org-crypt and PROPERTIES location

org-crypt and PROPERTIES location

[Peter Jones]

I've started playing with org-crypt, which is very nice.

Unfortunately the org-apply-to-entry-contents function assumes that the
PROPERTIES drawer is at the bottom of the heading's contents.  There
might be a variable in org mode to control this, but all my entries have
the PROPERTIES drawer at the top, with the content underneath.

Of course, this means that org-crypt doesn't actually encrypt anything
for me.

I'm posting this to the org-mode mailing list to see if there is a
better way to get to the contents of a heading, excluding any drawers.
I'm sure there is something in the export system we could reuse.  I say
"we" because I'm more than willing to submit patches.

One last note, the comments in the org-crypt file mention using an
after-save-hook to encrypt the appropriate headings.  Shouldn't that be
a before-save-hook?

Thanks to all.

-- 
Peter Jones, http://pmade.com
pmade inc.  Louisville, CO US

Re: org-crypt and PROPERTIES location

[Carsten Dominik]

Hi Peter,

we don't have a good function for this. Also, it is not so much about
extracting the contents, more about finding the boundaries.  In some  
ways I think it would actually be better to use a tag for encryption  
instead of a property, because this would allow to encrypt everything  
except the headline.  What about we change this to a tag crypt_KEY?   
Hmm, I am not sure if encryption keys would make valid tags.....

I'd very much like to see this improved and stabilized, so that we can
include this with Org, this would be valuable to have.  I am sure John
would not mind if you did the final steps with it.  Am I right, John?

- Carsten

On Mar 20, 2009, at 1:29 AM, Peter Jones wrote:

> I've started playing with org-crypt, which is very nice.
>
> Unfortunately the org-apply-to-entry-contents function assumes that  
> the
> PROPERTIES drawer is at the bottom of the heading's contents.  There
> might be a variable in org mode to control this, but all my entries  
> have
> the PROPERTIES drawer at the top, with the content underneath.
>
> Of course, this means that org-crypt doesn't actually encrypt anything
> for me.
>
> I'm posting this to the org-mode mailing list to see if there is a
> better way to get to the contents of a heading, excluding any drawers.
> I'm sure there is something in the export system we could reuse.  I  
> say
> "we" because I'm more than willing to submit patches.
>
> One last note, the comments in the org-crypt file mention using an
> after-save-hook to encrypt the appropriate headings.  Shouldn't that  
> be
> a before-save-hook?
>
> Thanks to all.
>
> -- 
> Peter Jones, http://pmade.com
> pmade inc.  Louisville, CO US
>
>
>
> _______________________________________________
> Emacs-orgmode mailing list
> Remember: use `Reply All' to send replies to the list.
> Emacs-orgmode@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-orgmode

Re: org-crypt and PROPERTIES location

[Peter Jones]

Carsten Dominik <dominik@science.uva.nl> writes:
> we don't have a good function for this. Also, it is not so much about
> extracting the contents, more about finding the boundaries.  In some
> ways I think it would actually be better to use a tag for encryption
> instead of a property, because this would allow to encrypt everything
> except the headline.  What about we change this to a tag crypt_KEY?
> Hmm, I am not sure if encryption keys would make valid tags.....

AFAIK, the key ID is only needed for encryption.  The encrypted text
will contain a pointer to the key needed for decryption.

Therefore, I suggest that we do this:

  - Have a defcustom for the default key ID to use (I think it will be
    very common that you'll use the same key to encrypt everything)

  - Allow overriding the key ID in an inheritable property (if you need
    to encrypt a heading using a different key than the default)

  - Use a tag matcher to specify headings to encrypt (another defcustom
    that will be used to find headings that should be encrypted
    before-save)

  - Encrypt everything under the heading, including properties (even if
    the property has the key ID, it won't be needed for decryption, so
    the entire contents can be encrypted)

Am I missing anything?

> I'd very much like to see this improved and stabilized, so that we can
> include this with Org, this would be valuable to have.  I am sure John
> would not mind if you did the final steps with it.  Am I right, John?

I had a private conversation with John, and he's completely open to me
putting the finishing touches on org-crypt and getting it included in
the org distribution.

Are there any org-crypt users out there that want to throw in their
$0.02?

-- 
Peter Jones, http://pmade.com
pmade inc.  Louisville, CO US

Re: Re: org-crypt and PROPERTIES location

[Carsten Dominik]

On Mar 20, 2009, at 6:27 PM, Peter Jones wrote:

> Carsten Dominik <dominik@science.uva.nl> writes:
>> we don't have a good function for this. Also, it is not so much about
>> extracting the contents, more about finding the boundaries.  In some
>> ways I think it would actually be better to use a tag for encryption
>> instead of a property, because this would allow to encrypt everything
>> except the headline.  What about we change this to a tag crypt_KEY?
>> Hmm, I am not sure if encryption keys would make valid tags.....
>
> AFAIK, the key ID is only needed for encryption.  The encrypted text
> will contain a pointer to the key needed for decryption.
>
> Therefore, I suggest that we do this:
>
>  - Have a defcustom for the default key ID to use (I think it will be
>    very common that you'll use the same key to encrypt everything)
>
>  - Allow overriding the key ID in an inheritable property (if you need
>    to encrypt a heading using a different key than the default)
>
>  - Use a tag matcher to specify headings to encrypt (another defcustom
>    that will be used to find headings that should be encrypted
>    before-save)
>
>  - Encrypt everything under the heading, including properties (even if
>    the property has the key ID, it won't be needed for decryption, so
>    the entire contents can be encrypted)
>
> Am I missing anything?

This sounds perfect to me.

You can use the mapping API to walk through the file
and do the encryption.  This has the advantage that you could,
in principle do de- and encryption in many files in one step.
Default should be the current file, though, exactly as you suggest.

- Carsten

>
>> I'd very much like to see this improved and stabilized, so that we  
>> can
>> include this with Org, this would be valuable to have.  I am sure  
>> John
>> would not mind if you did the final steps with it.  Am I right, John?
>
> I had a private conversation with John, and he's completely open to me
> putting the finishing touches on org-crypt and getting it included in
> the org distribution.
>
> Are there any org-crypt users out there that want to throw in their
> $0.02?
>
> -- 
> Peter Jones, http://pmade.com
> pmade inc.  Louisville, CO US
>
>
>
> _______________________________________________
> Emacs-orgmode mailing list
> Remember: use `Reply All' to send replies to the list.
> Emacs-orgmode@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-orgmode

Re: org-crypt and PROPERTIES location

[news]

Peter Jones <mlists@pmade.com> writes:

>
> Are there any org-crypt users out there that want to throw in their
> $0.02?
Hi Peter,
         I wrote a function for encrypting subtrees a couple of weeks
         ago. I didn't realize someone else was already on the case.
         I guess your solution is better than mine, but where can I get
         hold of it? I don't see it in worg.
-- 
aleblanc

Re: org-crypt and PROPERTIES location

[Peter Jones]

<news@aleblanc.cotse.net> writes:
>> Are there any org-crypt users out there that want to throw in their
>> $0.02?
> Hi Peter,
>          I wrote a function for encrypting subtrees a couple of weeks
>          ago. I didn't realize someone else was already on the case.
>          I guess your solution is better than mine, but where can I get
>          hold of it? I don't see it in worg.

Actually, org-crypt was written by John Wiegley, and I'm going to be
making some changes to slightly alter how it works.  You can get John's
version here:

  http://github.com/jwiegley/dot-emacs/raw/cf72e6e004034849bd683b467b51d960d80a3e66/org-crypt.el

I'll post a message after I've made my changes to it.

-- 
Peter Jones, http://pmade.com
pmade inc.  Louisville, CO US

Re: org-crypt and PROPERTIES location

[news]

Peter Jones <mlists@pmade.com> writes:

> Actually, org-crypt was written by John Wiegley, and I'm going to be
> making some changes to slightly alter how it works.  You can get John's
> version here:
>
>   http://github.com/jwiegley/dot-emacs/raw/cf72e6e004034849bd683b467b51d960d80a3e66/org-crypt.el
>
> I'll post a message after I've made my changes to it.

Thanks. 
Here's my $0.02: symmetric encryption (i.e. passphrase based encryption).

-- 
aleblanc

Re: org-crypt and PROPERTIES location

[John Wiegley]

On Mar 20, 2009, at 9:23 AM, Carsten Dominik wrote:

> I'd very much like to see this improved and stabilized, so that we can
> include this with Org, this would be valuable to have.  I am sure John
> would not mind if you did the final steps with it.  Am I right, John?

Quite correct, it's community property! :)

John