xml-coreutils-discuss

[Xml-coreutils-discuss] Problem using xml-grep

[Douglas H. <do...@do...>]

Hi Laird,

I've got xml-coreutils 0.8 running on Mac OS X (Darwin Kernel Version
9.8.0:xnu-1228.15.4~1/RELEASE_I386) and have a variety of problems
trying to grep elements out of a 3MB valid xml file.

Full output is below, but in summary I am trying to refit the xml
utilities into my usual pattern:
  cat file | grep 'expression'

I am not finding the utilities to work as I expect.  Perhaps this is
usage or documentation.  Can you advise?  Thanks.

I cannot provide the xml file itself.  Let me know if you need a test
case or if there is debug information I can obtain on my end.

Regards,
Doug

douglas-helds-macbook-pro-2:2010Q3 douglasheld$ xml-cat core_java.xml
| xml-grep --subtree 'ControlflowRule'
<?xml version="1.0"?>
<root>











</root>
douglas-helds-macbook-pro-2:2010Q3 douglasheld$ cat core_java.xml |
grep -m 1 'ControlflowRule'
            <ControlflowRule formatVersion="3.2" language="java">
douglas-helds-macbook-pro-2:2010Q3 douglasheld$ cat core_java.xml |
xml-grep --subtree 'ControlflowRule'
xml-grep: error: segmentation fault.
douglas-helds-macbook-pro-2:2010Q3 douglasheld$ echo $?
1
douglas-helds-macbook-pro-2:2010Q3 douglasheld$

-- 
Douglas Held
do...@do...
+447986527654

Re: [Xml-coreutils-discuss] Problem using xml-grep

[Laird B. <la...@lb...>]

Hi Doug,

To make sure we're on the same page, can you do your test with the
tarball I just uploaded to the files area called
xml-coreutils-0.8b.tar.gz ? (It's the current snapshot of the code)

Cheers,
Laird.
-- 
http://www.lbreyer.com
The mind boggles, but does the Boggle mind? The mind boggles...

Re: [Xml-coreutils-discuss] Problem using xml-grep

[Laird B. <la...@lb...>]

On Sep 13 2010, Douglas Held wrote:
> All right, I built 0.8b and the symptom looks similar.  Please let me
> know what debugging information I can provide. Thanks!

Thanks, now we're ready to talk :)

> douglas-helds-macbook-pro-2:2010Q3 douglasheld$ cat core_java.xml |
> xml-grep --subtree 'ControlflowRule'
> xml-grep: error: segmentation fault.

This is a bug in xml-grep. I can only track it down if I have a sample
XML which triggers it. Since you can't show me the XML, you can try
giving me a stack trace in gdb if you're comfortable with that.

gdb --args xml-grep 'ControlflowRule' core_java.xml

Type "run", and after it segfaults, type "where".

> douglas-helds-macbook-pro-2:2010Q3 douglasheld$ xml-cat core_java.xml
> | xml-grep --subtree 'ControlflowRule'
> <?xml version="1.0"?>
> <root>
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> </root>
> douglas-helds-macbook-pro-2:2010Q3 douglasheld$ cat core_java.xml |
> grep -m 1 'ControlflowRule'
>             <ControlflowRule formatVersion="3.2" language="java">

This is correct (ie intended) output. The xml-grep doesn't grep
the tags, but only the strings in attributes or text nodes. For
example, you should be able to get the output you expect with 

xml-cat core_java.xml | xml-grep '3.2'

Most commands accept an XPATH type expression that lets you drill
down in the tree structure.
   
Some other commands that might do what you're looking for in your 
example are

xml-ls '://ControlflowRule' 'core_java.xml'
xml-cp 'core_java.xml' '://ControlflowRule[1]' stdout

Laird.
-- 
http://www.lbreyer.com
The mind boggles, but does the Boggle mind? The mind boggles...

Re: [Xml-coreutils-discuss] Problem using xml-grep

[Laird B. <la...@lb...>]

On Sep 14 2010, Douglas Held wrote:
> OK, now we're talking indeed.
> 
> First, the problem is not reproducible.  I must have had the old
> libraries loaded, or old binaries in the buffer cache or
> i-don't-know-what.  But second, since trying the gdb command I get the
> correct output with the blanks.
> 
> Thanks for your time.  I'll study the examples and ask a more
> intelligent question next time.  If I can reproduce any problem I'll
> send a gdb stack trace or test case.

Excellent! Feel free to ask nonintelligent :) questions about how the
tools work, as it documents use-cases which can lead to UI improvements.

I've actually just remembered why xml-grep doesn't grep the tree
structure in the regex - it's to be specified on the command line after
the file name:

xml-grep --subtree '.*' 'core_java.xml' '://ControlflowRule'
xml-cat core_java.xml | xml-grep --subtree '.*' '://ControlflowRule'

If you "man 7 xml-coreutils", the section on the Common Unified
Command Line Convention should explain the gist of it.

Cheers,
Laird.
-- 
http://www.lbreyer.com
The mind boggles, but does the Boggle mind? The mind boggles...