On 04/22/10 06:20 PM, Jack Schwartz wrote:
Hi Joe and Keith.
Thanks for your comments. I've finally had time to go back and work
on this. Below is IMHO a solution which addresses all of your concerns:
On 04/ 5/10 07:44 AM, Keith Mitchell wrote:
On 04/ 5/10 05:16 AM, joseph.vl...@oracle.com wrote:
On 04/ 4/10 11:41 PM, Jack Schwartz wrote:
Hi everyone.
This easy webrev is for a fix to the DC finalizer script which
copies files into the boot_archive. It fixes how whole-directories
are excluded.
I have tested this fix by having built several images during Driver
Update testing.
Webrev:
http://cr.opensolaris.org/~schwartz/100404.1/webrev/index.html
Bug:
http://defect.opensolaris.org/bz/show_bug.cgi?id=15290
Thanks,
Jack
_______________________________________________
caiman-discuss mailing list
caiman-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
Hey Jack,
I see what I think might be a problem with your solution. What if
the item you are attempting to match is a substring match and not
the full match?
For example:
>>> excitem="AAAA/BBB/CCC"
>>> excitem.find("AA")
0
This matches "AA" even though I think you only want to match "AAAA"
I think this proposed solution might be safer:
>>> EXCLUDES = "one "
>>> for i in excitem.split('/'):
... if "AAA" == i:
... EXCLUDES = EXCLUDES + "more excludes"
...
>>> print EXCLUDES
one
>>>
Correctly does not match "AAA"
>>> for i in excitem.split('/'):
... if "AAAA" == i:
... EXCLUDES = EXCLUDES + "more excludes"
...
>>> print EXCLUDES
one more excludes
>>>
but does correcly match "AAAA"
What do you think?
If I append a / to the end of excitem and to the end of item if it
doesn't already end in /, then it all works nicely. The trailing / of
item properly "terminates" strings which would otherwise incorrectly
match. For example,
/var/pk will incorrectly show as an ancestor to /var/pkg/123,
but
/var/pk/ won't.
I tested using a test program with a pretty extensive list of test
cases, and then tested by running a build and checking added output.
If you are curious, test program and output is attached.
Please bless.
Thanks,
Jack
I'm not sure that'll work in all cases Joe? Such as if "foo/bar" were
included, and "bar" were excluded, it would match? But it's early on
a Monday and maybe I'm not thinking it through.
Another point to consider is, it doesn't *seem* like there would be
ill effects in the scenario you bring up. For example, consider the
following set of dirs:
foo
foo/bar
foo/bartwo
Suppose you wanted to include foo, include foo/bartwo, and exclude
foo/bar. Based on the code, "foo/bar" would 'match' against both
foo/bar and foo/bartwo as it looped, but at line 163 all that's added
is a "grep -v foo/bar" - so no harm done, in this case, in being
overzealous.
That said:
Jack, please use "if excitem.startswith(item):" instead of parsing
for "...find(..) == 0".
- Keith
Joe
_______________________________________________
caiman-discuss mailing list
caiman-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
Great idea Jack!
I'm good with this solution, other than the comment Keith made regarding
line 169.
Joe
_______________________________________________
caiman-discuss mailing list
caiman-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
