Re: qcad does not segfault when qtscript is compiled with -O0.
Thus said Stuart Henderson on Tue, 26 Dec 2023 23:25:42 +: > This was already diagnosed and fixed in -current, see > https://marc.info/?l=openbsd-ports&m=170160684730438&w=2 Indeed it was. While I do follow ports@ I must not have noticed because I only recently discovered the association with qtscript and -O0 today. Thanks for the response. Andy
Re: qcad does not segfault when qtscript is compiled with -O0.
This was already diagnosed and fixed in -current, see
https://marc.info/?l=openbsd-ports&m=170160684730438&w=2
https://github.com/openbsd/ports/blob/e10f7e1a39f85873183c232462850449f777e16f/x11/qt5/qtscript/patches/patch-src_3rdparty_javascriptcore_JavaScriptCore_interpreter_CallFrame_h
On 2023-12-26, Andy Bradford
wrote:
> Hello,
>
> A little over a year ago I reported[1] that qcad started having
> segfaults in 7.0 (or 7.1). I thought I would try to debug it today in
> 7.4 and so enabled DEBUG_PACKAGES and DEBUG variables in the cad/qcad
> Makefile and also the x11/qt5/qtscript Makefile hoping to debug the
> segfault. In the Makefile for both I set:
>
> DEBUG_PACKAGES= ${BUILD_PACKAGES}
> DEBUG=-O0 -g
>
> Then I built the packages, installed them and their debug packages and
> hoped to capture a segfault, but it doesn't crash. Apparently enabling
> -O0 in qtscript's Makefile makes it not crash. To test that hypothesis,
> I commented out the DEBUG line completely and rebuilt qtscript with just
> the defaults that DEBUG_PACKAGES inherits and sure enough, after
> installing that qtscript package built with -O2 qcad again crashes.
>
> (gdb) bt
> #0 0x03eae2f9a990 in QScriptEnginePrivate::mark (this=0x3eb4af54a80,
> markStack=...) at
> /usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/script/api/qscriptengine.cpp:1320
> #1 0x03eae2f14d6c in QTJSC::Heap::markRoots (this=0x3ea8c2129e8) at
> /usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/3rdparty/javascriptcore/JavaScriptCore/runtime/Collector.cpp:1167
> #2 0x03eae2f13ac5 in QTJSC::Heap::reset (this=0x3ea8c2129e8) at
> /usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/3rdparty/javascriptcore/JavaScriptCore/runtime/Collector.cpp:1269
> #3 QTJSC::Heap::allocate (this=0x3ea8c2129e8, s=) at
> /usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/3rdparty/javascriptcore/JavaScriptCore/runtime/Collector.cpp:405
> #4 0x03eae2fa3ac2 in QTJSC::JSCell::operator new (size=56,
> exec=) at
> /usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/3rdparty/javascriptcore/JavaScriptCore/runtime/JSCell.h:167
> #5 QScriptEnginePrivate::newVariant (this=0x3eb4af54a80, value=...) at
> /usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/script/api/qscriptengine.cpp:1797
> #6 0x03eae2fa51b5 in QScriptEngine::newVariant (this=,
> value=...) at
> /usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/script/api/qscriptengine.cpp:2233
> #7 0x03eade54f63f in qtscript_create_Qt_Axis_class
> (engine=0x3eaf5f91770, clazz=...) at
> ../../generated_cpp/com_trolltech_qt_core/qtscript_Qt.cpp:4157
> #8 0x03eade5445ee in qtscript_create_Qt_class (engine=0x3eaf5f91770) at
> ../../generated_cpp/com_trolltech_qt_core/qtscript_Qt.cpp:8386
> #9 0x03eade62a327 in qtscript_initialize_com_trolltech_qt_core_bindings
> (extensionObject=...) at
> ../../generated_cpp/com_trolltech_qt_core/com_trolltech_qt_core_init.cpp:312
> #10 0x03eade3a1d9e in com_trolltech_qt_core_ScriptPlugin::initialize
> (this=0x3eabb1f6100, key=..., engine=0x3eaf5f91770) at plugin.cpp:21
> #11 0x03eae2fabdf9 in QScriptEngine::importExtension (this=0x3eaf5f91770,
> extension=...) at
> /usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/script/api/qscriptengine.cpp:3769
> #12 0x03ea8abc35ab in RScriptHandlerEcma::RScriptHandlerEcma
> (this=0x3eabb1e94e0) at RScriptHandlerEcma.cpp:434
> #13 0x03ea8abd4919 in RScriptHandlerEcma::factory () at
> RScriptHandlerEcma.cpp:1074
> #14 0x03ead8eb2c3e in RScriptHandlerRegistry::createScriptHandler
> (extension=...) at RScriptHandlerRegistry.cpp:67
> #15 0x03ead8eb2caf in RScriptHandlerRegistry::getGlobalScriptHandler
> (extension=...) at RScriptHandlerRegistry.cpp:76
> #16 0x03e87ec05021 in main (argc=1, argv=0x7b524ae78618) at main.cpp:389
> (gdb) list
> 1315}
> 1316
> 1317if (q) {
> 1318QScriptContext *context = q->currentContext();
> 1319
> 1320while (context) {
> 1321JSC::ScopeChainNode *node =
> frameForContext(context)->scopeChain();
> 1322JSC::ScopeChainIterator it(node);
> 1323for (it = node->begin(); it != node->end(); ++it) {
> 1324JSC::JSObject *object = *it;
>
> Stepping through with -O2, it seems to go through one or two iterations of a
> while loop and then crashes:
>
> (gdb) print context
> $8 = (QScriptContext *) 0xf0800cd3288
> (gdb) print *context
> $9 = {d_ptr = 0x1}
> (gdb) next
> 1323for (it = node->begin(); it != node->end(); ++it) {
> (gdb)
> 1324JSC::JSObject *object = *it;
> (gdb)
> 1325if (object)
> (gdb)
> 1326markStack.append(object);
> (gdb)
> 1323
qcad does not segfault when qtscript is compiled with -O0.
Hello,
A little over a year ago I reported[1] that qcad started having
segfaults in 7.0 (or 7.1). I thought I would try to debug it today in
7.4 and so enabled DEBUG_PACKAGES and DEBUG variables in the cad/qcad
Makefile and also the x11/qt5/qtscript Makefile hoping to debug the
segfault. In the Makefile for both I set:
DEBUG_PACKAGES= ${BUILD_PACKAGES}
DEBUG= -O0 -g
Then I built the packages, installed them and their debug packages and
hoped to capture a segfault, but it doesn't crash. Apparently enabling
-O0 in qtscript's Makefile makes it not crash. To test that hypothesis,
I commented out the DEBUG line completely and rebuilt qtscript with just
the defaults that DEBUG_PACKAGES inherits and sure enough, after
installing that qtscript package built with -O2 qcad again crashes.
(gdb) bt
#0 0x03eae2f9a990 in QScriptEnginePrivate::mark (this=0x3eb4af54a80,
markStack=...) at
/usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/script/api/qscriptengine.cpp:1320
#1 0x03eae2f14d6c in QTJSC::Heap::markRoots (this=0x3ea8c2129e8) at
/usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/3rdparty/javascriptcore/JavaScriptCore/runtime/Collector.cpp:1167
#2 0x03eae2f13ac5 in QTJSC::Heap::reset (this=0x3ea8c2129e8) at
/usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/3rdparty/javascriptcore/JavaScriptCore/runtime/Collector.cpp:1269
#3 QTJSC::Heap::allocate (this=0x3ea8c2129e8, s=) at
/usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/3rdparty/javascriptcore/JavaScriptCore/runtime/Collector.cpp:405
#4 0x03eae2fa3ac2 in QTJSC::JSCell::operator new (size=56, exec=) at
/usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/3rdparty/javascriptcore/JavaScriptCore/runtime/JSCell.h:167
#5 QScriptEnginePrivate::newVariant (this=0x3eb4af54a80, value=...) at
/usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/script/api/qscriptengine.cpp:1797
#6 0x03eae2fa51b5 in QScriptEngine::newVariant (this=,
value=...) at
/usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/script/api/qscriptengine.cpp:2233
#7 0x03eade54f63f in qtscript_create_Qt_Axis_class (engine=0x3eaf5f91770,
clazz=...) at ../../generated_cpp/com_trolltech_qt_core/qtscript_Qt.cpp:4157
#8 0x03eade5445ee in qtscript_create_Qt_class (engine=0x3eaf5f91770) at
../../generated_cpp/com_trolltech_qt_core/qtscript_Qt.cpp:8386
#9 0x03eade62a327 in qtscript_initialize_com_trolltech_qt_core_bindings
(extensionObject=...) at
../../generated_cpp/com_trolltech_qt_core/com_trolltech_qt_core_init.cpp:312
#10 0x03eade3a1d9e in com_trolltech_qt_core_ScriptPlugin::initialize
(this=0x3eabb1f6100, key=..., engine=0x3eaf5f91770) at plugin.cpp:21
#11 0x03eae2fabdf9 in QScriptEngine::importExtension (this=0x3eaf5f91770,
extension=...) at
/usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/script/api/qscriptengine.cpp:3769
#12 0x03ea8abc35ab in RScriptHandlerEcma::RScriptHandlerEcma
(this=0x3eabb1e94e0) at RScriptHandlerEcma.cpp:434
#13 0x03ea8abd4919 in RScriptHandlerEcma::factory () at
RScriptHandlerEcma.cpp:1074
#14 0x03ead8eb2c3e in RScriptHandlerRegistry::createScriptHandler
(extension=...) at RScriptHandlerRegistry.cpp:67
#15 0x03ead8eb2caf in RScriptHandlerRegistry::getGlobalScriptHandler
(extension=...) at RScriptHandlerRegistry.cpp:76
#16 0x03e87ec05021 in main (argc=1, argv=0x7b524ae78618) at main.cpp:389
(gdb) list
1315}
1316
1317if (q) {
1318QScriptContext *context = q->currentContext();
1319
1320while (context) {
1321JSC::ScopeChainNode *node =
frameForContext(context)->scopeChain();
1322JSC::ScopeChainIterator it(node);
1323for (it = node->begin(); it != node->end(); ++it) {
1324JSC::JSObject *object = *it;
Stepping through with -O2, it seems to go through one or two iterations of a
while loop and then crashes:
(gdb) print context
$8 = (QScriptContext *) 0xf0800cd3288
(gdb) print *context
$9 = {d_ptr = 0x1}
(gdb) next
1323for (it = node->begin(); it != node->end(); ++it) {
(gdb)
1324JSC::JSObject *object = *it;
(gdb)
1325if (object)
(gdb)
1326markStack.append(object);
(gdb)
1323for (it = node->begin(); it != node->end(); ++it) {
(gdb)
1329context = context->parentContext();
(gdb)
1320while (context) {
(gdb) print context
$10 = (QScriptContext *) 0x1
(gdb) next
Thread 1 received signal SIGSEGV, Segmentation fault.
0x0f07f754b990 in QScriptEnginePrivate::mark (this=0xf079205c480,
markStack=...)
at
/usr/ports/pobj/qtscript-5.15.10/qtscript-everywhere-src-5.15.10/src/script/api/qscriptengine.cpp:1320
1320while (co
Re: self-hosted man.openbsd.org script?
Hi Paul,
Paul Pace wrote on Sun, Dec 24, 2023 at 05:25:55AM -0800:
> I have this vague memory of reading someone who posted a script, IIRC,
> to convert the system's man pages to HTML, or similar, into somewhere
> under /var/www and the pages worked just like the highly useful
> man.openbsd.org, and not like the plain text pages that everyone always
> posts to their websites.
>
> Does someone happen to know where that is?
I don't know about any such "script" and believe using a script for it
would be a bad idea - a dirty hack at best.
Converting a whole tree of manual pages to a different format
sounds like the job for the tradition catman(8) utility program
that Christoph Robitschko first implemented in 1993. NetBSD
and FreeBSD contain implementations by various other authors.
OpenBSD does not contain the catman(8) utility because it is rarely
needed by ordinary users and we tend to only include code in the base
system that is useful for many people.
However, the portable mandoc distribution does contain a version
that i wrote together with Michael Stapelberg (of the Debian project)
in 2017:
https://mandoc.bsd.lv/
https://mandoc.bsd.lv/man/catman.8.html
For your purpose, you may want to replace to line
options.fragment = 1;
in mandocd.c by something like
{ char style[] = "/usr/share/misc/mandoc.css";
options.style = style; }
before compiling such that you get complete HTML code including
, , , and elements.
Be careful to not clobber the system mandoc(1) installation by blindly
running "make install". Instead, just manually installing the
binary "mandocd" anywhere in the $PATH is enough. Installing
the "catman" binary in not necessary but won't hurt either.
Also note that viewing the results with a monster browser like
firefox or chrome may require some tweaking with respect to unveil(2),
see the respective files below /usr/local/share/doc/pkg-readmes/.
I freely admit all this is not particularly user-friendly but more
geared towards the needs of server admins. For example, the
server manpages.debian.org is essentially using something similar
to this method. For them, it's critical that this implementation
of catman(8) is much more efficient than the NetBSD and FreeBSD
implementations: it saves lots of time because it does *not* fork
and exec a new parser/formatter process for every manual page
but instead merely reinitialized and resuses the same parser and
formatter process over and over again, which is *much* faster.
With the huge amount of manual pages Debian has to format very
often, their server would not be able to keep up without that
optimization.
Yours,
Ingo
systat mbuf vs. netstat -m
In various places, netstat/mbuf.c does stuff like this against the hiwat value:
printf("%u/%lu mbuf %d byte clusters in use"
" (current/peak)\n",
mclpools[i].pr_nout,
(unsigned long)
mclpools[i].pr_hiwat * mclpools[i].pr_itemsperpage,
mclpools[i].pr_size);
for this result:
3/144 mbuf 2048 byte clusters in use (current/peak)
1011/83430 mbuf 2112 byte clusters in use (current/peak)
the hiwat according to systat mbuf is:
IFACERING LIVELOCKS SIZE ALIVE LWM HWM CWM
System mbufs 0 256 23115254
mcl2k2048 6 18
mcl2k2 2112 10145562
Is this sensible?
Re: OpenSSH update web page: typo
revision 1.147 date: 2023/12/20 17:30:01; author: millert; state: Exp; lines: +3 -3; commitid: nZ6tdVWYkmCCLb6k; Correct the links in the 9.6 section. Reported by Christos Zoulas. Are you guys kidding? :) On Tue, Dec 19, 2023 at 9:35 AM Alex Naumov wrote: > Hey, > > It seems there are two same update manuals for OpenSSH 9.5 and 9.6[1]. > Link to the tarball and the second shell command should be updated. > > Cheers, > Alex > > [1] https://www.openssh.com/openbsd.html >
