dbusbind.c patch, fixes possible segfault

dbusbind.c patch, fixes possible segfault

[David Hansen]

[-- Attachment #1: Type: text/plain, Size: 288 bytes --]


This one is similar to the previous dbusbind.c patch.  No need to copy
the string and if dbus_bus_get_unique_name () returns NULL strcpy will
segfault.

David

2008-04-13  David Hansen  <david.hansen@gmx.net>

	* dbusbind.c (dbus-get-unique-name): Remove extra copying of name
	string.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: Type: text/x-patch, Size: 949 bytes --]

*** dbusbind.c.~1.24.~	2008-04-05 23:46:55.000000000 +0200
--- dbusbind.c	2008-04-13 15:19:27.000000000 +0200
***************
*** 701,707 ****
       Lisp_Object bus;
  {
    DBusConnection *connection;
!   char name[DBUS_MAXIMUM_NAME_LENGTH];
  
    /* Check parameters.  */
    CHECK_SYMBOL (bus);
--- 701,707 ----
       Lisp_Object bus;
  {
    DBusConnection *connection;
!   const char *name;
  
    /* Check parameters.  */
    CHECK_SYMBOL (bus);
***************
*** 710,716 ****
    connection = xd_initialize (bus);
  
    /* Request the name.  */
!   strcpy (name, dbus_bus_get_unique_name (connection));
    if (name == NULL)
      xsignal1 (Qdbus_error, build_string ("No unique name available"));
  
--- 710,716 ----
    connection = xd_initialize (bus);
  
    /* Request the name.  */
!   name = dbus_bus_get_unique_name (connection);
    if (name == NULL)
      xsignal1 (Qdbus_error, build_string ("No unique name available"));
  

Re: dbusbind.c patch, fixes possible segfault

[Michael Albinus]

David Hansen <david.hansen@gmx.net> writes:

> This one is similar to the previous dbusbind.c patch.  No need to copy
> the string and if dbus_bus_get_unique_name () returns NULL strcpy will
> segfault.

Looks OK, I've committed it. Thanks a lot!

Btw, I'm a little bit curious how you've managed it NOT to have a unique
name?

> David

Best regards, Michael.

Re: dbusbind.c patch, fixes possible segfault

[David Hansen]

On Sun, 13 Apr 2008 18:43:13 +0200 Michael Albinus wrote:

> David Hansen <david.hansen@gmx.net> writes:
>
>> This one is similar to the previous dbusbind.c patch.  No need to copy
>> the string and if dbus_bus_get_unique_name () returns NULL strcpy will
>> segfault.
>
> Looks OK, I've committed it. Thanks a lot!
>
> Btw, I'm a little bit curious how you've managed it NOT to have a unique
> name?

I haven't.  I try to make myself familiar with the low level dbus api
and the emacs code is a nice reading.

David

Re: dbusbind.c patch, fixes possible segfault

[David Hansen]

BTW, no need to flag my changes as `tiny' (this one obviously was, the
other one might be questionable), papers are signed.  Wonder what "tiny"
is anyway.

David

Re: dbusbind.c patch, fixes possible segfault

[Eli Zaretskii]

> From: David Hansen <david.hansen@gmx.net>
> Date: Mon, 14 Apr 2008 06:04:47 +0200
> 
> Wonder what "tiny" is anyway.

It indicates that the contributor didn't have papers signed at the
time of the contribution, and that her contribution was found to be
small enough to not require papers.  This is so we could tell, years
later, how come someone's code is in Emacs without paperwork.

If you signed legal papers, your contributions after the papers were
signed should NOT be marked as "tiny".

Re: dbusbind.c patch, fixes possible segfault

[Michael Albinus]

David Hansen <david.hansen@gmx.net> writes:

> BTW, no need to flag my changes as `tiny' (this one obviously was, the
> other one might be questionable), papers are signed.  Wonder what "tiny"
> is anyway.

You're right, you're listed in copyright.list on fencepost. T thought I
did check it when you have contributed your first patch; maybe I've
overlooked your entry.

I've removed the "tiny" flag from your Changelog entries. Eli did
explain it already to you what they stand for; you can read also about
at <http://www.gnu.org/prep/maintain/maintain.html#Legally-Significant>.

Btw, when you have signed the papers, you should have write access to
the Emacs repository, shouldn't you?

> David

Best regards, Michael.

Re: dbusbind.c patch, fixes possible segfault

[David Hansen]

On Mon, 14 Apr 2008 20:47:52 +0200 Michael Albinus wrote:

> Btw, when you have signed the papers, you should have write access to
> the Emacs repository, shouldn't you?

No idea how this works, but I think that would be a little bit to much
people with write access.

David

Re: dbusbind.c patch, fixes possible segfault

[Stefan Monnier]

> Btw, when you have signed the papers, you should have write access to
> the Emacs repository, shouldn't you?

Not really: signing paperwork only means that your code is
legally acceptable.  But we may decide not to trust you w.r.t quality.
If you do have a fair bit of code already installed in Emacs
(e.g. a package), then as a general rule I'd most likely give out
write-access.


        Stefan

Re: dbusbind.c patch, fixes possible segfault

[Glenn Morris]

David Hansen wrote:

> On Mon, 14 Apr 2008 20:47:52 +0200 Michael Albinus wrote:
>
>> Btw, when you have signed the papers, you should have write access to
>> the Emacs repository, shouldn't you?
>
> No idea how this works, but I think that would be a little bit to much
> people with write access.

Some people with commit access don't have Emacs copyright assignments,
and some others seems to have contributed no, or almost no, code.
It's all a bit odd.

Re: dbusbind.c patch, fixes possible segfault

[Stefan Monnier]

>>> Btw, when you have signed the papers, you should have write access to
>>> the Emacs repository, shouldn't you?
>> 
>> No idea how this works, but I think that would be a little bit to much
>> people with write access.

> Some people with commit access don't have Emacs copyright assignments,

Sounds like errors.  Unless maybe they only have assignments for part of
the code.  In any case we should try and make sure that such situations
do not happen.  Can you figure out who those people are?


        Stefan

Re: dbusbind.c patch, fixes possible segfault

[Glenn Morris]

Stefan Monnier wrote:

> Sounds like errors.  Unless maybe they only have assignments for part of
> the code.  In any case we should try and make sure that such situations
> do not happen.  Can you figure out who those people are?

Christopher Stacy
Dhruva Krishnamurthy
Jaeyoun Chung 
Martin Baulig
Rudy Gevaert