113684 - Disable prepending of "-- \n" before .sig if a "-- \n" is already in the sig

Status: VERIFIED DUPLICATE of bug 229044

(MailNews Core :: Composition, enhancement)

Description

[Alexander Skwar]

It would be very good, if there were a way to disable prepending the "-- \n"
signature delimiter before a .sig.  My .sig alread contains a "-- \n", so the
one added by Mozilla is just not needed.

Additionally, it would be very good if Mozilla would detect if there's a a line
matching "^-- $" in the signature file.  If there's such a line, Mozilla should
not add the "-- \n" by itself.

Comment 1

[Boris Zbarsky [:bzbarsky]]

See bug 30208.  Is the "-- \n" not the first thing in your .sig file?

Comment 2

[Alexander Skwar]

No, "-- \n" is not the first line in my .sig.  The .sig I use is:

------ START -------
Alexander Skwar
-- 
How to quote: http://learn.to/quote (german) http://quote.6x.to (en)
   iso-top.de - Die günstige Art an Linux Distributionen zu kommen
        -=» Suche Wohnung in Gelsenkirchen und Umgebung! «=- 
                       Uptime: 1 day 3 hours 50 minutes
------ ENDE ---------

(Without the --START-- and --END--)

Comment 3

[Boris Zbarsky [:bzbarsky]]

setting status to new, but see discussion in bug 30208 about why we don't do
what you ask for...

Comment 4

[Alexander Skwar]

Well, do you want changes here in this bug, or in the other bug?

Comment 5

[Alexander Skwar]

Hmm, I'm not much of a C++ hacker, but to detect if the sigData contains "--
\n", wouldn't it be sufficient to change lines 3011, 3012
(http://lxr.mozilla.org/seamonkey/source/mailnews/compose/src/nsMsgCompose.cpp#3011)
to:

if (!(sigData.Find("-- \n", PR_TRUE, 0, 1) ||
      sigData.Find("-- \r", PR_TRUE, 0, 1)))

This should search sigData for any occurance of "-- \n" or "-- \r", and if it
finds one, it doesn't add the sigdashes.

Comment 6

[Daniel Bratell]

Probably, but that would scan the whole signature file twice in the normal case.
We at least want to keep it to one scan and probably limit the amount of data
scanned before giving up to.

Comment 7

[Alexander Skwar]

Hmm, yes, it scans twice, however, I don't see this as a major performance hit.
 But, you are right, it for sure will take longer than the current approach.

Hm, as I said, I'm not much of a C++ hacker, and I much less know the Mozilla
sources well enough.  However, here's another idea.

a) Split sigData into lines (delimited by some EOL, or "\n", or "\r")
b) Start scanning at line 0
c) Check if line length is equal to 3
d) If yes, check if line is equal to "-- "
d2) "Remember" that no sigdash should be inserted
d3) Bail out of the sigcheck
e) If not, get the next line
e2) Goto step c)

The difference with this approach is, that sigData needs to be split, and that
it needs to loop thru all the lines of sigData.

Comment 8

[Boris Zbarsky [:bzbarsky]]

Well... the simple thing to do is to seach sigData for "-- " using
FindInReadable and then check that the second iter points to a \n or \r.

The hard part is making this play nice with HTML signatures which may contain
embedded comments.

Comment 9

[Daniel Bratell]

Also, you must check so that the "-- " is alone on a line. It mustn't be
triggered by the text: 

"And then we waited some more -- "

Comment 10

[Ben Bucksch (:BenB)]

> The hard part is making this play nice with HTML signatures

We should not search for "-- " at all in the HTML sig. "-- " is for plaintext.

Comment 11

[Alexander Skwar]

Thanks, that's just what I though, and I also thought that this was said in the other 
bug.

Besides, does the current way work for HTML?

Because of the "problems" with 
lines in a .sig like "foo -- ", I'd still think that the right way to go was to split the lines.  Then, after the 
lines are split, loop thru all the lines and check the if the line is 3 chars long and is equal to "-- 
".

Dunno, but isn't it possible to get the lines of a nsAutoString (sigData is of this type)?

Comment 12

[Boris Zbarsky [:bzbarsky]]

*** Bug 115039 has been marked as a duplicate of this bug. ***

Comment 13

[Alexander Skwar]

Hm, I was thinking about this.  Why does there need to be any "inteligence" at
all?  I'd propose to throw away the autodetection and add a setting "[x] Prepend
signature delimiter before the signature".  This setting should be enabled by
default.

Comment 14

[Ben Bucksch (:BenB)]

- the less prefs UI the better
- We should (or even have to?) force a sig delim., esp. for newbies not knowing
the consequences of disabling it.

Comment 15

[Alexander Skwar]

Well, I do not think that the delim should be forced at all.  Yes, Mozilla
*should* use the delim, if the user wishes so.  Quite a lot of mail/news clients
allow to disable the deli.

To cater for the newbies, the setting really should default to using the delim.
 And this setting might/should be hidden in some "advanced" prefs.

Comment 16

[Marcin Orlowski]

*** Bug 142289 has been marked as a duplicate of this bug. ***

Comment 17

[Marcin Orlowski]

I'd keep the target far in future, as there's really a big pain in the **** for
someone who just migrated to Mozilla Mail and is now unable to keep old habits
and sigs.

Comment 18

[Christian :Biesinger (don't email me, ping me on IRC)]

well, ch.ey fixed this in new bug 229044.

*** This bug has been marked as a duplicate of 229044 ***

Comment 19

[Ben Bucksch (:BenB)]

vrfy.