UNC Pathname Handling within Applications

UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 1417 bytes --]

Hi,

I'm trying to find the best approach to solve a bug reported against
NEdit: 
https://sourceforge.net/tracker/index.php?func=detail&aid=873188&group_id=11005&atid=111005

The reporter complains that NEdit cannot open files from network
shares using Cygwin's double-slash notation. That is not suprising,
since NEdit silently drops any doubled slash before using the path.


I know that NEdit is not the only application ignoring multiplied
slashes, so I wonder how this is normally handled with Cygwin. Indeed,
a simple test on Linux seems to indicate that fopen() accepts any
number of slashes, so this is not isolated to NEdit.

- The user guide ("Mapping path names") does tell me something about
what's to expect but little about implementation.
- Googling brought me zilch, all involved words seem to be much too
generic to find anything, and Google does not search for slashes.
- I was also looking for another document (Posix?) to help here but
came up empty.


My questions:
- Is there any standard way to approach this problem? Has it come up
before in other applications?
- Do you know a resource which explains how Posix apps are expected to
handle paths like this?
- Does Cygwin offer another way around this problem? Could cygpath be
of any help?


Thank you very much for any pointers you can give me!


Thorsten
-- 
Auch Hunger ist Krieg.
    - Willy Brandt

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: UNC Pathname Handling within Applications

[Christopher Faylor]

On Thu, Jul 15, 2004 at 05:32:32AM +0200, Thorsten Haude wrote:
>I'm trying to find the best approach to solve a bug reported against
>NEdit: 
>https://sourceforge.net/tracker/index.php?func=detail&aid=873188&group_id=11005&atid=111005
>
>The reporter complains that NEdit cannot open files from network
>shares using Cygwin's double-slash notation. That is not suprising,
>since NEdit silently drops any doubled slash before using the path.
>
>I know that NEdit is not the only application ignoring multiplied
>slashes, so I wonder how this is normally handled with Cygwin. Indeed,
>a simple test on Linux seems to indicate that fopen() accepts any
>number of slashes, so this is not isolated to NEdit.

Why would the fact that linux can open files with any number of slashes
indicative of a problem with other applications?  Cygwin can open files
with any number of slashes too as long as they are not the first slash.

>- The user guide ("Mapping path names") does tell me something about
>what's to expect but little about implementation.

Implementation of what?  It isn't obvious what you are expecting cygwin
to do or how you are expecting it to help.  By the time Cygwin sees the
file spec the extra slash has been removed.  There is no amount of help
that cygwin can provide at that point.

>- Googling brought me zilch, all involved words seem to be much too
>generic to find anything, and Google does not search for slashes.
>- I was also looking for another document (Posix?) to help here but
>came up empty.
>
>
>My questions:
>- Is there any standard way to approach this problem?

This is an application problem, not a windows or cygwin problem.  Double
slashes at the beginnning of a path name are allowed to mean something
different and if an application doesn't recognize that fact, that is at
least arguably a bug.

So, no, there is no "standard" way to approach the problem.  You have
to modify the application if you want it to recognize this syntax.

Alternately, you can just mount your network shares:

mount -b '//foo/bar' /bar

and refer to files as /bar/blah , bypassing the need to use double slashes.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 2341 bytes --]

Hi,

* Christopher Faylor wrote (2004-07-15 06:15):
>On Thu, Jul 15, 2004 at 05:32:32AM +0200, Thorsten Haude wrote:
>>I know that NEdit is not the only application ignoring multiplied
>>slashes, so I wonder how this is normally handled with Cygwin. Indeed,
>>a simple test on Linux seems to indicate that fopen() accepts any
>>number of slashes, so this is not isolated to NEdit.
>
>Why would the fact that linux can open files with any number of slashes
>indicative of a problem with other applications?  Cygwin can open files
>with any number of slashes too as long as they are not the first slash.

I only wanted to point out that this might not be the first time that
this particular problem came up.


>>- The user guide ("Mapping path names") does tell me something about
>>what's to expect but little about implementation.
>
>Implementation of what?  It isn't obvious what you are expecting cygwin
>to do or how you are expecting it to help.  By the time Cygwin sees the
>file spec the extra slash has been removed.  There is no amount of help
>that cygwin can provide at that point.

Well, Cygwin does already provide services for converting filenames,
so I thought it might be worthwhile to ask whether similar services
exist for the point in question.


>>- Googling brought me zilch, all involved words seem to be much too
>>generic to find anything, and Google does not search for slashes.
>>- I was also looking for another document (Posix?) to help here but
>>came up empty.
>>
>>
>>My questions:
>>- Is there any standard way to approach this problem?
>
>This is an application problem, not a windows or cygwin problem.  Double
>slashes at the beginnning of a path name are allowed to mean something
>different and if an application doesn't recognize that fact, that is at
>least arguably a bug.

Where is it defined that double slashes at the beginning of a path are
allowed to mean something? Is this C? Posix? To find the source of
this convention could really help to find the best way to remove this
problem.


Thanks!


Thorsten
-- 
Rarely do we find people who willingly engage in hard, solid thinking.
There is an almost universal quest for easy answers and half-baked
solutions. Nothing pains some people more than having to think.
    - Martin Luther King

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: UNC Pathname Handling within Applications

[Christopher Faylor]

On Thu, Jul 15, 2004 at 06:48:36AM +0200, Thorsten Haude wrote:
>Hi,
>
>* Christopher Faylor wrote (2004-07-15 06:15):
>>On Thu, Jul 15, 2004 at 05:32:32AM +0200, Thorsten Haude wrote:
>>>I know that NEdit is not the only application ignoring multiplied
>>>slashes, so I wonder how this is normally handled with Cygwin. Indeed,
>>>a simple test on Linux seems to indicate that fopen() accepts any
>>>number of slashes, so this is not isolated to NEdit.
>>
>>Why would the fact that linux can open files with any number of slashes
>>indicative of a problem with other applications?  Cygwin can open files
>>with any number of slashes too as long as they are not the first slash.
>
>I only wanted to point out that this might not be the first time that
>this particular problem came up.
>
>
>>>- The user guide ("Mapping path names") does tell me something about
>>>what's to expect but little about implementation.
>>
>>Implementation of what?  It isn't obvious what you are expecting cygwin
>>to do or how you are expecting it to help.  By the time Cygwin sees the
>>file spec the extra slash has been removed.  There is no amount of help
>>that cygwin can provide at that point.
>
>Well, Cygwin does already provide services for converting filenames,
>so I thought it might be worthwhile to ask whether similar services
>exist for the point in question.

The only services that are available are for translating cygwin path
names to and from Windows path names.  Filenames beginning with two
slashes are both valid cygwin path names and windows path names so there
is no conversion routine necessary.

>>This is an application problem, not a windows or cygwin problem.  Double
>>slashes at the beginnning of a path name are allowed to mean something
>>different and if an application doesn't recognize that fact, that is at
>>least arguably a bug.
>
>Where is it defined that double slashes at the beginning of a path are
>allowed to mean something?  Is this C?  Posix?  To find the source of
>this convention could really help to find the best way to remove this
>problem.

http://www.opengroup.org/onlinepubs/007908799/xbd/glossary.html#tag_004_000_196

This says that an implementation is allowed to consider two leading
slashes as special.  You still have to look at Nedit's code and fix it
so that two leading slashes are not compressed into one, however.  I
don't know why knowing that someone says it's official helps much.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 329 bytes --]

Hi,

* Christopher Faylor wrote (2004-07-15 07:18):
>I don't know why knowing that someone says it's official helps much.

It doesn't I just wanted to make sure whether there are any other
issues that might come up later.

Thank you for your help.


Thorsten
-- 
Guns don't protect freedom, people protect freedom.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: UNC Pathname Handling within Applications

[Shankar Unni]

Thorsten Haude wrote:

> - Is there any standard way to approach this problem? Has it come up
> before in other applications?

I don't know why you are even trying to normalize the paths like this. 
Just hand the thing off to the OS. Usually, both the user and the OS 
know what it is they are trying to do.

For your program to try to mediate between them, and try to "understand 
and correct" the user's request is almost always inappropriate.

Don't even try to normalize "\" to "/" or vice-versa. Windows APIs (the 
low-level ones) know perfectly well how to handle forward-slashes as 
directory separators..

> - Do you know a resource which explains how Posix apps are expected to
> handle paths like this?

It doesn't matter. Just pass the names, as is, to the underlying OS API.

POSIX specifically says that *two* leading slashes "implementation 
defined" behavior, and this is specifically to allow NetBIOS-like (and 
AFS-like) names in POSIX implementations.


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 608 bytes --]

Hi,

* Shankar Unni wrote (2004-07-16 00:52):
>Thorsten Haude wrote:
>>- Is there any standard way to approach this problem? Has it come up
>>before in other applications?
>
>I don't know why you are even trying to normalize the paths like this. 
>Just hand the thing off to the OS. Usually, both the user and the OS 
>know what it is they are trying to do.

The path is also displayed at various places. The user might be
surprised to see surplus slashes.


Thorsten
-- 
Das Briefgeheimnis sowie das Post- und Fernmeldegeheimnis sind unverletzlich.
    - Grundgesetz, Artikel 10, Abs. 1 

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: UNC Pathname Handling within Applications

[Larry Hall]

At 07:30 PM 7/15/2004, you wrote:
>Hi,
>
>* Shankar Unni wrote (2004-07-16 00:52):
>>Thorsten Haude wrote:
>>>- Is there any standard way to approach this problem? Has it come up
>>>before in other applications?
>>
>>I don't know why you are even trying to normalize the paths like this. 
>>Just hand the thing off to the OS. Usually, both the user and the OS 
>>know what it is they are trying to do.
>
>The path is also displayed at various places. The user might be
>surprised to see surplus slashes.



Better that they be surprised and have things work than be surprised and
complain when things don't work.  Don't shield them from the reality of the
real world.  It's a learning experience and depriving them of that won't
help anyone.

That ends the philosophy lesson for today.  Please go back to your homes
now and meditate on this. ;-)


--
Larry Hall                              http://www.rfk.com
RFK Partners, Inc.                      (508) 893-9779 - RFK Office
838 Washington Street                   (508) 893-9889 - FAX
Holliston, MA 01746                     


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 1470 bytes --]

Hi,

please send me every mail only once. Thanks.

* Larry Hall wrote (2004-07-16 04:21):
>At 07:30 PM 7/15/2004, you wrote:
>>* Shankar Unni wrote (2004-07-16 00:52):
>>>Thorsten Haude wrote:
>>>>- Is there any standard way to approach this problem? Has it come up
>>>>before in other applications?
>>>
>>>I don't know why you are even trying to normalize the paths like this. 
>>>Just hand the thing off to the OS. Usually, both the user and the OS 
>>>know what it is they are trying to do.
>>
>>The path is also displayed at various places. The user might be
>>surprised to see surplus slashes.
>
>Better that they be surprised and have things work than be surprised and
>complain when things don't work.  Don't shield them from the reality of the
>real world.  It's a learning experience and depriving them of that won't
>help anyone.

I think some people might see the surprising path, *think* the
software is broken and stop using it. Another problem ist the toolkit.

I agree that the leading extra slash should be supported, but it might
not be as easy as it sounds.


>That ends the philosophy lesson for today.  Please go back to your homes
>now and meditate on this. ;-)

Thank you, O wise one!


Thorsten
-- 
As we enjoy great advantages from the inventions of others we should
be glad of an opportunity to serve others by any invention of ours,
and this we should do freely and generously.
    - Benjamin Franklin

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: UNC Pathname Handling within Applications

[Corinna Vinschen]

On Jul 16 14:06, Thorsten Haude wrote:
> I think some people might see the surprising path, *think* the
> software is broken and stop using it. Another problem ist the toolkit.
> 
> I agree that the leading extra slash should be supported, but it might
> not be as easy as it sounds.

Thorsten,

even if I repeat other postings, please note that double slashes are
absolutely fine according to the standards.  Leading double slashes may
have a special meaning on POSIX implementations and the *well* *known*
Windows specific meaning is an UNC path for accessing network (CIFS)
shares.  Honestly, there should be no surprise in that fact to users
of your software.  If you really want to mangle file names in your
application instead of just leaving them alone, then your application
should honor that fact, to be *portable*.

This isn't us being mean, as usual, that's just fact.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Co-Project Leader          mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 853 bytes --]

Hi,

* Corinna Vinschen wrote (2004-07-16 15:01):
>even if I repeat other postings, please note that double slashes are
>absolutely fine according to the standards.  Leading double slashes may
>have a special meaning on POSIX implementations and the *well* *known*
>Windows specific meaning is an UNC path for accessing network (CIFS)
>shares.  Honestly, there should be no surprise in that fact to users
>of your software.  If you really want to mangle file names in your
>application instead of just leaving them alone, then your application
>should honor that fact, to be *portable*.
>
>This isn't us being mean, as usual, that's just fact.

I already agreed that it should be supported, so what's missing?


Thorsten
-- 
The man who does not read good books has no
advantage over the man who cannot read them.
    - Mark Twain

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: UNC Pathname Handling within Applications

[Corinna Vinschen]

On Jul 16 15:10, Thorsten Haude wrote:
> Hi,
> 
> * Corinna Vinschen wrote (2004-07-16 15:01):
> >even if I repeat other postings, please note that double slashes are
> >absolutely fine according to the standards.  Leading double slashes may
> >have a special meaning on POSIX implementations and the *well* *known*
> >Windows specific meaning is an UNC path for accessing network (CIFS)
> >shares.  Honestly, there should be no surprise in that fact to users
> >of your software.  If you really want to mangle file names in your
> >application instead of just leaving them alone, then your application
> >should honor that fact, to be *portable*.
> >
> >This isn't us being mean, as usual, that's just fact.
> 
> I already agreed that it should be supported, so what's missing?

Erm... nothing, apparently.  I must have misinterpreted your posting.

Sorry,
Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Co-Project Leader          mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Larry Hall]

Thorsten,

The best way to insure that you get only one copy of any response from the 
list is to set your reply-to to point to the list.  Depending on the software
you use, you may find that you can easily filter out duplicates too.  I don't
second-guess the implied intent of posters to whom I respond by editing the
recipient list but my email client automatically honors when "reply-to" is 
set.  Most do.  I'd recommend setting it when you correspond with this list or any that you subscribe to.

Larry

PS - I've edited the recipient list in this case for you, given your stated
preference.


At 08:06 AM 7/16/2004, you wrote:
>Hi,
>
>please send me every mail only once. Thanks.
>
>* Larry Hall wrote (2004-07-16 04:21):
>>At 07:30 PM 7/15/2004, you wrote:
>>>* Shankar Unni wrote (2004-07-16 00:52):
>>>>Thorsten Haude wrote:
>>>>>- Is there any standard way to approach this problem? Has it come up
>>>>>before in other applications?
>>>>
>>>>I don't know why you are even trying to normalize the paths like this. 
>>>>Just hand the thing off to the OS. Usually, both the user and the OS 
>>>>know what it is they are trying to do.
>>>
>>>The path is also displayed at various places. The user might be
>>>surprised to see surplus slashes.
>>
>>Better that they be surprised and have things work than be surprised and
>>complain when things don't work.  Don't shield them from the reality of the
>>real world.  It's a learning experience and depriving them of that won't
>>help anyone.
>
>I think some people might see the surprising path, *think* the
>software is broken and stop using it. Another problem ist the toolkit.
>
>I agree that the leading extra slash should be supported, but it might
>not be as easy as it sounds.
>
>
>>That ends the philosophy lesson for today.  Please go back to your homes
>>now and meditate on this. ;-)
>
>Thank you, O wise one!
>
>
>Thorsten
>-- 
>As we enjoy great advantages from the inventions of others we should
>be glad of an opportunity to serve others by any invention of ours,
>and this we should do freely and generously.
>    - Benjamin Franklin


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 4974 bytes --]

Hi,

I sent the sollowing message to Larry via private mail:

- - - Schnipp - - -
Hi Larry,

please don't use tofu mails. http://www.vranx.de/mail/tofu.html

* Larry Hall wrote (2004-07-16 17:41):
>The best way to insure that you get only one copy of any response from the
>list is to set your reply-to to point to the list.

I think the Reply-To header is quite unfit for that.

For once, it it mungled often enough that I had to check with every
list whether I could use it or not.

Second, the Reply-To header has a different purpose. If would eg. post
to the list from a mobile phone, I would want a way to ensure that I
don't get the entire list on this mobile, accumulating huge costs.
Reply-To ist that way.

Third, Reply-To is a much stronger way to redirect replies. If this
mail would have been sent to the list, I might set the Reply-To header
to indicate that further discussion should be done off-list.


Instead I use the Mail-Followup-To header as described here:
http://cr.yp.to/proto/replyto.html


>Depending on the software you use, you may find that you can easily
>filter out duplicates too.

That wouldn't be entirely satisfactory either for two reasons:
- The list mail would usually arrive second and be destroyed. That
leaves not-list messages in my list folder.
- I might be lead to think that it *is* a personal reply (like this
one) and answer personally where a list answer might be more
approppriate.


>I don't second-guess the implied intent of posters to whom I respond
>by editing the recipient list but my email client automatically
>honors when "reply-to" is set.  Most do.  I'd recommend setting it
>when you correspond with this list or any that you subscribe to.

Actually, I wonder what possible reasons could exist to send every
mail twice, and even by default. I haven't checked, but I guess you
cannot subscribe this list or any other write-only, so in what
possible situation would the second mail enhance communication?


So my recommendation to you, O wise one, is to just skip the second
mail, reply to the list *or* in personal way as preferred for a
particular mail and be done with it.


Thorsten
--
As we enjoy great advantages from the inventions of others we should
be glad of an opportunity to serve others by any invention of ours,
and this we should do freely and generously.
    - Benjamin Franklin
- - - Schnapp - - -


In an ironic twist I got this reply from the list manager:
- - - Schnipp - - -
Hi! This is the ezmlm program. I'm managing the
cygwin@cygwin.com mailing list.

This is a generic help message. The message I received wasn't sent to
any of my command addresses.


--- Administrative commands for the cygwin list ---

I can handle administrative requests automatically. Please
DO NOT SEND THEM TO THE LIST ADDRESS! If you do, I will not
see them and other subscribers will be annoyed. Instead, send
your message to the correct command address:


To subscribe to the list, send a message to:
   <cygwin-subscribe@cygwin.com>

To remove your address from the list, send a message to:
   <cygwin-unsubscribe@cygwin.com>

Send mail to the following for info and FAQ for this list:
   <cygwin-info@cygwin.com>
   <cygwin-faq@cygwin.com>

Similar addresses exist for the digest list:
   <cygwin-digest-subscribe@cygwin.com>
   <cygwin-digest-unsubscribe@cygwin.com>

To get messages 123 through 145 (a maximum of 100 per request), mail:
   <cygwin-get.123_145@cygwin.com>

To get an index with subject and author for messages 123-456 , mail:
   <cygwin-index.123_456@cygwin.com>

They are always returned as sets of 100, max 2000 per request,
so you'll actually get 100-499.

To receive all messages with the same subject as message 12345,
send an empty message to:
   <cygwin-thread.12345@cygwin.com>

The messages do not really need to be empty, but I will ignore
their content. Only the ADDRESS you send to is important.

You can start a subscription for an alternate address,
for example "john@host.domain", just add a hyphen and your
address (with '=' instead of '@') after the command word:
<cygwin-subscribe-john=host.domain@cygwin.com>

To stop subscription for this address, mail:
<cygwin-unsubscribe-john=host.domain@cygwin.com>

In both cases, I'll send a confirmation message to that address. When
you receive it, simply reply to it to complete your subscription.

If despite following these instructions, you do not get the
desired results, please contact my owner at
cygwin-owner@cygwin.com. Please be patient, my owner is a
lot slower than I am ;-)

--- Enclosed is a copy of the request I received.

(...)
- - - Schnapp - - -


Thorsten
-- 
As we enjoy great advantages from the inventions of others we should
be glad of an opportunity to serve others by any invention of ours,
and this we should do freely and generously.
    - Benjamin Franklin

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: UNC Pathname Handling within Applications

[Shankar Unni]

Thorsten Haude wrote:

> The path is also displayed at various places. The user might be
> surprised to see surplus slashes.

[Starting to drift OT, sorry..]

Who's generating the surplus slashes? Is this because of trailing "/" in 
paths, and you're internally compositing names somehow?

I thought the only names you would *show* to the user would be the names 
of editor buffers, etc., something which the user has already specified 
explicitly anyway. It's not like you're saying "nedit myfile", and 
you're searching all over the filesystem to locate myfile..

If the name is something the user is already specifying as a relative or 
absolute path, then there's no need to further normalize them.

If you *do* ever show to the user a path that nedit has composed itself 
(say, some property display, like "where is the foobar package that is 
automatically loaded?"), then the right thing is to only normalize *at 
the point where you do the composition*.

I.e. if you're doing "prefix + basename", then the only normalization 
should be between prefix and basename (i.e. strip trailing slashes (any 
kind) from prefix, and append a "known slash" and basename).

You shouldn't be doing any further normalization (e.g. /x/../y -> /y or 
any such), because if it *is* a synthesized name, then there's great 
value in knowing *how* the name was synthesized. Normalizing it loses 
this information.


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Larry Hall]

At 02:13 PM 7/16/2004, you wrote:
>Hi,
>
>I sent the sollowing message to Larry via private mail:


Sorry.  Everything I post to the list I want to be on the list.  So only
the list version got through.  I guess you figured that out by now.  If
you would like to continue this or some other discussion off-list (and 
continuing it much beyond this is likely getting OT so that would make 
sense), reply to the list and let me know that's the direction you want
to go.  I'll then reply to you directly and we can finish up off list.


>- - - Schnipp - - -
>Hi Larry,
>
>please don't use tofu mails. http://www.vranx.de/mail/tofu.html


Ah-huh.


>* Larry Hall wrote (2004-07-16 17:41):
>>The best way to insure that you get only one copy of any response from the
>>list is to set your reply-to to point to the list.
>
>I think the Reply-To header is quite unfit for that.
>
>For once, it it mungled often enough that I had to check with every
>list whether I could use it or not.


OK, it is not munged by this list.


>Second, the Reply-To header has a different purpose. If would eg. post
>to the list from a mobile phone, I would want a way to ensure that I
>don't get the entire list on this mobile, accumulating huge costs.
>Reply-To ist that way.


Sorry, you've left this a bit vague.  If you're implying that you want 
to get a reply on your mobile and that that address is different than 
the one you subscribed to the list with, I agree.  But that doesn't seem
to be particularly pertinent other than its a way that Reply-To can be 
used. 


>Third, Reply-To is a much stronger way to redirect replies. If this
>mail would have been sent to the list, I might set the Reply-To header
>to indicate that further discussion should be done off-list.


That's fine, if that's what you want.



>Instead I use the Mail-Followup-To header as described here:
>http://cr.yp.to/proto/replyto.html
>


So you do.  I guess my email client doesn't honor that. :-(  Sorry about 
that.  Well, looks like we're at a bit of an impasse here then until one 
of us makes a change, unless I *always* remember to edit my replies to you
(not likely unfortunately).


>>Depending on the software you use, you may find that you can easily
>>filter out duplicates too.
>
>That wouldn't be entirely satisfactory either for two reasons:
>- The list mail would usually arrive second and be destroyed. That
>leaves not-list messages in my list folder.
>- I might be lead to think that it *is* a personal reply (like this
>one) and answer personally where a list answer might be more
>approppriate.


People do this with their email clients directly or through the help
of procmail.  Obviously, there are ways to address your concerns, though
I'm not suggesting that this is what you need to do.  Just that it is an
option for you as a way to deal with those who can't or won't honor your 
settings/requests.


>>I don't second-guess the implied intent of posters to whom I respond
>>by editing the recipient list but my email client automatically
>>honors when "reply-to" is set.  Most do.  I'd recommend setting it
>>when you correspond with this list or any that you subscribe to.
>
>Actually, I wonder what possible reasons could exist to send every
>mail twice, and even by default. I haven't checked, but I guess you
>cannot subscribe this list or any other write-only, so in what
>possible situation would the second mail enhance communication?


You guessed wrong.  That's exactly why I use reply-to-all.  You don't have 
to be subscribed to the list to post to it.


>So my recommendation to you, O wise one, is to just skip the second
>mail, reply to the list *or* in personal way as preferred for a
>particular mail and be done with it.


OK, fair recommendation, considering I had ones for you (well I would have
called my suggestions but what's semantics between friends ;-) ).  Just 
like you, I have reasons for replying as I do and they have served me well 
on this list low these many years.  I don't plan on changing them.  I 
have added "followup-to" to my list of important features that my next
email client has, once I make a switch though.  Thanks for the pointer to
that.



--
Larry Hall                              http://www.rfk.com
RFK Partners, Inc.                      (508) 893-9779 - RFK Office
838 Washington Street                   (508) 893-9889 - FAX
Holliston, MA 01746                     


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 3547 bytes --]

Hi,

* Larry Hall wrote (2004-07-16 21:19):
>At 02:13 PM 7/16/2004, you wrote:
>>I sent the sollowing message to Larry via private mail:
>
>Sorry.  Everything I post to the list I want to be on the list.

I expected that to be a misconfiguration of some kind, now you say you
do that on purpose. It's your choice to prefer discussions on the
list, but I it's highly annoying to send people off to some
non-existing address that you not even own.


>If you would like to continue this or some other discussion off-list,
>reply to the list and let me know that's the direction you want to go.

What other loops have I got to jump through trying to communicate with
you? If you force this discussion on the list by avoiding to open
other means of communication, that's what you get.


>(and continuing it much beyond this is likely getting OT so that
>would make sense)>

Which is exactly the reason I wanted to take this off-list. Your
Reply-To antics is what prevented it.


>>I think the Reply-To header is quite unfit for that.
>>
>>For once, it it mungled often enough that I had to check with every
>>list whether I could use it or not.
>
>OK, it is not munged by this list.

Yeah, but I don't want to check every time somebody insists to send me
every mail twice.


>>Instead I use the Mail-Followup-To header as described here:
>>http://cr.yp.to/proto/replyto.html
>
>So you do.  I guess my email client doesn't honor that.

Your email client *sets* Mail-Followup-To, so it must be pretty crappy
to ignore it on inbound mail.


>>>Depending on the software you use, you may find that you can easily
>>>filter out duplicates too.
>>
>>That wouldn't be entirely satisfactory either for two reasons:
>>- The list mail would usually arrive second and be destroyed. That
>>leaves not-list messages in my list folder.
>>- I might be lead to think that it *is* a personal reply (like this
>>one) and answer personally where a list answer might be more
>>approppriate.
>
>People do this with their email clients directly or through the help
>of procmail.

People do *what* with Procmail? I've never seen an MDA recipe that
deletes the *first* mail of an identical set. It's rather hard to do,
too, since the MDA would have to remember where that first mail went.


>>>I don't second-guess the implied intent of posters to whom I respond
>>>by editing the recipient list but my email client automatically
>>>honors when "reply-to" is set.  Most do.  I'd recommend setting it
>>>when you correspond with this list or any that you subscribe to.
>>
>>Actually, I wonder what possible reasons could exist to send every
>>mail twice, and even by default. I haven't checked, but I guess you
>>cannot subscribe this list or any other write-only, so in what
>>possible situation would the second mail enhance communication?
>
>You guessed wrong.

Please tell me how I activate the write-only subscription.


>That's exactly why I use reply-to-all.  You don't have to be
>subscribed to the list to post to it.

Ok, so find a way to identify these so that you don't have to send
each mail out twice by default. If these people do not care enough to
subscribe, why should you care to accomodate them?


>(well I would have called my suggestions but what's semantics between
>friends ;-) )

Yeah, let's discuss our knowledge of the english language. Is there
any lower point you could have made?


Thorsten
-- 
Guns don't protect freedom, people protect freedom.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 1632 bytes --]

Hi,

* Shankar Unni wrote (2004-07-16 20:36):
>Thorsten Haude wrote:
>
>>The path is also displayed at various places. The user might be
>>surprised to see surplus slashes.
>
>[Starting to drift OT, sorry..]
>
>Who's generating the surplus slashes? Is this because of trailing "/" in 
>paths, and you're internally compositing names somehow?

Yes, we never cared for that because we weren't aware of this leading
slash thing. (I wasn't at least, and never heard about it before.) So
we have to check all internals for problems.

There will probably also be problems with the toolkit.


>I thought the only names you would *show* to the user would be the names 
>of editor buffers, etc., something which the user has already specified 
>explicitly anyway. It's not like you're saying "nedit myfile", and 
>you're searching all over the filesystem to locate myfile..

The filename is used in several parts of the application. I expect
some of them to make false assumptions about leading slashes.


>I.e. if you're doing "prefix + basename", then the only normalization 
>should be between prefix and basename (i.e. strip trailing slashes (any 
>kind) from prefix, and append a "known slash" and basename).
>
>You shouldn't be doing any further normalization (e.g. /x/../y -> /y or 
>any such), because if it *is* a synthesized name, then there's great 
>value in knowing *how* the name was synthesized. Normalizing it loses 
>this information.

These are the things we have to think about, yes.


Thorsten
-- 
Getting a thrill out of some stupid quote is a sign of idiocy.
    - turmeric

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: UNC Pathname Handling within Applications

[Larry Hall]

At 03:55 PM 7/16/2004, you wrote:

<snip>


>>>Actually, I wonder what possible reasons could exist to send every
>>>mail twice, and even by default. I haven't checked, but I guess you
>>>cannot subscribe this list or any other write-only, so in what
>>>possible situation would the second mail enhance communication?
>>
>>You guessed wrong.
>
>Please tell me how I activate the write-only subscription.
>

<snip>

You don't need to be subscribed to the Cygwin list to send email to it.
See <http://cygwin.com/lists.html>.

Unless you have some further pertinent point to make on this inquiry, I 
think we're done with this thread.   Been nice talking with you. 


Larry Hall                              http://www.rfk.com
RFK Partners, Inc.                      (508) 893-9779 - RFK Office
838 Washington Street                   (508) 893-9889 - FAX
Holliston, MA 01746                     


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Igor Pechtchanski]

On Fri, 16 Jul 2004, Thorsten Haude wrote:

> Please tell me how I activate the write-only subscription.

See <http://cygwin.com/ml/lists.html#rbl-sucks>.
	Igor
-- 
				http://cs.nyu.edu/~pechtcha/
      |\      _,,,---,,_		pechtcha@cs.nyu.edu
ZZZzz /,`.-'`'    -.  ;-;;,_		igor@watson.ibm.com
     |,4-  ) )-,_. ,\ (  `'-'		Igor Pechtchanski, Ph.D.
    '---''(_/--'  `-'\_) fL	a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

"I have since come to realize that being between your mentor and his route
to the bathroom is a major career booster."  -- Patrick Naughton

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 1065 bytes --]

Hi,

* Larry Hall wrote (2004-07-16 22:29):
>At 03:55 PM 7/16/2004, you wrote:
>>>>Actually, I wonder what possible reasons could exist to send every
>>>>mail twice, and even by default. I haven't checked, but I guess you
>>>>cannot subscribe this list or any other write-only, so in what
>>>>possible situation would the second mail enhance communication?
>>>
>>>You guessed wrong.
>>
>>Please tell me how I activate the write-only subscription.
>
><snip>
>
>You don't need to be subscribed to the Cygwin list to send email to it.
>See <http://cygwin.com/lists.html>.

I referred to that fact in the part you generously snipped.


>Unless you have some further pertinent point to make on this inquiry, I 
>think we're done with this thread.

You don't figure to be a guy caring about points.


Thorsten
-- 
There must be a parallel universe where Starbucks sells cocaine in dilute,
controlled, taxed doses, but where (crack) caffiene is sniffed and smoked
in huge binges, and smuggled in by evil Columbian cartels.
    - Rogerborg

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: UNC Pathname Handling within Applications

[Christopher Faylor]

On Fri, Jul 16, 2004 at 09:55:49PM +0200, Thorsten Haude wrote:
>* Larry Hall wrote (2004-07-16 21:19):
>Your email client *sets* Mail-Followup-To, so it must be pretty crappy
>to ignore it on inbound mail.

This field is added by the ezmlm software that is used to manage our
mailing lists.

>I've never seen an MDA recipe that deletes the *first* mail of an
>identical set.  It's rather hard to do, too, since the MDA would have
>to remember where that first mail went.

More things in heaven and earth, Horatio?

 From the procmailex manpage:

   If  you are subscribed to several mailinglists and people cross-post to
   some of them, you usually receive several  duplicate  mails  (one  from
   every  list).   The following simple recipe eliminates duplicate mails.
   It tells formail to keep an 8KB cache file in which it will  store  the
   Message-IDs  of  the most recent mails you received.  Since Message-IDs
   are guaranteed to be unique for every new mail, they are ideally suited
   to  weed  out  duplicate mails.  Simply put the following recipe at the
   top of your rcfile, and no duplicate mail will get past it.

	  :0 Wh: msgid.lock
	  | formail -D 8192 msgid.cache


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 1834 bytes --]

Hi,

* Christopher Faylor wrote (2004-07-16 23:19):
>On Fri, Jul 16, 2004 at 09:55:49PM +0200, Thorsten Haude wrote:
>>* Larry Hall wrote (2004-07-16 21:19):
>>Your email client *sets* Mail-Followup-To, so it must be pretty crappy
>>to ignore it on inbound mail.
>
>This field is added by the ezmlm software that is used to manage our
>mailing lists.

Ah, so he is just ignoring good advice given by the list software.
Thanks for clarifying that.

Sorry Larry, your mailer is not crappy.


>>I've never seen an MDA recipe that deletes the *first* mail of an
>>identical set.  It's rather hard to do, too, since the MDA would have
>>to remember where that first mail went.
>
>More things in heaven and earth, Horatio?
>
> From the procmailex manpage:
>
>   If  you are subscribed to several mailinglists and people cross-post to
>   some of them, you usually receive several  duplicate  mails  (one  from
>   every  list).   The following simple recipe eliminates duplicate mails.
>   It tells formail to keep an 8KB cache file in which it will  store  the
>   Message-IDs  of  the most recent mails you received.  Since Message-IDs
>   are guaranteed to be unique for every new mail, they are ideally suited
>   to  weed  out  duplicate mails.  Simply put the following recipe at the
>   top of your rcfile, and no duplicate mail will get past it.
>
>	  :0 Wh: msgid.lock
>	  | formail -D 8192 msgid.cache

I don't intend to install procmail to test this, but are you sure that
this recipe deletes the *first* of the two mails? From the manpage, it
looks like the second mail would be removed.


Thorsten
-- 
The smart way to keep people passive and obedient is to
strictly limit the spectrum of acceptable opinion, but
allow very lively debate within that spectrum.
    - Noam Chomsky

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: UNC Pathname Handling within Applications

[Christopher Faylor]

On Fri, Jul 16, 2004 at 11:32:56PM +0200, Thorsten Haude wrote:
>Hi,
>
>* Christopher Faylor wrote (2004-07-16 23:19):
>>On Fri, Jul 16, 2004 at 09:55:49PM +0200, Thorsten Haude wrote:
>>>* Larry Hall wrote (2004-07-16 21:19):
>>>Your email client *sets* Mail-Followup-To, so it must be pretty crappy
>>>to ignore it on inbound mail.
>>
>>This field is added by the ezmlm software that is used to manage our
>>mailing lists.
>
>Ah, so he is just ignoring good advice given by the list software.
>Thanks for clarifying that.
>
>Sorry Larry, your mailer is not crappy.

You must have forgotten.  He said his mailer does not honor this header
not that he, personally, is ignoring it.

Larry said that he would try to avoid sending you an extra copy in the
future and he has explained the reason for his email habits.  You don't
have to agree with his reasons but continuing to carp at him is
pointless.

>>>I've never seen an MDA recipe that deletes the *first* mail of an
>>>identical set.  It's rather hard to do, too, since the MDA would have
>>>to remember where that first mail went.
>>
>>More things in heaven and earth, Horatio?
>>
>> From the procmailex manpage:
>>
>>   If  you are subscribed to several mailinglists and people cross-post to
>>   some of them, you usually receive several  duplicate  mails  (one  from
>>   every  list).   The following simple recipe eliminates duplicate mails.
>>   It tells formail to keep an 8KB cache file in which it will  store  the
>>   Message-IDs  of  the most recent mails you received.  Since Message-IDs
>>   are guaranteed to be unique for every new mail, they are ideally suited
>>   to  weed  out  duplicate mails.  Simply put the following recipe at the
>>   top of your rcfile, and no duplicate mail will get past it.
>>
>>	  :0 Wh: msgid.lock
>>	  | formail -D 8192 msgid.cache
>
>I don't intend to install procmail to test this, but are you sure that
>this recipe deletes the *first* of the two mails? From the manpage, it
>looks like the second mail would be removed.

You're right.  It doesn't delete the first.  I didn't think you really
meant to delete the first message since that seemed to imply some kind
of precognitive ability -- how could a program ever know when there are
potentially N copies of a message being sent?  I guess you could delay
every email message delivery for an hour or so to make sure that you've
gotten them all but that isn't foolproof.

I don't understand why it matters that the first message would be
deleted since you can't predict if the message would be directly to you
or to the cygwin list.  However, please don't enlighten me since I can
live with not knowing.  More importantly, your personal email
preferences and my ignorance of same are really off-topic here.

Perhaps most importantly, shouldn't you be hard at work on Nedit now
that you know that there is no way for cygwin to reach inside the
program and stop the Nedit path handling software from deleting needed
double slashes?

Hmm.  Maybe there's a market there.  If we could develop that
precognitive mail software, how hard would developing a DLL that
telepathically changed the program flow of path handling software be?

*pause*

Nuts.  psychicsoftware.com is already taken.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: UNC Pathname Handling within Applications

[Thorsten Haude]

[-- Attachment #1: Type: text/plain, Size: 761 bytes --]

Hi,

* Christopher Faylor wrote (2004-07-17 03:50):
>I don't understand why it matters that the first message would be
>deleted since you can't predict if the message would be directly to you
>or to the cygwin list.  However, please don't enlighten me since I can
>live with not knowing.  More importantly, your personal email
>preferences and my ignorance of same are really off-topic here.

I explained that point before, but since you also seem to deprecate
personal communication (and even make this point clear), I won't
follow up on that.


Thorsten
-- 
He who receives an idea from me, receives instruction himself without lessening
mine; as he who lights his taper at mine, receives light without darkening me.
    - Thomas Jefferson

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

RE: UNC Pathname Handling within Applications

[Hannu E K Nevalainen]

cygwin-owner@cygwin.com wrote:
> Hi,
>
> * Christopher Faylor wrote (2004-07-17 03:50):
>> I don't understand why it matters that the first message would be
>> deleted since you can't predict if the message would be directly to
>> you or to the cygwin list.  However, please don't enlighten me since
>> I can live with not knowing.  More importantly, your personal email
>> preferences and my ignorance of same are really off-topic here.
>
> I explained that point before, but since you also seem to deprecate
> personal communication (and even make this point clear), I won't
> follow up on that.
>
>
> Thorsten

Wow, I'm impressed over the level of squabbling -

In general:
I would be very much IMPRESSED if ANY/ALL posters stopped attempting
anything else than having cygwin@<ditto>.com in the To: field and clearing
the other related fields.

"Peronal replies" is mentioned in old netiquette docs, look up the docs and
have a read - then please follow the recommendations.


/Hannu E K Nevalainen, B.Sc. EE - 59+16.37'N, 17+12.60'E
--76-->

** on a mailing list; please keep replies on that particular list **

-- printf("LocalTime: UTC+%02d\n",(DST)? 2:1); --
--END OF MESSAGE--


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/