cpio make it out?]

-- 
]       ON HUMILITY: to err is human. To moo, bovine.           |  firewalls  [
]   Michael Richardson,    Xelerance Corporation, Ottawa, ON    |net architect[
] mcr@xelerance.com      http://www.sandelman.ottawa.on.ca/mcr/ |device driver[
] panic("Just another Debian GNU/Linux using, kernel hacking, security guy"); [

[-- Attachment #2: Type: application/pgp-signature, Size: 480 bytes --]

Re: [PATCH 1/2] --cvs-exclude for mkfs.jffs2

[Wolfgang Denk]

In message <v0sltfeb45.fsf@marajade.sandelman.ca> you wrote:
>
>     Wolfgang> You use CVS, others use GIT, or Subversion, or ClearCase, or
>     Wolfgang> Monotone, or darcs, or ... - it would not be really practical
>     Wolfgang> to add special rules for everybody's favorite tool.
> 
>   It would very much be practical to do this.
>   I'd rather have a more general --exclude.

Now *this*  is  something  completely  different.  If  you  implement
generic  exclude feature, like the "--exclude=PATTERN" optioon in GNU
tar, than this would be really useful.

Just hard-coding a specific directory name  is  nothing  else  but  a
quick and dirty hack and should IMHO not be accepted for public code.

>   Of course, you argue for removing equivalent options from rsync and other
> things.  Accept the patch or not. 

Note that "rsync and other things"  do  not  hard-code  any  specific
names.

My vote is clear: this patch should not be accepted as is.

Best regards,

Wolfgang Denk

-- 
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de
But it's real. And if it's real it can be affected ... we may not  be
able  to break it, but, I'll bet you credits to Navy Beans we can put
a dent in it.
	-- deSalle, "Catspaw", stardate 3018.2