Hi Tony,
You are right.
Thanks. Dmitry.
> -Original Message-
> From: Antony Dovgal [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, June 01, 2005 10:13 PM
> To: Ilia Alshanetsky
> Cc: [EMAIL PROTECTED]; php-cvs@lists.php.net
> Subject: Re: [PHP-CVS] cvs: php-src /main
Yes. I foubd the error using this test file.
Thanks. Dmitry.
> -Original Message-
> From: Ilia Alshanetsky [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, June 01, 2005 9:31 PM
> To: Dmitry Stogov
> Cc: php-cvs@lists.php.net
> Subject: Re: [PHP-CVS] cvs: php-src /main
On Wed, 01 Jun 2005 12:20:51 -0400
Ilia Alshanetsky <[EMAIL PROTECTED]> wrote:
> Is there a test case or an example of a bug that required this fix?
I guess it fixes stream_get_line() buffer overrun that I tried to fix couple of
weeks ago.
At least I can't reproduce it anymore.
> ilia
>
> Dmit
:21 PM
To: Dmitry Stogov
Cc: php-cvs@lists.php.net
Subject: Re: [PHP-CVS] cvs: php-src /main/streams streams.c
Is there a test case or an example of a bug that required this fix?
ilia
Dmitry Stogov wrote:
dmitry Wed Jun 1 11:11:44 2005 EDT
Modified files:
/php-src
No. It fixes valgrind errors.
Thanks. Dmitry.
> -Original Message-
> From: Ilia Alshanetsky [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, June 01, 2005 8:21 PM
> To: Dmitry Stogov
> Cc: php-cvs@lists.php.net
> Subject: Re: [PHP-CVS] cvs: php-src /main/streams streams.c
Is there a test case or an example of a bug that required this fix?
ilia
Dmitry Stogov wrote:
dmitry Wed Jun 1 11:11:44 2005 EDT
Modified files:
/php-src/main/streams streams.c
Log:
Disabled search of recored separator outside of buffer
http://cvs.php.
On 5/23/05, Antony Dovgal <[EMAIL PROTECTED]> wrote:
> > (The|a) correct fix for 32810 is to tag the stream with the
> > plain_files_wrapper in _php_stream_fopen_tmpfile().
>
> If you know the correct fix why don't you just fix it then?
Pick one of:
- Busy
- Mostly without internet access this w
On 5/23/05, Antony Dovgal <[EMAIL PROTECTED]> wrote:
> > 32810 is a bug, but that's because tmpfile() is implemented as a
> > wrapper-less plain file.
>
> As far as I understand wrapper-less streams should not exist at all, right?
Wrapper-less streams are perfectly ok and acceptable, and are the
On Mon, 23 May 2005 11:31:05 -0400
Wez Furlong <[EMAIL PROTECTED]> wrote:
> 30441 sounds like a bogus bug report because no one can reproduce what
> that guy is seeing.
> 29023 sounds like a valid bug.
> 32649 is a db2 bug report ;)
I mean #32469 (With COM ports (Windows), fread sees nothing unti
30441 sounds like a bogus bug report because no one can reproduce what
that guy is seeing.
29023 sounds like a valid bug.
32649 is a db2 bug report ;)
32810 is a bug, but that's because tmpfile() is implemented as a
wrapper-less plain file.
(The|a) correct fix for 32810 is to tag the stream with t
On Mon, 23 May 2005 10:08:39 -0400
Wez Furlong <[EMAIL PROTECTED]> wrote:
> What tests did you run, and what problem were you trying to solve?
>
> You've made reads non-greedy for streams with wrappers (except for
> plain files), and made reads greedy for streams without wrappers.
>
> The correc
What tests did you run, and what problem were you trying to solve?
You've made reads non-greedy for streams with wrappers (except for
plain files), and made reads greedy for streams without wrappers.
The correct semantic is: only reads on local files should be greedy
(for BC); everything else sho
WTF!???
Please revert, immediately.
--Wez.
On 5/23/05, Antony Dovgal <[EMAIL PROTECTED]> wrote:
> tony2001Mon May 23 07:51:53 2005 EDT
>
> Modified files:
> /php-src/main/streams streams.c
> Log:
> fix bug #32810 (fread after tmpfile() reads only 8192 bytes)
>
>
On Mon, 23 May 2005 09:54:44 -0400
Wez Furlong <[EMAIL PROTECTED]> wrote:
> WTF!???
>
> Please revert, immediately.
Please care to explain why.
I tested it with all kinds of wrappers and it works fine for me.
> On 5/23/05, Antony Dovgal <[EMAIL PROTECTED]> wrote:
> > tony2001Mon
Andi Gutmans wrote:
Seems to me best to revert.
Antony, maybe you can get from Ilia the script that reproduced the
valgrind exposure of the problem.
The test case for stream_get_line() (bug #30362) exposes the valgrind error.
Ilia
--
PHP CVS Mailing List (http://www.php.net/)
To unsubscribe, visit
Seems to me best to revert.
Antony, maybe you can get from Ilia the script that reproduced the valgrind
exposure of the problem.
At 01:33 AM 4/8/2005 +0300, Antony Dovgal wrote:
If I should - just tell me, I'll do it.
On Thu, 7 Apr 2005 22:01:14 +0300 (EEST)
Jani Taskinen <[EMAIL PROTECTED]> wrot
If I should - just tell me, I'll do it.
On Thu, 7 Apr 2005 22:01:14 +0300 (EEST)
Jani Taskinen <[EMAIL PROTECTED]> wrote:
>
> Should Antony revert his patch then..?
> (wasn't Wez basically saying that? :)
>
> --Jani
>
>
> On Wed, 6 Apr 2005, Ilia Alshanetsky wrote:
>
> > I ch
Should Antony revert his patch then..?
(wasn't Wez basically saying that? :)
--Jani
On Wed, 6 Apr 2005, Ilia Alshanetsky wrote:
I checked the code with valgrind and it seems to me like the problem lies else
where. According to valgrind (at some point, probably the last line) we have a
I checked the code with valgrind and it seems to me like the problem
lies else where. According to valgrind (at some point, probably the last
line) we have a Conditional jump or move depends on uninitialised
value(s) php_stream_get_record (/php5/main/streams/streams.c:847)
Ilia
Antony Dovgal wr
On Wed, 6 Apr 2005 09:05:35 -0400
Wez Furlong <[EMAIL PROTECTED]> wrote:
> What problem does this fix?
> EOF can be a temporary condition for network streams, depending on
> non-blocking status, and the read call is used to re-enable it. (yay
> for backward compatibility).
stream_get_line() give
What problem does this fix?
EOF can be a temporary condition for network streams, depending on
non-blocking status, and the read call is used to re-enable it. (yay
for backward compatibility).
--Wez.
On Apr 6, 2005 4:26 AM, Antony Dovgal <[EMAIL PROTECTED]> wrote:
> tony2001Wed A
Nice catch!
At 04:34 PM 7/13/2004 +, Wez Furlong wrote:
wez Tue Jul 13 12:34:57 2004 EDT
Modified files:
/php-src/main/streams streams.c
Log:
Looks like Sara forgot to change this too.
http://cvs.php.net/diff.php/php-src/main/streams/streams.c?r1=1.60&r2=1.61&ty=u
In
t; <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
>Sent: Thursday, February 05, 2004 1:31 AM
>Subject: Re: [PHP-CVS] cvs: php-src /main/streams streams.c transports.c
>xp_socket.c
>
>
>>
>> I'm propably missing the point..but that test completes
>&
t;[EMAIL PROTECTED]>
Cc: "Wez Furlong" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Sent: Thursday, February 05, 2004 1:31 AM
Subject: Re: [PHP-CVS] cvs: php-src /main/streams streams.c transports.c
xp_socket.c
>
> I'm propably missing the point..but that test com
g (it won't be simple since
>the guts have changed from 4.3 to 5).
>
>--Wez.
>
>- Original Message -
>From: "Jani Taskinen" <[EMAIL PROTECTED]>
>To: "Wez Furlong" <[EMAIL PROTECTED]>
>Cc: <[EMAIL PROTECTED]>
>Sent: Thursday,
Jani Taskinen" <[EMAIL PROTECTED]>
To: "Wez Furlong" <[EMAIL PROTECTED]>
Cc: <[EMAIL PROTECTED]>
Sent: Thursday, February 05, 2004 12:53 AM
Subject: Re: [PHP-CVS] cvs: php-src /main/streams streams.c transports.c
xp_socket.c
>
> Does this fix have p
Does this fix have possibly anything to do with bug #26863 ??
And how about doing a nice little MFH?
--Jani
On Wed, 4 Feb 2004, Wez Furlong wrote:
>wezWed Feb 4 17:46:45 2004 EDT
>
> Modified files:
>/php-src/main/streams streams.c tran
Ah, it's char **path_for_open
I thought there was some bad pointer stuff going on there.
--Wez.
> #ifdef PHP_WIN32
> if (*(*path_for_open + 1) != ':')
> #endif
> (*path_for_open)--;
> And this last says: If the next character is a colon, we're dealing with
a
> win
> That win32 branch of the code looks suspect
> (I know it is not part of the commit).
> Anyone have time to check it out?
>
Looks to me like it's dealing with the case where file:///C:/foo.ext is
supplied.
*path_for_open = (char*)path + n + 1;
This'll put us at the first ':'
while (*(++*path_fo
That win32 branch of the code looks suspect
(I know it is not part of the commit).
Anyone have time to check it out?
--Wez.
> #ifdef PHP_WIN32
> if (*(*path_for_open + 1) != ':')
> #endif
> - *path_for_open--;
> + (*path_for_open)--;
--
PHP CVS Mailing List (http://www.php.net/)
To unsubscr
30 matches
Mail list logo
