pike issueshttps://git.lysator.liu.se/pikelang/pike/-/issues2009-04-16T14:11:39Zhttps://git.lysator.liu.se/pikelang/pike/-/issues/3912multi-part CMOD with storage dumps core2009-04-16T14:11:39ZPeter Bortasmulti-part CMOD with storage dumps coreImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3912
Reported by BW, Riverweb <hww3@riverweb.com>
When writing modules made up of multiple cmod files, the top level module
class has to have CVARs before PIKEVARS, other...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3912
Reported by BW, Riverweb <hww3@riverweb.com>
When writing modules made up of multiple cmod files, the top level module
class has to have CVARs before PIKEVARS, otherwise the module will cause a
segfault when loading. The following module dumps core (condensed to one
file for clarity), but if the CVAR and PIKEVAR are swapped, loads properly.
module.cmod:
```
typedef struct
{
int t;
} OBJECT_DATA;
CVAR OBJECT_DATA *object_data;
PIKEVAR int i;
pike_init_pam_main()
{
OBJECT_DATA * dta;
dta =
(OBJECT_DATA*)malloc(sizeof(OBJECT_DATA));
if (!dta)
Pike_error("init_test: Out of memory!\n");
THIS->object_data = dta;
}
pike_exit_pam_main()
{
if(THIS->object_data)
{
free(THIS->object_data);
}
}
void pike_module_init()
{
INIT
pike_init_pam_main();
}
void pike_module_exit()
{
EXIT
pike_exit_pam_main();
}
```Pike 7.8Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3911BACKPORT: CMOD module level storage not allocated2009-04-16T14:11:39ZPeter BortasBACKPORT: CMOD module level storage not allocatedImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3911
Reported by BW, Riverweb <hww3@riverweb.com>
The following CMOD (not within a PIKECLASS) will dump core:
```
typedef struct
{
char * data;
int started;
} OB...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3911
Reported by BW, Riverweb <hww3@riverweb.com>
The following CMOD (not within a PIKECLASS) will dump core:
```
typedef struct
{
char * data;
int started;
} OBJECT_DATA;
CVAR OBJECT_DATA *object_data;
PIKEFUN string foo()
{
if(THIS->object_data->data != NULL)
push_text(THIS->object_data->data);
else
push_int(0);
}
```
PIKEFUN void create()
{
}
```
INIT
{
char * blah = strdup("wahoo!");
OBJECT_DATA * dta =
(OBJECT_DATA*)malloc(sizeof(OBJECT_DATA));
if (!dta)
Pike_error("init_test: Out of memory!\n");
THIS->object_data = dta;
THIS->object_data->data = blah;
}
EXIT
{
if(THIS->object_data)
{
if(THIS->object_data->data != NULL)
free(THIS->object_data->data);
free(THIS->object_data);
}
}
```
It appears that the top level storage is not allocated. Grubba's
suggestion to add the following code to the bottom of the module:
```
PIKE_MODULE_INIT
{
_storage_offset = ADD_STORAGE(OBJECT_DATA);
INIT;
}
PIKE_MODULE_EXIT
{
EXIT;
}
```
Solves the problem with the following caveat... the prototype for
event_handler() must be provided, as precompile.pike generates code out of
order for it to be implicitly defined:
```
static void event_handler(int ev);
```Pike 7.6Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3907Optimizer kills loops dependant on variable2009-04-16T14:11:39ZPeter BortasOptimizer kills loops dependant on variableImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3907
Reported by Marcus Agehall <agehall@roxen.com>
The following program exits at once in a recent Pike 7.7 CVS:
```
int run = 1;
```
void gmain() {
```
while(run);...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3907
Reported by Marcus Agehall <agehall@roxen.com>
The following program exits at once in a recent Pike 7.7 CVS:
```
int run = 1;
```
void gmain() {
```
while(run);
exit(0);
}
Thread.Thread t;
int main(int argc, array(string) argv) {
t = Thread(gmain);
return -1;
}
```Pike 7.8Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3894Image.JPEG crashes with corrupt image2022-08-28T23:36:23ZPeter BortasImage.JPEG crashes with corrupt imageImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3894
Reported by Thomas Bopp, Uni-GH Paderborn <astra@upb.de>
The crash only happens on the IA64 architecture (with redhat). We
encountered it when moving to a new server...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3894
Reported by Thomas Bopp, Uni-GH Paderborn <astra@upb.de>
The crash only happens on the IA64 architecture (with redhat). We
encountered it when moving to a new server. I have written a standalone
test program which reproduces the crash. A single call causes only an error
(invalid marker or something like that) and the second call crashes the
program:
*** glibc detected *** double free or corruption (!prev):
0x6000000000245700 ***Pike 7.6Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3862Destructed object not properly treated as zero2011-02-03T15:59:35ZPeter BortasDestructed object not properly treated as zeroImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3862
Reported by Fredrik Noring, Roxen Internet Software <noring@roxen.com>
Reloading the Sitebuilder Main module twice gave this backtrace:
```
Incompatible types
Unkno...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3862
Reported by Fredrik Noring, Roxen Internet Software <noring@roxen.com>
Reloading the Sitebuilder Main module twice gave this backtrace:
```
Incompatible types
Unknown program: `+("VCFile(N/E:0:[(unknown)]:",0)
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/VCFile.pike
(version 1.337):58: Sitebuilder.FS.VCFile()->_sprintf()
pike/lib/pike/master.pike (version 1.259):117: master()->error("Failed to
find name of unencodable object %O.\n",Sitebuilder.FS.VCFile())
pike/lib/pike/master.pike (version 1.259):3490:
master()->nameof(Sitebuilder.FS.VCFile(),UNDEFINED)
base_server/cache.pike (version 1.85):216:
/usr/local/tmp/service/roxen/server-4.0.308/base_server/cache()->cache_clean()
base_server/roxen.pike (version 1.886):1004: roxen->bg_process_queue()
base_server/roxen.pike (version 1.886):642: roxen->handler_thread(1)
Incompatible types
Unknown program: `+("NoVCFile(N/E:0:[(unknown)]:",0)
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/NoVCFile.pike
(version 1.15):58: Sitebuilder.FS.NoVCFile()->_sprintf()
pike/lib/pike/master.pike (version 1.259):117: master()->error("Failed to
find name of unencodable object %O.\n",Sitebuilder.FS.NoVCFile())
pike/lib/pike/master.pike (version 1.259):3490:
master()->nameof(Sitebuilder.FS.NoVCFile(),UNDEFINED)
base_server/cache.pike (version 1.85):216:
/usr/local/tmp/service/roxen/server-4.0.308/base_server/cache()->cache_clean()
base_server/roxen.pike (version 1.886):1004: roxen->bg_process_queue()
base_server/roxen.pike (version 1.886):642: roxen->handler_thread(1)
Incompatible types
Unknown program: `+("VCFile(N/E:2:[(unknown)]:",0)
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/VCFile.pike
(version 1.337):58: Sitebuilder.FS.VCFile()->_sprintf()
pike/lib/pike/master.pike (version 1.259):117: master()->error("Failed to
find name of unencodable object %O.\n",Sitebuilder.FS.VCFile())
pike/lib/pike/master.pike (version 1.259):3490:
master()->nameof(Sitebuilder.FS.VCFile(),UNDEFINED)
base_server/cache.pike (version 1.85):216:
/usr/local/tmp/service/roxen/server-4.0.308/base_server/cache()->cache_clean()
base_server/roxen.pike (version 1.886):1004: roxen->bg_process_queue()
base_server/roxen.pike (version 1.886):642: roxen->handler_thread(1)
Incompatible types
Unknown program: `+("VCFile(N/E:0:[(unknown)]:",0)
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/VCFile.pike
(version 1.337):58: Sitebuilder.FS.VCFile()->_sprintf()
pike/lib/pike/master.pike (version 1.259):117: master()->error("Failed to
find name of unencodable object %O.\n",Sitebuilder.FS.VCFile())
pike/lib/pike/master.pike (version 1.259):3490:
master()->nameof(Sitebuilder.FS.VCFile(),UNDEFINED)
base_server/cache.pike (version 1.85):216:
/usr/local/tmp/service/roxen/server-4.0.308/base_server/cache()->cache_clean()
base_server/roxen.pike (version 1.886):1004: roxen->bg_process_queue()
base_server/roxen.pike (version 1.886):642: roxen->handler_thread(1)
Incompatible types
Unknown program: `+("NoVCFile(N/E:0:[(unknown)]:",0)
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/NoVCFile.pike
(version 1.15):58: Sitebuilder.FS.NoVCFile()->_sprintf()
pike/lib/pike/master.pike (version 1.259):117: master()->error("Failed to
find name of unencodable object %O.\n",Sitebuilder.FS.NoVCFile())
pike/lib/pike/master.pike (version 1.259):3490:
master()->nameof(Sitebuilder.FS.NoVCFile(),UNDEFINED)
base_server/cache.pike (version 1.85):216:
/usr/local/tmp/service/roxen/server-4.0.308/base_server/cache()->cache_clean()
base_server/roxen.pike (version 1.886):1004: roxen->bg_process_queue()
base_server/roxen.pike (version 1.886):642: roxen->handler_thread(1)
Incompatible types
Unknown program: `+("VCFile(N/E:2:[(unknown)]:",0)
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/VCFile.pike
(version 1.337):58: Sitebuilder.FS.VCFile()->_sprintf()
pike/lib/pike/master.pike (version 1.259):117: master()->error("Failed to
find name of unencodable object %O.\n",Sitebuilder.FS.VCFile())
pike/lib/pike/master.pike (version 1.259):3490:
master()->nameof(Sitebuilder.FS.VCFile(),UNDEFINED)
base_server/cache.pike (version 1.85):216:
/usr/local/tmp/service/roxen/server-4.0.308/base_server/cache()->cache_clean()
base_server/roxen.pike (version 1.886):1004: roxen->bg_process_queue()
base_server/roxen.pike (version 1.886):642: roxen->handler_thread(1)
```Pike 7.4Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3855mktime crashes pike on NT2009-04-16T14:11:39ZPeter Bortasmktime crashes pike on NTImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3855
Reported by Marcus Wellhardh <wellhard@roxen.com>
The following test crashes pike on windows:
```
C:\Roxen CMS\4.0\server-4.0.296\pike\bin>pike
Pike v7.4 release 24...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3855
Reported by Marcus Wellhardh <wellhard@roxen.com>
The following test crashes pike on windows:
```
C:\Roxen CMS\4.0\server-4.0.296\pike\bin>pike
Pike v7.4 release 240 running Hilfe v3.5 (Incremental Pike Frontend)
> mktime((["sec":0,"min":0,"hour":0,"mday":1,"mon":0,"year":1900-1934]));
```Pike 7.4Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3820Calendar.Month() offsets works in an unnatural way2009-04-23T14:08:31ZPeter BortasCalendar.Month() offsets works in an unnatural wayImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3820
Reported by Marcus Wellhardh <wellhard@roxen.com>
```
Reported in RT#5759.
```
The date tag can give unlogic results when using the months attribute. Example:
<d...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3820
Reported by Marcus Wellhardh <wellhard@roxen.com>
```
Reported in RT#5759.
```
The date tag can give unlogic results when using the months attribute. Example:
<date iso-time="2004-07-31" months="-1" strftime="%Y-%m-%d"/>
->
2004-07-31
Mysql has the following behaviour:
select DATE_SUB('2004-07-31', INTERVAL 1 month)
->
2004-06-30Pike 7.4https://git.lysator.liu.se/pikelang/pike/-/issues/3813Tiff image crashes pike on Windows2009-04-16T14:11:39ZPeter BortasTiff image crashes pike on WindowsImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3813
Reported by Marcus Wellhardh <wellhard@roxen.com>
```
The attached image crashes pike on Windows (Roxen CMS 4.0.172-release3).
Reported in RT#5738.
C:\Roxen CMS\4.0...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3813
Reported by Marcus Wellhardh <wellhard@roxen.com>
```
The attached image crashes pike on Windows (Roxen CMS 4.0.172-release3).
Reported in RT#5738.
C:\Roxen CMS\4.0\server-4.0.172\pike\bin>pike
Pike v7.4 release 162 running Hilfe v3.5 (Incremental Pike Frontend)
> object o = Image.load_layer("c:\\Logotyp27.tif");
```
The same image works in Roxen 3.4.136 on Windows.Pike 7.4Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3793memory-leak in oracle module2009-04-16T14:11:39ZPeter Bortasmemory-leak in oracle moduleImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3793
Reported by Michael Stenitzer, E.V.A. <webmaster@eva.wsr.ac.at>
we have watched a memory-leak in Pike v7.4 release 139 with Roxen WebServer
4.0.141 most likly residi...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3793
Reported by Michael Stenitzer, E.V.A. <webmaster@eva.wsr.ac.at>
we have watched a memory-leak in Pike v7.4 release 139 with Roxen WebServer
4.0.141 most likly residing in oracle module's C code.
after retrieving about 800 pages in three parallel recursive wgets for 2
hours from a website with an oracle db memory demand grew up to 550 MB (and
was then manually stopped, because it would have grown even more).
"Pike memory usage information" showed that pike's memory demand has been
more or less constant so we are assuming that growing memory demand is
coming from c-code.
this growing memory demand has not been seen with websites using mysql
instead of oracle.Pike 7.4Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3789Assignment of out-idexed arrays too lenient. (Was: Type system too gullible f...2009-04-16T14:11:39ZPeter BortasAssignment of out-idexed arrays too lenient. (Was: Type system too gullible for soft casts of constants.)Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3789
Reported by Martin Nilsson, IDA <nilsson@pike.ida.liu.se>
The following prompts a compilation error
void main() { array bar; bar=1; }
while the following only a wa...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3789
Reported by Martin Nilsson, IDA <nilsson@pike.ida.liu.se>
The following prompts a compilation error
void main() { array bar; bar=1; }
while the following only a warning.
void main() { array bar; bar=[array]1; }
This causes the following to pass in hilfe
```
> array bar; bar=16;
(1) Result: 16
```
but strangely enough not the following
```
> string bar; bar=16;
Compiler Error: 1:Bad type in assignment.
Compiler Error: 1:Expected: string
Compiler Error: 1:Got : int(16..16)
```Pike 7.8Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3788Parser.XML.Tree namespaces wrong2011-02-08T13:34:54ZPeter BortasParser.XML.Tree namespaces wrongImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3788
Reported by Martin Nilsson, IDA <nilsson@pike.ida.liu.se>
Given that I understand the API correctly I think
```
> object x=Parser.XML.Tree.parse_input("<foo:x xmlns...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3788
Reported by Martin Nilsson, IDA <nilsson@pike.ida.liu.se>
Given that I understand the API correctly I think
```
> object x=Parser.XML.Tree.parse_input("<foo:x xmlns:foo='uri://bar/'/>")[0];
> x->get_tag_name();
(1) Result: "x"
> x->get_short_name();
(2) Result: "x"
> x->get_full_name();
(3) Result: "foo:x"
```
should return
"x", "foo:x" and "uri://bar/x".Pike 7.8Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3775Oracle problems in 4.02009-04-16T14:11:39ZPeter BortasOracle problems in 4.0Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3775
Reported by Marcus Wellhardh <wellhard@roxen.com>
The Oracle support does not work in Roxen 4.0. The following works in 3.4:
```
Pike v7.4 release 124 running Hilfe...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3775
Reported by Marcus Wellhardh <wellhard@roxen.com>
The Oracle support does not work in Roxen 4.0. The following works in 3.4:
```
Pike v7.4 release 124 running Hilfe v3.5 (Incremental Pike Frontend)
> object db = Sql.oracle("TEST", 0, "scott", "tiger");
> destruct(db->big_query("select * from tab"));
Attempting to access parent of destructed object.
Unknown program: unknown function()
... last 1 frames above repeated 1 times ...
HilfeInput:1: HilfeInput()->___HilfeWrapper()
```Pike 7.4https://git.lysator.liu.se/pikelang/pike/-/issues/3761MySQL 4.1.3 changes mysql_shutdown() function2009-04-16T14:11:39ZPeter BortasMySQL 4.1.3 changes mysql_shutdown() functionImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3761
Reference: http://dev.mysql.com/doc/mysql/en/mysql_shutdown.html
Reported by Dan Nelson <dnelson_1901@yahoo.com>
The 4.1.3 release of mysql adds another argument to...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3761
Reference: http://dev.mysql.com/doc/mysql/en/mysql_shutdown.html
Reported by Dan Nelson <dnelson_1901@yahoo.com>
The 4.1.3 release of mysql adds another argument to mysql_shutdown().
Currently the only allowed value is SHUTDOWN_DEFAULT. I assume later
versions will add other options, but since there is currently only one,
this patch should be sufficient to allow pike to build against both old and
new versions.
```
===================================================================
RCS file: /cvs/Pike/7.6/src/modules/Mysql/mysql.c,v
retrieving revision 1.81
diff -u -r1.81 mysql.c
--- mysql.c 19 Mar 2004 14:49:44 -0000 1.81
+++ mysql.c 13 Jul 2004 16:43:47 -0000
@@ -59,6 +59,10 @@
#endif /* HAVE_MYSQL_ERRMSG_H */
#endif /* HAVE_ERRMGS_H */
+#if MYSQL_VERSION_ID >= 40103
+/* 4.1.3 added an extra shutdown_level argument */
+#define mysql_shutdown(m) mysql_shutdown(m, SHUTDOWN_DEFAULT);
+#endif
```
#endif /* HAVE_MYSQL */Pike 7.6Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3759File stat on smb shares does not work2020-03-01T15:33:04ZPeter BortasFile stat on smb shares does not workImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3759
Reported by Marcus Wellhardh <wellhard@roxen.com>
The following returnes false even if the share exists:
```
Stdio.exists("\\\\xps12.roxen.com\\cms40");
```
The ...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3759
Reported by Marcus Wellhardh <wellhard@roxen.com>
The following returnes false even if the share exists:
```
Stdio.exists("\\\\xps12.roxen.com\\cms40");
```
The following returnes true:
```
Stdio.exists("\\\\xps12.roxen.com\\cms40\Basic.rep");
```Pike 7.4Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3755Support for TLS 1.12009-04-16T14:11:39ZPeter BortasSupport for TLS 1.1Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3755
Reported by @grubba
```
From: "Yngve N. Pettersen (Developer Opera Software ASA)" <yngve@opera.com>
Organization: Opera Software AS
To: "grubba@roxen.com" <grubba@ro...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3755
Reported by @grubba
```
From: "Yngve N. Pettersen (Developer Opera Software ASA)" <yngve@opera.com>
Organization: Opera Software AS
To: "grubba@roxen.com" <grubba@roxen.com>,
"nilsson@roxen.com" <nilsson@roxen.com>, "bill@roxen.com"
<bill@roxen.com>
Cc: Yngve N. Pettersen <yngve@opera.com>, Developer Opera Software ASA:
Date: Sun, 04 Jul 2004 19:35:57 +0200
Subject: Roxen/Pike is not TLS (Transport Layer Security) forward
compatible with TLS 1.1
```
Hi,
I noticed your names in recent updates of the file "handshake.pike" file
of the Roxen/Pike CVS repository, and I am sending this email directly to
you as I assume you would be the ones to handle this report anyway. Just
in case I'm writing this in English.
I am a developer with Opera Software ASA, and one of my work areas is the
Opera browser's SSL/TLS protocol support.
I've recently been testing Opera's new TLS 1.1 implementation (presently
an IETF draft status document) and have unfortunately found that the
Roxen/Pike server is not (at least according to my understanding) properly
implementing RFC 2246.
The result is that no TLS 1.1 capable client is able to connect to a
Roxen/Pike TLS server, *unless* TLS 1.1 is disabled.
If I read the source at
http://community.roxen.com/_internal/cvsview!0/94970/1.49/1/handshake.pike
correctly, any version number higher than 3.1 (TLS 1.0) will result in a
closed connection.
RFC 2246 (and the SSL v3 specification before it) clearly states (RFC 2246
sec. 7.4.1.2 and 7.4.1.3):
client_version
The version of the TLS protocol by which the client wishes to
communicate during this session. This should be the latest
(highest valued) version supported by the client. For this
version of the specification, the version will be 3.1
server_version
This field will contain the lower of that suggested by the client
in the client hello and the highest supported by the server. For
this version of the specification, the version is 3.1
Similar language exists in the TLS 1.1 draft, except the version number is
3.2.
This means that a client even if the client sends version 3.0 the server
should respond with 3.0, but it also imply that if the client sends
version 3.2 the server should send 3.1 (if TLS 1.0 is the highest version
supported).
According to my reading, Roxen/Pike does not do this at present, it just
terminates the connection.
What your code should do is to keep the version number sent by the client
in a separate version number storage, as it is needed to validate the RSA
encrypted premaster secret message block (RFC 2246 sec 7.4.7.1), and then
set the actual version number to be used to the lowest version number of
the higest supported by the client and the highest supported by the
server. In your case that means falling back to TLS 1.0.
IMO you should fix this problem as soon as possible.
I expect that TLS 1.1 will become an RFC within 3-12 months (just my
opinion, something might delay it), and Opera will most likely start
shipping final versions with TLS 1.1 shortly afterwards. AFAIK there are
at least two TLS 1.1 capable software libraries available: GnuTLS and
cryptolib.
References:
TLS 1.0: http://www.ietf.org/rfc/rfc2246.txt
TLS 1.1 draft:
http://www.ietf.org/internet-drafts/draft-ietf-tls-rfc2246-bis-06.txt
--
Sincerely,
Yngve N. Pettersen
********************************************************************
Senior Developer Email: yngve@opera.com
Opera Software ASA http://www.opera.com/
Phone: +47 24 16 42 60 Fax: +47 24 16 40 01
********************************************************************Pike 7.4Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3745Float.MAX segmentation fault2011-02-08T13:32:09ZPeter BortasFloat.MAX segmentation faultImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3745
Reported by M. Ferecatu, none <mferecat@numericable.fr>
System: Slackware Linux, 9.1 and Redhat 9.0, Pentium III, 600MHz
Compilers: gcc-3.2.2 and gcc-3.3.4
Pike vers...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3745
Reported by M. Ferecatu, none <mferecat@numericable.fr>
System: Slackware Linux, 9.1 and Redhat 9.0, Pentium III, 600MHz
Compilers: gcc-3.2.2 and gcc-3.3.4
Pike version: current cvs (7.7.5 snapshot from 2004/06/23), Pike 7.6.6 release
Compiled with './configure --with-double-precision' pike segfaults on the
following simple sequence:
```
$ pike
Pike v7.7 release 5 running Hilfe v3.5 (Incremental Pike Frontend)
> Float.MAX;
Segmentation fault
```
All other constants from module float seem to work. Also, when compiled
without double precision the bug dissapears.
I Posted a gdb session on pike@roxen.org mailing-list.
Cheers,
M.
Cheers.Pike 7.8Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3695Stricter type checks aren't turned off in compat mode2011-02-08T11:07:00ZPeter BortasStricter type checks aren't turned off in compat modeImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3695
Reported by Martin Stjernholm <mast@roxen.com>
I get type errors like the following when I run pike 7.6 with -V7.4. It
doesn't happen in pike 7.4.
```
Failed to com...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3695
Reported by Martin Stjernholm <mast@roxen.com>
I get type errors like the following when I run pike 7.6 with -V7.4. It
doesn't happen in pike 7.4.
```
Failed to compile module
/home/mast/Roxen/4.0/server/modules/sitebuilder/modules/manager/content_editor.pike:
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:45
Error: Bad type in assignment.
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:45
Error: Expected: { SBStat = object(implements 66514) }
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:45
Error: Got : { Stat = object(implements 10) }
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:45
Error: Identifier "_sprintf" is incompatible.
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:45
Error: Expected: function( : string)
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:45
Error: Got : function(int, void | mapping : string)
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:49
Error: Wrong return type.
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:49
Error: Expected: { Stat = object(implements 10) }
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:49
Error: Got : { SBStat = object(implements 66514) }
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:49
Error: Identifier "_sprintf" is incompatible.
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:49
Error: Expected: function( : string)
modules/sitebuilder/pike-modules/Sitebuilder.pmod/FS.pmod/SBFileData.pike:49
Error: Got : function(int, void | mapping : string)
```Pike 7.6Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3693BACKPORT: copy_value() doesn't preserve the identity relation2011-02-03T14:27:39ZPeter BortasBACKPORT: copy_value() doesn't preserve the identity relationImported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3693
Reported by @grubba
```
From: PeterPan <peterpan@airinbox.com>
Resent-From: pike@roxen.com
To: "pike@roxen.com" <pike@roxen.com>
Date: Thu, 27 May 2004 10:05:45 +080...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3693
Reported by @grubba
```
From: PeterPan <peterpan@airinbox.com>
Resent-From: pike@roxen.com
To: "pike@roxen.com" <pike@roxen.com>
Date: Thu, 27 May 2004 10:05:45 +0800
Resent-Date: Thu, 27 May 2004 04:06:11 +0200 (MEST)
Subject: seems a bug
look this:
[peterpan@server01 ~/pikenv]$pike
Pike v7.6 release 6 running Hilfe v3.5 (Incremental Pike Frontend)
> array a1=({([])})*2;a1[0]==a1[1];
(1) Result: 0
> int t=2;array a2=({([])})*t;a2[0]==a2[1];
(2) Result: 1
```
the items in a1 are not same, but the items in a2 are same.Pike 7.6Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3692LDAP via SSL not working.2009-04-16T14:11:39ZPeter BortasLDAP via SSL not working.Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3692
Reported by Anders Johansson <anders@roxen.com>
This could be a variant of [bug #3611].
```
16:22:53 : AC:LDAPimport: open_dir: FALSE! Reason: [Attempt to call the...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3692
Reported by Anders Johansson <anders@roxen.com>
This could be a variant of [bug #3611].
```
16:22:53 : AC:LDAPimport: open_dir: FALSE! Reason: [Attempt to call the
NULL-value
4m14.8s : ]
: AC:LDAPimport: Attempt to call the NULL-value
: Unknown program: 0(0)
: /home/anders/Pike/7.4/lib/modules/SSL.pmod/sslfile.pike
(version 1.55):541: SSL.sslfile(0,SSL.context())->update_callbacks()
: /home/anders/Pike/7.4/lib/modules/SSL.pmod/sslfile.pike
(version 1.55):147: SSL.sslfile(0,SSL.context())->queue_write()
16:22:54 : /home/anders/Pike/7.4/lib/modules/SSL.pmod/sslfile.pike
(version 1.55):354: SSL.sslfile(0,SSL.context())->write_blocking()
4m14.9s : /home/anders/Pike/7.4/lib/modules/SSL.pmod/sslfile.pike
(version 1.55):326:
SSL.sslfile(0,SSL.context())->get_blocking_to_handshake_finished_state()
: /home/anders/Pike/7.4/lib/modules/SSL.pmod/sslfile.pike
(version 1.55):679:
SSL.sslfile(0,SSL.context())->create(Fd(),SSL.context(),1,1)
:
/home/anders/Pike/7.4/lib/modules/Protocols.pmod/LDAP.pmod/client.pike
(version 1.47):412: Protocols.LDAP.client("socket", "ldap.roxen.com:636",
777 /* fd=41
*/)->create("ldaps://ldap.roxen.com/ou=People,dc=roxen,dc=com??sub?",SSL.context())
: modules/sitebuilder/modules/ac/acimport_ldap.pike (version
1.49):38: RoxenModule(basic/acimport_ldap#0)->open_dir()
: modules/sitebuilder/modules/ac/acimport_ldap.pike (version
1.49):85: RoxenModule(basic/acimport_ldap#0)->getusers()
16:22:54 : modules/sitebuilder/modules/ac/acimport_ldap.pike (version
1.49):575: RoxenModule(basic/acimport_ldap#0)->get_data(AC_DB[7])
4m14.9s : modules/sitebuilder/modules/ac/acimport_ldap.pike (version
1.49):952: RoxenModule(basic/acimport_ldap#0)->periodic_update(0)
[...]
```Pike 7.4Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbströmhttps://git.lysator.liu.se/pikelang/pike/-/issues/3687An other segfault in image module.2013-12-11T18:18:26ZPeter BortasAn other segfault in image module.Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3687
Reported by Anders Johansson <anders@roxen.com>
This code generates a segfault, probably due to the rotate operation when
the layer size is zero.
<gxml format="jpeg...Imported from http://bugzilla.roxen.com/bugzilla/show_bug.cgi?id=3687
Reported by Anders Johansson <anders@roxen.com>
This code generates a segfault, probably due to the rotate operation when
the layer size is zero.
<gxml format="jpeg">
<load-image src="/roxen.jpg"/>
<poly color="black">
<c x='0' y='0' />
<c x='10' y='0' />
<c x='10' y='0' />
<c x='0' y='0' />
</poly>
<rotate degrees="-90"/>
</gxml>Pike 7.4Henrik (Grubba) GrubbströmHenrik (Grubba) Grubbström