unoccluded names in gnats mail archives?

unoccluded names in gnats mail archives?

[Christopher Faylor]

Does anyone know where to look to fix this problem?

cgf

----- Forwarded message from Ralph Smith <smith@arete-dc.com> -----

From: Ralph Smith <smith@arete-dc.com>
To: gdb-prs-owner@sources.redhat.com
Date: Fri, 3 Nov 2000 09:10:48 -0500 (EST)

Hello,

I have noticed that most of the addresses in message headers for the
gdb... mailing lists are rewritten to discourage spam.  However, the
"Resent-Reply-To" header which is automatically inserted in gdb-prs
messages is not being rewritten.  It seems that this should be made
consistent.

While the gdb-prs mailing list is still new and small, would it be
feasible to manually fix addresses in existing messages before archives or
digests are compiled?

Thanks for your attention,

				Ralph Smith

----- End forwarded message -----

Re: unoccluded names in gnats mail archives?

[Jason Molenda]

On Fri, Nov 03, 2000 at 04:09:19PM -0500, Christopher Faylor wrote:
> Does anyone know where to look to fix this problem?
> 


infra/ml-archives/messages.mrc.

I use the AddressModifyCode feature to do the address munging; it
doesn't do its magic on Resent-Cc and Resent-Reply-To (I don't
think these are "real" headers, defined by the mail RFCs).  If you
look at messages.mrc, you'll see a huge list of excluded headers
(Excs)[1], I'd recommend adding these two headers to that list.

[1] One of my big gripes about most mail archives on the net - they
are wasteful of vertical space.  You'll have all kinds of navigation
doodads at the top, then you'll have all of the headers in the mail
note, and the finally you'll see some quoted text at the top of the
message.

Mhonarc doesn't give you a way to say "Only show me THESE headers",
so I've had to add every dumb-ass header that people's MUA's stick in
mail notes over the years. :-)



Jason

Re: unoccluded names in gnats mail archives?

[Hans-Peter Nilsson]

On Fri, 3 Nov 2000, Jason Molenda wrote:
> On Fri, Nov 03, 2000 at 04:09:19PM -0500, Christopher Faylor wrote:
> > Does anyone know where to look to fix this problem?

> infra/ml-archives/messages.mrc.

I hope saying this is unnecessary, but just don't forget that this is
CVS-controlled.  Please don't modify "manually".  Right, Jason?

brgds, H-P

Re: unoccluded names in gnats mail archives?

[Chris Faylor]

On Fri, Nov 03, 2000 at 04:56:17PM -0500, Hans-Peter Nilsson wrote:
>On Fri, 3 Nov 2000, Jason Molenda wrote:
>> On Fri, Nov 03, 2000 at 04:09:19PM -0500, Christopher Faylor wrote:
>> > Does anyone know where to look to fix this problem?
>
>> infra/ml-archives/messages.mrc.
>
>I hope saying this is unnecessary, but just don't forget that this is
>CVS-controlled.  Please don't modify "manually".  Right, Jason?

That seems to be where the rule for the translation happens but I don't
understand why the translation is selective.  Most of the headers
are munged correctly.  It only seems to be the Resent-Reply-To part
that is not.

cgf

Re: unoccluded names in gnats mail archives?

[Jim Kingdon]

> Resent-Cc and Resent-Reply-To (I don't think these are "real" headers,
> defined by the mail RFCs).

Resent-cc and Resent-reply-to are as real as they come, RFC-wise.  See
RFC822.  It doesn't get much more real than that :-).

The problem is that if you just omit resent-* headers, you end up with
a misleading result, since the message is actually delivered to the
"resent-cc" rather than the "cc" if both are present.  Maybe not a big
deal in the context of the archiver, though, since we already know
what list it was sent to....

Re: unoccluded names in gnats mail archives?

[Chris Faylor]

On Fri, Nov 03, 2000 at 05:07:51PM -0500, Chris Faylor wrote:
>On Fri, Nov 03, 2000 at 04:56:17PM -0500, Hans-Peter Nilsson wrote:
>>On Fri, 3 Nov 2000, Jason Molenda wrote:
>>> On Fri, Nov 03, 2000 at 04:09:19PM -0500, Christopher Faylor wrote:
>>> > Does anyone know where to look to fix this problem?
>>
>>> infra/ml-archives/messages.mrc.
>>
>>I hope saying this is unnecessary, but just don't forget that this is
>>CVS-controlled.  Please don't modify "manually".  Right, Jason?
>
>That seems to be where the rule for the translation happens but I don't
>understand why the translation is selective.  Most of the headers
>are munged correctly.  It only seems to be the Resent-Reply-To part
>that is not.
>>
>>infra/ml-archives/messages.mrc.
>>
>>I use the AddressModifyCode feature to do the address munging; it
>>doesn't do its magic on Resent-Cc and Resent-Reply-To (I don't
>>think these are "real" headers, defined by the mail RFCs).  If you
>>look at messages.mrc, you'll see a huge list of excluded headers
>>(Excs)[1], I'd recommend adding these two headers to that list.


Duh.  You mean I have to move my eyes up several lines and read what's
right on the screen in front of me?

I've added those headers too.

Personally, I would have implemented this as "change all occurrences of
an '@' in the header to 'whatever'" rather than making it only work on
specific parts of the header.  The consequences of allowing an actual
email into the archives is pretty dire, imo.

cgf

Re: unoccluded names in gnats mail archives?

[Jason Molenda]

On Fri, Nov 03, 2000 at 05:18:18PM -0500, Chris Faylor wrote:

> I've added those headers too.
> 
> Personally, I would have implemented this as "change all occurrences of
> an '@' in the header to 'whatever'" rather than making it only work on
> specific parts of the header.  The consequences of allowing an actual
> email into the archives is pretty dire, imo.
> 

 [ and JimK informs us that Resent-* are in RFC822 ]

It could be considered a bug in mhonarc, the header munging feature
is a recent addition and we're probably the first to notice this
omission.  At least Resent-* should be modified if they're listed
in 822.  mhonarc is just a pile o' perl code if someone wants to
do the change, I'm sure it wouldn't be especially difficult to spot
the munged-header-list and add these.

Hans-Peter wrote:

> I hope saying this is unnecessary, but just don't forget that this is
> CVS-controlled.  Please don't modify "manually".  Right, Jason?

Oh yes, I probably should have said that explicitly.  Thanks for the
reminder.

J

Re: unoccluded names in gnats mail archives?

[Christopher Faylor]

On Fri, Nov 03, 2000 at 02:29:20PM -0800, Jason Molenda wrote:
>> I hope saying this is unnecessary, but just don't forget that this is
>> CVS-controlled.  Please don't modify "manually".  Right, Jason?
>
>Oh yes, I probably should have said that explicitly.  Thanks for the
>reminder.

That's ok.  I knew this already.

I just had to figure out how to change those stupid ,v files in the CVS
directory to put in all of the right info.  It really shouldn't be so
hard to make changes.  It's really hard to "cat" those files correctly.

cgf
:-)

Re: unoccluded names in gnats mail archives?

[Phil Edwards]

On Fri, Nov 03, 2000 at 05:52:27PM -0500, Christopher Faylor wrote:
> 
> I just had to figure out how to change those stupid ,v files in the CVS
> directory to put in all of the right info.  It really shouldn't be so
> hard to make changes.  It's really hard to "cat" those files correctly.
> 
> cgf
> :-)

And when working on source code, it's even harder to get them to compile.
Took me for-EVER to get a bootstrap the first time I looked at this project.


Phil

Re: unoccluded names in gnats mail archives?

[Chris Faylor]

On Fri, Nov 03, 2000 at 04:56:17PM -0500, Hans-Peter Nilsson wrote:
>On Fri, 3 Nov 2000, Jason Molenda wrote:
>> On Fri, Nov 03, 2000 at 04:09:19PM -0500, Christopher Faylor wrote:
>> > Does anyone know where to look to fix this problem?
>
>> infra/ml-archives/messages.mrc.
>
>I hope saying this is unnecessary, but just don't forget that this is
>CVS-controlled.  Please don't modify "manually".  Right, Jason?

That seems to be where the rule for the translation happens but I don't
understand why the translation is selective.  Most of the headers
are munged correctly.  It only seems to be the Resent-Reply-To part
that is not.

cgf

Re: unoccluded names in gnats mail archives?

[Chris Faylor]

On Fri, Nov 03, 2000 at 05:07:51PM -0500, Chris Faylor wrote:
>On Fri, Nov 03, 2000 at 04:56:17PM -0500, Hans-Peter Nilsson wrote:
>>On Fri, 3 Nov 2000, Jason Molenda wrote:
>>> On Fri, Nov 03, 2000 at 04:09:19PM -0500, Christopher Faylor wrote:
>>> > Does anyone know where to look to fix this problem?
>>
>>> infra/ml-archives/messages.mrc.
>>
>>I hope saying this is unnecessary, but just don't forget that this is
>>CVS-controlled.  Please don't modify "manually".  Right, Jason?
>
>That seems to be where the rule for the translation happens but I don't
>understand why the translation is selective.  Most of the headers
>are munged correctly.  It only seems to be the Resent-Reply-To part
>that is not.
>>
>>infra/ml-archives/messages.mrc.
>>
>>I use the AddressModifyCode feature to do the address munging; it
>>doesn't do its magic on Resent-Cc and Resent-Reply-To (I don't
>>think these are "real" headers, defined by the mail RFCs).  If you
>>look at messages.mrc, you'll see a huge list of excluded headers
>>(Excs)[1], I'd recommend adding these two headers to that list.


Duh.  You mean I have to move my eyes up several lines and read what's
right on the screen in front of me?

I've added those headers too.

Personally, I would have implemented this as "change all occurrences of
an '@' in the header to 'whatever'" rather than making it only work on
specific parts of the header.  The consequences of allowing an actual
email into the archives is pretty dire, imo.

cgf

Re: unoccluded names in gnats mail archives?

[Christopher Faylor]

On Fri, Nov 03, 2000 at 02:29:20PM -0800, Jason Molenda wrote:
>> I hope saying this is unnecessary, but just don't forget that this is
>> CVS-controlled.  Please don't modify "manually".  Right, Jason?
>
>Oh yes, I probably should have said that explicitly.  Thanks for the
>reminder.

That's ok.  I knew this already.

I just had to figure out how to change those stupid ,v files in the CVS
directory to put in all of the right info.  It really shouldn't be so
hard to make changes.  It's really hard to "cat" those files correctly.

cgf
:-)

Re: unoccluded names in gnats mail archives?

[Phil Edwards]

On Fri, Nov 03, 2000 at 05:52:27PM -0500, Christopher Faylor wrote:
> 
> I just had to figure out how to change those stupid ,v files in the CVS
> directory to put in all of the right info.  It really shouldn't be so
> hard to make changes.  It's really hard to "cat" those files correctly.
> 
> cgf
> :-)

And when working on source code, it's even harder to get them to compile.
Took me for-EVER to get a bootstrap the first time I looked at this project.


Phil

unoccluded names in gnats mail archives?

[Christopher Faylor]

Does anyone know where to look to fix this problem?

cgf

----- Forwarded message from Ralph Smith <smith@arete-dc.com> -----

From: Ralph Smith <smith@arete-dc.com>
To: gdb-prs-owner@sources.redhat.com
Date: Fri, 3 Nov 2000 09:10:48 -0500 (EST)

Hello,

I have noticed that most of the addresses in message headers for the
gdb... mailing lists are rewritten to discourage spam.  However, the
"Resent-Reply-To" header which is automatically inserted in gdb-prs
messages is not being rewritten.  It seems that this should be made
consistent.

While the gdb-prs mailing list is still new and small, would it be
feasible to manually fix addresses in existing messages before archives or
digests are compiled?

Thanks for your attention,

				Ralph Smith

----- End forwarded message -----

Re: unoccluded names in gnats mail archives?

[Jason Molenda]

On Fri, Nov 03, 2000 at 05:18:18PM -0500, Chris Faylor wrote:

> I've added those headers too.
> 
> Personally, I would have implemented this as "change all occurrences of
> an '@' in the header to 'whatever'" rather than making it only work on
> specific parts of the header.  The consequences of allowing an actual
> email into the archives is pretty dire, imo.
> 

 [ and JimK informs us that Resent-* are in RFC822 ]

It could be considered a bug in mhonarc, the header munging feature
is a recent addition and we're probably the first to notice this
omission.  At least Resent-* should be modified if they're listed
in 822.  mhonarc is just a pile o' perl code if someone wants to
do the change, I'm sure it wouldn't be especially difficult to spot
the munged-header-list and add these.

Hans-Peter wrote:

> I hope saying this is unnecessary, but just don't forget that this is
> CVS-controlled.  Please don't modify "manually".  Right, Jason?

Oh yes, I probably should have said that explicitly.  Thanks for the
reminder.

J

Re: unoccluded names in gnats mail archives?

[Hans-Peter Nilsson]

On Fri, 3 Nov 2000, Jason Molenda wrote:
> On Fri, Nov 03, 2000 at 04:09:19PM -0500, Christopher Faylor wrote:
> > Does anyone know where to look to fix this problem?

> infra/ml-archives/messages.mrc.

I hope saying this is unnecessary, but just don't forget that this is
CVS-controlled.  Please don't modify "manually".  Right, Jason?

brgds, H-P

Re: unoccluded names in gnats mail archives?

[Jason Molenda]

On Fri, Nov 03, 2000 at 04:09:19PM -0500, Christopher Faylor wrote:
> Does anyone know where to look to fix this problem?
> 


infra/ml-archives/messages.mrc.

I use the AddressModifyCode feature to do the address munging; it
doesn't do its magic on Resent-Cc and Resent-Reply-To (I don't
think these are "real" headers, defined by the mail RFCs).  If you
look at messages.mrc, you'll see a huge list of excluded headers
(Excs)[1], I'd recommend adding these two headers to that list.

[1] One of my big gripes about most mail archives on the net - they
are wasteful of vertical space.  You'll have all kinds of navigation
doodads at the top, then you'll have all of the headers in the mail
note, and the finally you'll see some quoted text at the top of the
message.

Mhonarc doesn't give you a way to say "Only show me THESE headers",
so I've had to add every dumb-ass header that people's MUA's stick in
mail notes over the years. :-)



Jason

Re: unoccluded names in gnats mail archives?

[Jim Kingdon]

> Resent-Cc and Resent-Reply-To (I don't think these are "real" headers,
> defined by the mail RFCs).

Resent-cc and Resent-reply-to are as real as they come, RFC-wise.  See
RFC822.  It doesn't get much more real than that :-).

The problem is that if you just omit resent-* headers, you end up with
a misleading result, since the message is actually delivered to the
"resent-cc" rather than the "cc" if both are present.  Maybe not a big
deal in the context of the archiver, though, since we already know
what list it was sent to....