Thread: [Htmlvalidator-help] tidy FreeBSD
Brought to you by:
mgueury
From: [LoN]Kamikaze <LoN...@gm...> - 2006-06-08 13:38:16
Attachments:
signature.asc
|
When trying to install the extension as root an alert pops up saying: Html Validator-002: Failed autoRegistering renamed platform specific nsti= dy.dll Does it mean it's looking for nstidy.dll instead of libnstidy.so? PS: Any feedback on the new line numbering algorithm? |
From: Marc G. <mg...@sk...> - 2006-06-08 14:07:50
|
Hi Dominic, I think that what happens is that the detection of the OS does not work for freeBSD. There is a code in the tidyInstall.js that detect the OS. function tidyInstallGetPlatform() { var platform = navigator.platform.toLowerCase(); if (platform.indexOf('linux') != -1) { return 'linux'; } if (platform.indexOf('mac') != -1) { return 'mac'; } return 'windows'; } Based on this, the extension tries to detect the OS and file extension. I think that it will not work for FreeBSD ? Just I do not know how to do it in a better way. If you could correct this, I will include the fix in the next beta version that I finally succeeded to compile and run on Linux. Marc [LoN]Kamikaze wrote: > When trying to install the extension as root an alert pops up saying: > > Html Validator-002: Failed autoRegistering renamed platform specific nstidy.dll > > Does it mean it's looking for nstidy.dll instead of libnstidy.so? > > PS: Any feedback on the new line numbering algorithm? > > > ------------------------------------------------------------------------ > > _______________________________________________ > Htmlvalidator-help mailing list > Htm...@li... > https://lists.sourceforge.net/lists/listinfo/htmlvalidator-help > |
From: [LoN]Kamikaze <LoN...@gm...> - 2006-06-08 15:05:46
Attachments:
signature.asc
|
Marc Gueury wrote: > Hi Dominic, >=20 > I think that what happens is that the detection of the OS does not work= > for freeBSD. > There is a code in the tidyInstall.js that detect the OS. >=20 > function tidyInstallGetPlatform() > { > var platform =3D navigator.platform.toLowerCase(); > if (platform.indexOf('linux') !=3D -1) > { > return 'linux'; > } > if (platform.indexOf('mac') !=3D -1) > { > return 'mac'; > } > return 'windows'; > } >=20 > Based on this, the extension tries to detect the OS and file extension.= > I think that it will not work for FreeBSD ? > Just I do not know how to do it in a better way. If you could correct > this, I will include the fix in the next beta version > that I finally succeeded to compile and run on Linux. >=20 > Marc >=20 It should be enought to add: if (platform.indexOf('bsd') !=3D -1) { return 'bsd'; } Your tidyInstallGetPlatformName() falls back to lib{name}.so so that it r= eturns the right file name. Adding this I now get the following error message: Html Validator-002: Failed autoRegistering renamed platform specific libn= stidy.so Upon the next start of Firefox I get the usual library not found page. Th= e JavaScript console holds the following messages: Warning: reference to undefined property Components.classes['@mozilla.org= /tidy;1'] Source File: chrome://tidy/content/tidyUtil.js Line: 119 Error: Components.classes['@mozilla.org/tidy;1'] has no properties Source File: chrome://tidy/content/tidyUtil.js Line: 119 Warning: anonymous function does not always return a value Source File: chrome://dta/content/dta/addingFunctions.js Line: 90, Column: 1 Source Code: }, Error: oTidyUtil.tidy has no properties Source File: chrome://tidy/content/tidyBrowser.js Line: 1194 Error: oTidyUtil.tidy has no properties Source File: chrome://tidy/content/tidyBrowser.js Line: 1194 Error: oTidyUtil.tidy has no properties Source File: chrome://tidy/content/tidyBrowser.js Line: 188 Error: oTidyUtil.tidy has no properties Source File: chrome://tidy/content/tidyBrowser.js Line: 1194 Error: oTidyUtil.tidy has no properties Source File: chrome://tidy/content/tidyBrowser.js Line: 1194 Error: oTidyUtil.tidy has no properties Source File: chrome://tidy/content/tidyBrowser.js Line: 1194 |
From: Marc G. <mg...@sk...> - 2006-06-08 16:17:46
|
Hi Dominic, I added the 'BSD' check and reuploaded a new version of the source here: (08 june 2006) http://users.skynet.be/mgueury/mozilla/preview_080.html It contains several change to make the extension work better on Unix system. > Like not trying to load the .so library since it does not exist anyway. > And do not try to write in the firefox program directory Currently, it works fine on Linux, I still need to update the web site to put the 1rst linux version there. and some explanations of the things I would like to get fixed before to make a prod version of the 0.8x. Marc [LoN]Kamikaze wrote: > Marc Gueury wrote: > >> Hi Dominic, >> >> I think that what happens is that the detection of the OS does not work >> for freeBSD. >> There is a code in the tidyInstall.js that detect the OS. >> >> function tidyInstallGetPlatform() >> { >> var platform = navigator.platform.toLowerCase(); >> if (platform.indexOf('linux') != -1) >> { >> return 'linux'; >> } >> if (platform.indexOf('mac') != -1) >> { >> return 'mac'; >> } >> return 'windows'; >> } >> >> Based on this, the extension tries to detect the OS and file extension. >> I think that it will not work for FreeBSD ? >> Just I do not know how to do it in a better way. If you could correct >> this, I will include the fix in the next beta version >> that I finally succeeded to compile and run on Linux. >> >> Marc >> >> > > It should be enought to add: > > if (platform.indexOf('bsd') != -1) > { > return 'bsd'; > } > > > Your tidyInstallGetPlatformName() falls back to lib{name}.so so that it returns the right file name. > > Adding this I now get the following error message: > > Html Validator-002: Failed autoRegistering renamed platform specific libnstidy.so > > > Upon the next start of Firefox I get the usual library not found page. The JavaScript console holds the following messages: > > Warning: reference to undefined property Components.classes['@mozilla.org/tidy;1'] > Source File: chrome://tidy/content/tidyUtil.js > Line: 119 > > Error: Components.classes['@mozilla.org/tidy;1'] has no properties > Source File: chrome://tidy/content/tidyUtil.js > Line: 119 > > Warning: anonymous function does not always return a value > Source File: chrome://dta/content/dta/addingFunctions.js > Line: 90, Column: 1 > Source Code: > }, > > Error: oTidyUtil.tidy has no properties > Source File: chrome://tidy/content/tidyBrowser.js > Line: 1194 > > Error: oTidyUtil.tidy has no properties > Source File: chrome://tidy/content/tidyBrowser.js > Line: 1194 > > Error: oTidyUtil.tidy has no properties > Source File: chrome://tidy/content/tidyBrowser.js > Line: 188 > > Error: oTidyUtil.tidy has no properties > Source File: chrome://tidy/content/tidyBrowser.js > Line: 1194 > > Error: oTidyUtil.tidy has no properties > Source File: chrome://tidy/content/tidyBrowser.js > Line: 1194 > > Error: oTidyUtil.tidy has no properties > Source File: chrome://tidy/content/tidyBrowser.js > Line: 1194 > > > ------------------------------------------------------------------------ > > _______________________________________________ > Htmlvalidator-help mailing list > Htm...@li... > https://lists.sourceforge.net/lists/listinfo/htmlvalidator-help > |
From: [LoN]Kamikaze <LoN...@gm...> - 2006-06-08 23:17:54
Attachments:
signature.asc
|
Marc Gueury wrote: > Hi Dominic, >=20 > I added the 'BSD' check and reuploaded a new version of the source here= : > (08 june 2006) >=20 > http://users.skynet.be/mgueury/mozilla/preview_080.html >=20 > It contains several change to make the extension work better on Unix sy= stem. >> Like not trying to load the .so library since it does not exist anyway= =2E >> And do not try to write in the firefox program directory >=20 > Currently, it works fine on Linux, I still need to update the web site > to put the 1rst linux version there. > and some explanations of the things I would like to get fixed before to= > make a prod version of the 0.8x. >=20 > Marc >=20 Thank you for those changes, especially for removing the evilness that re= quired moving the sgml stuff around. My primary problem remains, though. Firefox does not load "libnstidy.so".= It is not even mentioned in "compreg.dat". The only way I managed to mak= e Firefox acknowledge its existance was by copying it into the plugins fo= lder. Which obviously causes an error, since tidy is not a plugin. One thing I recognized, though is that "install.js" contains the line: const X_COM_FILE =3D "libnstidy.so"; whereas "install.rdf" does not mention the library. Is there a way to exp= licitly declare the existence of a library to Firefox? I have found no me= ntion of that in various Firefox-extension tutorials. |
From: [LoN]Kamikaze <LoN...@gm...> - 2006-06-09 10:59:15
Attachments:
signature.asc
|
I rebuilt Firefox with debugging and it looks like this is causing my pro= blems: nsNativeComponentLoader: SelfRegisterDll(libnstidy.so) Load FAILED with e= rror: /home/kamikaze/.mozilla/firefox/eze6qb8b.default/extensions/{3b56bc= c7-54e5-44a2-9b44-66c3ef58c13e}/components/libnstidy.so: Undefined symbol= "__cxa_get_exception_ptr" |
From: [LoN]Kamikaze <LoN...@gm...> - 2006-06-09 16:14:46
Attachments:
signature.asc
|
[LoN]Kamikaze wrote: > I rebuilt Firefox with debugging and it looks like this is causing my p= roblems: >=20 > nsNativeComponentLoader: SelfRegisterDll(libnstidy.so) Load FAILED with= error: /home/kamikaze/.mozilla/firefox/eze6qb8b.default/extensions/{3b56= bcc7-54e5-44a2-9b44-66c3ef58c13e}/components/libnstidy.so: Undefined symb= ol "__cxa_get_exception_ptr" >=20 >=20 I was able to resolve this by using another version of gcc. However, now = Firefox segfaults when the extension is installed. This is the gdb output= : Starting program: /usr/X11R6/lib/firefox/firefox-bin=20 warning: Unable to get location for thread creation breakpoint: generic e= rror [New LWP 101142] [New Thread 0x8079000 (LWP 101142)] Type Manifest File: /home/kamikaze/.mozilla/firefox/eze6qb8b.default/xpti= =2Edat *** Registering Apprunner components (all right -- a generic module!) nsNativeComponentLoader: autoregistering begins. nsNativeComponentLoader: autoregistering succeeded nsNativeComponentLoader: registering deferred (0) pldhash: for the table at address 0x8131820, the given entrySize of 44 pr= obably favors chaining over double hashing. WARNING: NS_ENSURE_TRUE(NS_SUCCEEDED(rv)) failed, file nsChromeRegistry.c= pp, line 1245 GFX: dpi=3D81 t2p=3D0.0555556 p2t=3D18 depth=3D16 ++WEBSHELL =3D=3D 1 ++DOMWINDOW =3D=3D 1 WARNING: NS_ENSURE_TRUE(NS_SUCCEEDED(rv)) failed, file nsStringBundle.cpp= , line 273 ++DOMWINDOW =3D=3D 2 ++WEBSHELL =3D=3D 2 ++DOMWINDOW =3D=3D 3 ++DOMWINDOW =3D=3D 4 Note: styleverifytree is disabled Note: frameverifytree is disabled Note: verifyreflow is disabled Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x8079000 (LWP 101125)] 0x2a6f46f8 in nsTidyImpl_GetInterfacesHelper (count=3D0x0, array=3D0x0) at nsTidy.cpp:94 94=09 (gdb) bt #0 0x2a6f46f8 in nsTidyImpl_GetInterfacesHelper (count=3D0x0, array=3D0x= 0) at nsTidy.cpp:94 #1 0x281c79ed in nsGenericFactory::GetInterfaces (this=3D0x8790100,=20 countp=3D0xbfbfd300, array=3D0x2a83a4ec) at nsGenericFactory.cpp:99 #2 0x28d4877c in XPCNativeSet::GetNewOrUsed (ccx=3D@0xbfbfd638,=20 classInfo=3D0x8790104) at xpcwrappednativeinfo.cpp:604 #3 0x28d50891 in XPCWrappedNativeProto::GetNewOrUsed (ccx=3D@0xbfbfd638,= =20 Scope=3D0x8422400, ClassInfo=3D0x8790104, ScriptableCreateInfo=3D0xbf= bfd538,=20 ForceNoSharing=3D0) at xpcprivate.h:3363 #4 0x28d3e03d in XPCWrappedNative::GetNewOrUsed (ccx=3D@0xbfbfd638,=20 Object=3D0x8790280, Scope=3D0x8422400, Interface=3D0x81e19a0,=20 resultWrapper=3D0xbfbfd5c4) at xpcprivate.h:3366 #5 0x28d1cb58 in XPCConvert::NativeInterface2JSObject (ccx=3D@0xbfbfd638= ,=20 dest=3D0xbfbfd748, src=3D0x8790280, iid=3D0xbfbfd798, scope=3D0x84224= 00,=20 allowNativeWrapper=3D0, pErr=3D0xbfbfd634) at xpcprivate.h:3362 #6 0x28cfe09b in nsXPConnect::WrapNative (this=3D0x80ed7c0,=20 aJSContext=3D0x8425400, aScope=3D0xbfbfd638, aCOMObj=3D0x8790280, aII= D=3D@0x0,=20 _retval=3D0xbfbfd748) at nsXPConnect.cpp:586 #7 0x28d26c61 in nsJSCID::CreateInstance (this=3D0x86ee100, _retval=3D0x= bfbfda80) at nsCOMPtr.h:1149 #8 0x2826c72d in XPTC_InvokeByIndex (that=3D0x86ee100, methodIndex=3D10,= =20 paramCount=3D713270508, params=3D0xbfbfda80) at xptcinvoke_unixish_x8= 6.cpp:179 #9 0x28d4336e in XPCWrappedNative::CallMethod (ccx=3D@0xbfbfdb60,=20 mode=3DXPCWrappedNative::CALL_METHOD) at xpcwrappednative.cpp:2151 #10 0x28d4eb5a in XPC_WN_CallMethod (cx=3D0x8425400, obj=3D0x0, argc=3D0,= =20 argv=3D0xbfbfdb60, vp=3D0x0) at xpcwrappednativejsops.cpp:1445 #11 0x280e730a in js_Invoke (cx=3D0x8425400, argc=3D0, flags=3D0) at jsin= terp.c:1188 #12 0x280f40cd in js_Interpret (cx=3D0x8425400, pc=3D0x870620f ":",=20 result=3D0xbfbfdfbc) at jsinterp.c:3583 #13 0x280e7381 in js_Invoke (cx=3D0x8425400, argc=3D1, flags=3D1) at jsin= terp.c:1208 #14 0x280efc7d in js_Interpret (cx=3D0x8425400, pc=3D0x86ec53a "#",=20 result=3D0xbfbfe2f0) at jsinterp.c:3128 #15 0x280e7cd8 in js_Execute (cx=3D0x8425400, chain=3D0x837b690, script=3D= 0x879d000,=20 down=3D0x0, flags=3D0, result=3D0x0) at jsinterp.c:1434 #16 0x280b737a in JS_ExecuteScript (cx=3D0x8425400, obj=3D0x0, script=3D0= x0,=20 rval=3D0x0) at jsapi.c:4012 #17 0x2a1f3532 in nsJSContext::ExecuteScript (this=3D0x84223c0,=20 aScriptObject=3D0x0, aScopeObject=3D0x837b690, aRetValue=3D0x0, aIsUn= defined=3D0x0) at nsJSEnvironment.cpp:1216 #18 0x2a1c6402 in nsXULDocument::ExecuteScript (this=3D0x837b690,=20 aScriptObject=3D0x8744118) at nsXULDocument.cpp:3558 #19 0x2a1c609e in nsXULDocument::OnStreamComplete (this=3D0x840d000,=20 aLoader=3D0x873e700, context=3D0x0, aStatus=3D0, stringLen=3D14214784= 0,=20 string=3D0x8790100 "=C2=A8=C3=B1((=C3=AC=C3=B1((\002") at nsXULDocume= nt.cpp:3452 #20 0x28e677be in nsStreamLoader::OnStopRequest (this=3D0x873e700, reques= t=3D0x0,=20 ctxt=3D0x0, aStatus=3D142147840) at nsCOMPtr.h:1149 #21 0x290b8ee9 in nsJARChannel::OnStopRequest (this=3D0x86e0000, req=3D0x= 877eb00,=20 ctx=3D0x0, status=3D0) at nsCOMPtr.h:1149 #22 0x28e36e29 in nsInputStreamPump::OnStateStop (this=3D0x877eb00) at nsCOMPtr.h:1149 #23 0x28e367e7 in nsInputStreamPump::OnInputStreamReady (this=3D0x877eb00= ,=20 stream=3D0x874ae0c) at nsInputStreamPump.cpp:343 #24 0x28217d70 in nsInputStreamReadyEvent::EventHandler (plevent=3D0x0) at nsStreamUtils.cpp:119 #25 0x2823e37b in PL_HandleEvent (self=3D0x873e8c4) at plevent.c:688 #26 0x2823e238 in PL_ProcessPendingEvents (self=3D0x80e5040) at plevent.c= :623 #27 0x2824174b in nsEventQueueImpl::ProcessPendingEvents (this=3D0x80e500= 0) at nsEventQueue.cpp:417 #28 0x290f4625 in event_processor_callback (source=3D0x8386ac0,=20 condition=3DG_IO_IN, data=3D0x8790100) at nsAppShell.cpp:67 #29 0x2890e17f in g_io_channel_unix_get_fd () from /usr/local/lib/libglib-2.0.so.0 #30 0x288e5a8c in g_main_context_dispatch () from /usr/local/lib/libglib-2.0.so.0 #31 0x288e8ddf in g_main_context_check () from /usr/local/lib/libglib-2.0= =2Eso.0 #32 0x288e91a7 in g_main_loop_run () from /usr/local/lib/libglib-2.0.so.0= #33 0x2841e251 in gtk_main () from /usr/X11R6/lib/libgtk-x11-2.0.so.0 #34 0x0842ff10 in ?? () #35 0x00000001 in ?? () #36 0x00000000 in ?? () |
From: [LoN]Kamikaze <LoN...@gm...> - 2006-06-09 23:11:23
Attachments:
signature.asc
|
I managed to link spvalid dynamically to an OpenSP-1.5 installation, but = the segfaults remain. I'm really out of ideas, about what to do here. If = I link dynamically and deinstall opensp afterwords firefox complains abou= t the lack of libosp.so.3 instead of segfaulting. This seems to indicate = that the problem is part of the opensp component. |
From: Marc G. <mg...@sk...> - 2006-06-11 22:02:02
|
Hi Dominic, I have no idea neither why it is crashing. Maybe some questions to help: 1) Just by curiosity, which Firefox do you use after the compilation of the extension ? The one in mozilla/dist/bin that you compiled or the one coming on your system ? Does one work better than the other ? 2) Did you succeed to compile Firefox 1.5 and does this mozilla/dist/bin version work ? 3) which FreeBSD version ? Which CPU ? hardware ? 4) does the tidy 0.76 version of FreeBSD cause the same issue for you (I am convinced that the problem is not related to libosp.so but more to a compiler version or a library problem). 5) Does ltrace and strace work FreeBSD, maybe try to run ltrace to see which C call is failing ? Marc [LoN]Kamikaze wrote: > I managed to link spvalid dynamically to an OpenSP-1.5 installation, but the segfaults remain. I'm really out of ideas, about what to do here. If I link dynamically and deinstall opensp afterwords firefox complains about the lack of libosp.so.3 instead of segfaulting. This seems to indicate that the problem is part of the opensp component. > > > ------------------------------------------------------------------------ > > _______________________________________________ > Htmlvalidator-help mailing list > Htm...@li... > https://lists.sourceforge.net/lists/listinfo/htmlvalidator-help > |
From: [LoN]Kamikaze <LoN...@gm...> - 2006-06-11 22:35:53
Attachments:
signature.asc
|
Marc Gueury wrote: > Hi Dominic, >=20 > I have no idea neither why it is crashing. Maybe some questions to help= : >=20 > 1) Just by curiosity, which Firefox do you use after the compilation > of the extension ? The one in mozilla/dist/bin that you compiled > or the one coming on your system ? Does one work better than the othe= r ? I am using 1.5.0.4, both work the same. > 2) Did you succeed to compile Firefox 1.5 and does this mozilla/dist/bi= n > version > work ? Yes I succeeded and it works if I remove Tidy. > 3) which FreeBSD version ? Which CPU ? hardware ? FreeBSD 6.1-STABLE, Pentium-M 1300, Thinkpad R40 2722-5MG. > 4) does the tidy 0.76 version of FreeBSD cause the same issue for you > (I am convinced that the problem is not related to libosp.so but more > to a compiler version or a library problem). Tidy 0.76 also segfaults. So it's not an OSP issue. > 5) Does ltrace and strace work FreeBSD, maybe try to run ltrace to see > which C call is failing ? =20 I just installed ltrace and strace on my system. I will let you know what= they tell me. |
From: [LoN]Kamikaze <LoN...@gm...> - 2006-08-13 22:08:12
Attachments:
signature.asc
|
Marc Gueury wrote: > Hi Dominic, >=20 > There is a long time since we search about the FreeBSD issue to use the= > HTML Validator. An problem with the same stack trace and symptoms than > you on Debian Linux make me think there is maybe a simple solution. > Please read this: >=20 > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=3D378667 >=20 > Marc I will give that a try, thank you. If that works, I will patch tidy to use a dynamic OpenSP library (that only requires a change in the Makefile= ). There's also a tidy-lib in the FreeBSD Ports, but that will need more fiddling, I think. |
From: [LoN]Kamikaze <LoN...@gm...> - 2006-08-13 23:07:53
Attachments:
signature.asc
|
[LoN]Kamikaze wrote: > Marc Gueury wrote: >> Hi Dominic, >> >> There is a long time since we search about the FreeBSD issue to use th= e >> HTML Validator. An problem with the same stack trace and symptoms than= >> you on Debian Linux make me think there is maybe a simple solution. >> Please read this: >> >> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=3D378667 >> >> Marc >=20 > I will give that a try, thank you. If that works, I will patch tidy to > use a dynamic OpenSP library (that only requires a change in the Makefi= le). >=20 > There's also a tidy-lib in the FreeBSD Ports, but that will need more > fiddling, I think. >=20 Setting LD_PRELOAD does not help at all. I am at a total loss. |
From: Marc G. <mg...@sk...> - 2006-08-13 21:31:59
|
Hi Dominic, There is a long time since we search about the FreeBSD issue to use the HTML Validator. An problem with the same stack trace and symptoms than you on Debian Linux make me think there is maybe a simple solution. Please read this: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=378667 Marc [LoN]Kamikaze wrote: > Marc Gueury wrote: > >> Hi Dominic, >> >> I have no idea neither why it is crashing. Maybe some questions to help: >> >> 1) Just by curiosity, which Firefox do you use after the compilation >> of the extension ? The one in mozilla/dist/bin that you compiled >> or the one coming on your system ? Does one work better than the other ? >> > > I am using 1.5.0.4, both work the same. > > >> 2) Did you succeed to compile Firefox 1.5 and does this mozilla/dist/bin >> version >> work ? >> > > Yes I succeeded and it works if I remove Tidy. > > >> 3) which FreeBSD version ? Which CPU ? hardware ? >> > > FreeBSD 6.1-STABLE, Pentium-M 1300, Thinkpad R40 2722-5MG. > > >> 4) does the tidy 0.76 version of FreeBSD cause the same issue for you >> (I am convinced that the problem is not related to libosp.so but more >> to a compiler version or a library problem). >> > > Tidy 0.76 also segfaults. So it's not an OSP issue. > > >> 5) Does ltrace and strace work FreeBSD, maybe try to run ltrace to see >> which C call is failing ? >> > > I just installed ltrace and strace on my system. I will let you know what they tell me. > > > ------------------------------------------------------------------------ > > _______________________________________________ > Htmlvalidator-help mailing list > Htm...@li... > https://lists.sourceforge.net/lists/listinfo/htmlvalidator-help > |