Created attachment 104393 [details] error After update from 4.2.5.2 to 4.3.0.4 Application Error soffice.bin, version 4.3.0.4, module sal3.dll, version 4.3.0.4, address 0x000268d5. LО doesn't start ( Ошибка приложения soffice.bin, версия 4.3.0.4, модуль sal3.dll, версия 4.3.0.4, адрес 0x000268d5.
Something like this problem: http://go.mail-archive.com/5fqF79Qe7kgmiSzrxL7OYteZQy4= ? Set status to NEEDINFO. Please set it back to UNCONFIRMED once you have provided requested informations. Thank you for your understanding. Best regards. JBF
Hi, JBF! Yes, I think it's the same error. problems are present: LibreOffice 4.3.1.1 Can you fix it? Thank you! Computer Operating System Microsoft Windows XP Professional OS Service Pack Service Pack 3 Internet Explorer 8.0.6001.18702 Motherboard CPU Type AMD Athlon XP, 1746 MHz 2100+ Motherboard Name Abit NF7(-S) v2.0 (5 PCI, 1 AGP, 3 DIMM, Audio, LAN) Motherboard Chipset nVIDIA nForce2 Ultra 400 System Memory 2304 MB BIOS Type Award (11/22/04) CPUID Properties CPUID Manufacturer AuthenticAMD CPUID CPU Name AMD Sempron(tm) 2500+ CPUID Revision 0681h Extended CPUID Revision 0781h Platform ID CBh (Socket A) IA CPU Serial Number Unknown Instruction Set AMD 3DNow! Supported AMD 3DNow! Professional Supported AMD Enhanced 3DNow! Supported AMD Extended MMX Supported AMD64 Not Supported Cyrix Extended MMX Not Supported IA-64 Not Supported IA MMX Supported IA SSE Supported IA SSE 2 Not Supported IA SSE 3 Not Supported CLFLUSH Instruction Not Supported CMPXCHG8B Instruction Supported Conditional Move Instruction Supported MONITOR / MWAIT Instruction Not Supported CPUID Features 36-bit Page Size Extension Supported Automatic Clock Control Not Supported CPL Qualified Debug Store Not Supported Debug Trace Store Not Supported Debugging Extension Supported Enhanced SpeedStep Technology Not Supported Fast Save & Restore Supported Fast System Call Supported Hyper-Threading Technology Not Supported L1 Context ID Not Supported Local APIC On Chip Supported LongRun Not Supported LongRun Table Interface Not Supported Machine-Check Architecture Supported Machine-Check Exception Supported Memory Type Range Registers Supported Model Specific Registers Supported Page Attribute Table Supported Page Global Extension Supported Page Size Extension Supported Pending Break Event Not Supported Physical Address Extension Supported Processor Duty Cycle Control Not Supported Processor Serial Number Not Supported Self-Snoop Not Supported Thermal Monitor Not Supported Thermal Monitor 2 Not Supported Time Stamp Counter Supported Virtual Mode Extension Supported Enhanced Power Management Frequency ID Control Not Supported Software Thermal Control Not Supported Temperature Sensing Diode Supported Thermal Monitoring Not Supported Thermal Trip Not Supported Voltage ID Control Not Supported
Same problem with update from 4.2.5.2 to 4.3.0.4 Update to 4.3.0.4 Ok the programm does not run using XP SP3 processeur AMD Athlon XP1800 SSE I see that in 4.3 Release Notes that compilation is made of 4.3 on Windows Server 2012R2 with Visual Studio 2012 and it seems that the default build with VS 2012 is for SSE2 ? Processor 1 ID = 0 Number of cores 1 (max 1) Number of threads 1 (max 1) Name AMD Athlon XP Codename Thoroughbred Specification AMD Athlon(tm) Package Socket A (462) CPUID 6.8.0 Extended CPUID 7.8 Core Stepping A0 Technology 0.13 um Core Speed 1152.9 MHz Multiplier x FSB 11.5 x 100.3 MHz Rated Bus speed 200.5 MHz Instructions sets MMX (+), 3DNow! (+), SSE Mainboard Model MS-6380E (0x000001E3 - 0x00BE1AEA) Memory 2x 512 MBytes
So confirmed. See also the discussion on the developer list: http://nabble.documentfoundation.org/v4-3-0-4-not-working-on-SSE1-CPUs-anymore-tp4118212.html Best regards. JBF
potential patch for this is here, don't have hardware to test it though and also nobody knows what sort of performance impact will result from not using SSE2: https://gerrit.libreoffice.org/#/c/10838/
Most of our performance problems are not at the micro level wrt. SSE2 optimization but crazy big algorithmic issues that SSE2 won't help for ;-) so I'd say - go for it ...
Hi, Is there a link to test a build with this new patch SSE ? Thanks
Michael Stahl committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=8bd6bf93b7711a7ac7c5cbd7c3bb980481570ebd fdo#82430: configure: MSVC build: avoid using SSE2 instructions The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Michael Stahl committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=36ce22f41a754fa405804434899a08a23081e721 fdo#82430: MSVC 2010 refuses both -clr and -arch in the same command The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Michael Stahl committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=a18ff3d5c75c2b468c48bd19439dee0689d24d67 fdo#82430: gcc-wrappers: don't stop if REAL_CXX_FLAGS is empty The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
do the latest daily builds from here run on the old CPUs? http://dev-builds.libreoffice.org/daily/master/Win-x86@42/
Hi, >do the latest daily builds from here run on the old CPUs? download the libo-master~2014-08-19_05.34.52_LibreOfficeDev_4.4.0.0.alpha0_Win_x86.msi install on XP SP3 processeur AMD Athlon XP1800 SSE Install OK Run the program, the scren "LibreOffice" appears then a warning box "The application cannot be stared..." see http://postimg.org/image/llm4q0od5/ Try Repair warning box idem Try Modify uncheck Windows Explorer Extension ans Extensions warning box idem
Michael Stahl committed a patch related to this issue. It has been pushed to "libreoffice-4-3": http://cgit.freedesktop.org/libreoffice/core/commit/?id=f83c7358964cf9f265f21e7dfa72e82a26923b3e&h=libreoffice-4-3 fdo#82430: configure: MSVC build: avoid using SSE2 instructions It will be available in LibreOffice 4.3.2. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Michael Stahl committed a patch related to this issue. It has been pushed to "libreoffice-4-3-1": http://cgit.freedesktop.org/libreoffice/core/commit/?id=31c1ceaf50f911b5ea22a1f7770b961343483230&h=libreoffice-4-3-1 fdo#82430: configure: MSVC build: avoid using SSE2 instructions It will be available already in LibreOffice 4.3.1. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Created attachment 105285 [details] message after run libo-43~2014-08-25_08.45.25_LibreOfficeDev_4.3.2.0.0 libo-43~2014-08-25_08.45.25_LibreOfficeDev_4.3.2.0.0_Win_x86.msi test install and run from XP SP3 processeur AMD Athlon XP1800 SSE
>It will be available in LibreOffice 4.3.2. download the libo-43~2014-08-25_08.45.25_LibreOfficeDev_4.3.2.0.0_Win_x86.msi install on XP SP3 processeur AMD Athlon XP1800 SSE Install OK Run the program, and an alert dialog "libreoffice a cessé de fonctionner" see file 469f_appcompat.txt
(In reply to comment #15) > libo-43~2014-08-25_08.45.25_LibreOfficeDev_4.3.2.0.0_Win_x86.msi > test install and run from XP SP3 processeur AMD Athlon XP1800 SSE thanks for testing, but i think this build is too old and does not contain the commit yet; i hope today's build (08-26) should have it.
>i hope today's build (08-26) should have it I download the libo-43~2014-08-26_10.56.01_LibreOfficeDev_4.3.2.0.0_Win_x86.msi install on XP SP3 processeur AMD Athlon XP1800 SSE Install OK Run the program, the screen "LibreOffice" appears, but stop with a warning box LibreOfficeDec 4.3 - Fatal Error "The application cannot be started Extension Manager: exception during enabledExtension" It is idem the Comment 12 Try Repair : warning box idem Try Modify and uncheck all the components : warning box idem Now I cannot test before 16 September, I tomorrow goes on vacation ;)
"Extension Manager: exception during enabledExtension" that is unfortunate, something is caught in desktop/source/deployment/manager/dp_extensionmanager.cxx which is not a regular UNO exception. maybe some of our bundled external libraries are not compiled with the -arch:SSE?
Hi, I download the http://dev-builds.libreoffice.org/pre-releases/win/x86/LibreOffice_4.3.2.1_Win_x86.msi install on XP SP3 processeur AMD Athlon XP1800 SSE Install OK Run the program, the screen "LibreOffice" appears, but stop with a warning box Fatal Error "The application cannot be started Extension Manager: exception during enabledExtension" It is idem the Comment 12 and Comment 18 So Go back tho the version 4.2.6.2 This bug is NOT solved ;)
hi2all Yes. LibreOffice_4.3.1_Win_x86.msi Run the program, the screen "LibreOffice" appears, but stop with a warning box Fatal Error "The application cannot be started Extension Manager: exception during enabledExtension" This bug isn't solved.
I download the lastest daily/libreoffice-4-3/Win-x86@42/current libo-43~2014-09-22_07.54.32_LibreOfficeDev_4.3.3.0.0_Win_x86.msi install on XP SP3 processeur AMD Athlon XP1800 SSE Install OK Run the program, the screen "LibreOffice" appears, but stop with a warning box Fatal Error "The application cannot be started Extension Manager: exception during enabledExtension" It is idem the Comment 12 and Comment 18 and 20 is there a solution to have a working version 4.3 on XP SSE ?
do you think Bug 84316 is a duplicate of this one?
sorry was occupied with other tasks for the last weeks... there's a script that can determine instructions used: http://dev.gentoo.org/~solar/misc/analyse-x86.pl based on that i've just tweaked the most important external libraries, let's hope this will help to at least start soffice successfully: Disassembling ./URE/bin/libxml2.dll, please wait... i486: 0 i586: 0 ppro: 0 mmx: 570 sse: 204 sse2: 690 ./URE/bin/libxml2.dll will run on Pentium IV (pentium4) or higher processor. Checking vendor_id string...GenuineIntel Disassembling ./URE/bin/icuuc53.dll, please wait... i486: 0 i586: 0 ppro: 0 mmx: 1 sse: 14 sse2: 173 ./URE/bin/icuuc53.dll will run on Pentium IV (pentium4) or higher processor. Disassembling ./program/libxslt.dll, please wait... i486: 0 i586: 4 ppro: 0 mmx: 55 sse: 61 sse2: 258 This binary was found to contain the cpuid instruction. It may be able to conditionally execute instructions if they are supported on the host (i586+). ./program/libxslt.dll will run on Pentium IV (pentium4) or higher processor. Checking vendor_id string...GenuineIntel Disassembling ./program/libexslt.dll, please wait... i486: 0 i586: 0 ppro: 0 mmx: 111 sse: 24 sse2: 482 ./program/libexslt.dll will run on Pentium IV (pentium4) or higher processor. Checking vendor_id string...GenuineIntel Disassembling ./program/libxmlsec.dll, please wait... i486: 0 i586: 0 ppro: 0 mmx: 71 sse: 26 sse2: 4 ./program/libxmlsec.dll will run on Pentium IV (pentium4) or higher processor. Disassembling ./program/icuin53.dll, please wait... i486: 0 i586: 4 ppro: 2 mmx: 1 sse: 2 sse2: 3676 This binary was found to contain the cpuid instruction. It may be able to conditionally execute instructions if they are supported on the host (i586+). ./program/icuin53.dll will run on Pentium IV (pentium4) or higher processor. Disassembling ./program/nssckbi.dll, please wait... i486: 0 i586: 0 ppro: 0 mmx: 52 sse: 1 sse2: 14 ./program/nssckbi.dll will run on Pentium IV (pentium4) or higher processor. Disassembling ./program/nspr4.dll, please wait... i486: 3 i586: 0 ppro: 2 mmx: 206 sse: 38 sse2: 274 ./program/nspr4.dll will run on Pentium IV (pentium4) or higher processor. Disassembling ./program/ssl3.dll, please wait... i486: 0 i586: 0 ppro: 0 mmx: 376 sse: 13 sse2: 7 ./program/ssl3.dll will run on Pentium IV (pentium4) or higher processor. Disassembling ./program/sqlite3.dll, please wait... i486: 0 i586: 4 ppro: 1 mmx: 650 sse: 159 sse2: 682 This binary was found to contain the cpuid instruction. It may be able to conditionally execute instructions if they are supported on the host (i586+). ./program/sqlite3.dll will run on Pentium IV (pentium4) or higher processor. Disassembling ./program/python3.dll, please wait... i486: 0 i586: 4 ppro: 0 mmx: 10 sse: 0 sse2: 60 This binary was found to contain the cpuid instruction. It may be able to conditionally execute instructions if they are supported on the host (i586+). ./program/python3.dll will run on Pentium IV (pentium4) or higher processor. Disassembling ./program/lcms2.dll, please wait... i486: 0 i586: 4 ppro: 0 mmx: 635 sse: 1163 sse2: 5897 This binary was found to contain the cpuid instruction. It may be able to conditionally execute instructions if they are supported on the host (i586+). ./program/lcms2.dll will run on Pentium IV (pentium4) or higher processor. Checking vendor_id string...GenuineIntel Disassembling ./program/avmediaogl.dll, please wait... i486: 0 i586: 0 ppro: 6 mmx: 706 sse: 4327 sse2: 554 ./program/avmediaogl.dll will run on Pentium IV (pentium4) or higher processor.
installsets with the latest patch https://gerrit.libreoffice.org/#/c/11769/ have been provided here: http://dev-builds.libreoffice.org/tmp/fdo_82430/ please give them a quick test.
*** Bug 84316 has been marked as a duplicate of this bug. ***
(In reply to Christian Lohmaier from comment #25) > installsets with the latest patch https://gerrit.libreoffice.org/#/c/11769/ > have been provided here: http://dev-builds.libreoffice.org/tmp/fdo_82430/ Thanks Download LibreOfficeDev_4.3.3.0.0_Win_x86_en-US_de_ar_ja_qtz.msi Install on XP SP3 processeur AMD Athlon XP1800 SSE Install OK Run the program,start ok see http://postimg.org/image/c0ptn0wi7/ test some files ods, odt, rtf loading is Ok but sometimes LibreOffice crash just by clicking on the page see message http://postimg.org/image/dikk6x1hn/ Test export PDF same message
Could be possible that bug 83277 is a duplicate of this one ? Best regards. JBF
Hi, is the version LibreOffice_4.3.3.2_Win_x86.msi 23-Oct-2014 21:01 from http://dev-builds.libreoffice.org/pre-releases/win/x86/ fixes this bug?
Michael Stahl committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=c79d2dbe3a75e3376a5997444b6a3a5f947da107 fdo#82430: MSVC build: disable a few more cases of SSE2 in externals It will be available in 4.4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
did a little more investigation... found that in the last patch i've unfortunately only adjusted ICU's CFLAGS and not CXXFLAGS so the majority of ICU code was still compiled for SSE2; this most likely causes lots of crashes in spell checking, word counting etc. this should be fixed on master now. there are still 3 kinds of externals that the analyze-x86.pl script complains about: 1. firebird and 2. coinmp are not so important because they are new experimental features anyway; also we patch horrible things into coinmp's build system which needs cleaning up... the 3. is basically "everything built with gcc-wrapper", i.e. mostly the more obscure import filters, and i have no idea how to fix that, evidently the -arch:SSE is passed to cl.exe; i *hope* it's a false positive from the analyze-x86.pl. * * * there is apparently a tool from Intel called "Software Development Emulator" that can emulate various x86 CPUs in a fast way with dynamic translation, so i thought it should be possible to use it for testing... unfortunately it turns out to be severely under-documented and i couldn't get it to run soffice, it was always aborting on a SSE2 instruction in MSVCR110.DLL which indicates that the CPUID still claims SSE2 support, despite setting it to Pentium3 (the -cpuid-in parameter is undocumented). /home/sde-external-7.8.0-2014-10-02-win/sde.exe -chip-check PENTIUM3 -cpuid-in c:/cygwin/home/ms/sde-external-7.8.0-2014-10-02-win/misc/cpuid/pentium3/cpuid.def -- instdir/program/soffice
Michael Stahl committed a patch related to this issue. It has been pushed to "libreoffice-4-3": http://cgit.freedesktop.org/libreoffice/core/commit/?id=0fae44a2c9c17d6e3325c7ca2d56d4f90ad539c9&h=libreoffice-4-3 fdo#82430: MSVC build: disable a few more cases of SSE2 in externals It will be available in 4.3.3. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
download the file libo-43~2014-11-09_07.56.38_LibreOfficeDev_4.3.4.0.0_Win_x86.msi on http://dev-builds.libreoffice.org/daily/libreoffice-4-3/Win-x86@42/ Install it Run it and LBO run well on XP SP3 processeur AMD Athlon XP1800 SSE Version: 4.3.4.0.0+ Build ID: ce807b9143baa7a686ab9b2a95899f2c1817a715 Well ... thanks and "Bravo to super" Michael Stahl ;) Now tests : - test open files ods, odt, doc, .ppt OK - test open files xls -> message Read-error. Internal import error Waiting for the results of other feedback
(In reply to JCE from comment #33) > download the file > libo-43~2014-11-09_07.56.38_LibreOfficeDev_4.3.4.0.0_Win_x86.msi > on http://dev-builds.libreoffice.org/daily/libreoffice-4-3/Win-x86@42/ > Install it > Run it and LBO run well on XP SP3 processeur AMD Athlon XP1800 SSE > Now tests : > - test open files ods, odt, doc, .ppt OK > - test open files xls -> message Read-error. Internal import error Test an open file spreadsheet: Predim4.ods an try to save it on Excel 97 .XLS message : Error saving the document Predim4: Read-Error. Internal import error.
(This is an automated message.) LibreOffice development currently prioritizes bugs with the so called MAB (most annoying bugs) -- as this bug has not run through that process (including writing a short rationale for this bug being a candidate and other who are watching the tracker bug silently approving that rationale etc.) its priority is set to high. Note this is effectively no change in the urgency assigned to this bug, as we are currently not making a difference between high and highest and severity is untouched. You can find out more about MABs and how the process works by contacting libreoffice qa on irc: http://webchat.freenode.net/?channels=libreoffice-qa The QA wiki page also gives you hints on how to get in contact with the team (if IRC fails you, your next best choice is the mailing list): https://wiki.documentfoundation.org/QA
download the file libo-44~2015-02-02_15.42.43_LibreOfficeDev_4.4.1.0.0_Win_x86.msi Install it : OK Run it and LBO run well on XP SP3 processeur AMD Athlon XP1800 SSE same problem as reported
*** Bug 88374 has been marked as a duplicate of this bug. ***
I've just spent the better part of a day trying to install LibreOffice on an old PC with a clean Windows 7 installation and I couldn't figure out why it wouldn't work. My other Athlon XP CPU is running v4.2 without issue. Anyway, Athlon XP 2800+ (no SSE2 of course), Windows 7 x86: Loading .XLS documents generates Read Error: Internal Import Error. Loading .PPS hangs the program. I'm sure there are other issues, but it was enough for me to uninstall LO. v4.2 is fine. This is a lightweight program, it should run on light weight systems. This bug was reported almost 6 months ago, can we please elevate it? Not running at all is a SERIOUS bug, not minor.
it's not obvious where the remaining problem is, none of the developers has such old hardware, and to have any hope of making progress we would need to locate the faulting instructions with a stack trace from windbg or VisualStudio. for windbg there are some instructions on this here, but i don't know if it stops automatically when an invalid instruction is executed or if that needs some special commands... https://wiki.documentfoundation.org/How_to_get_a_backtrace_with_WinDbg in Visual Studio there is a dialog in "Debug->Exceptions" with a lot of checkboxes about which should stop the execution, there is under "Win32 exceptions" the "0xc0000001d Illegal Instruction" which i guess should be enabled.
(In reply to Michael Stahl from comment #39) > it's not obvious where the remaining problem is, none of the developers has > such old hardware, and to have any hope of making progress we would need to > locate the faulting instructions with a stack trace from windbg or > VisualStudio. > > for windbg there are some instructions on this here, but i don't > know if it stops automatically when an invalid instruction is executed > or if that needs some special commands... > > https://wiki.documentfoundation.org/How_to_get_a_backtrace_with_WinDbg > > in Visual Studio there is a dialog in "Debug->Exceptions" with > a lot of checkboxes about which should stop the execution, > there is under "Win32 exceptions" the "0xc0000001d Illegal Instruction" > which i guess should be enabled. I will admit I know little about debugging. Typically in cases where SSE2 is the issue it is because the compiler was set to use that instruction set and a rebuild clearing that option can fix it. Sorry, but I was just a big put off that I spent a log of time bouncing against this bug and the tone of this thread makes it appear that fixing it is a low priority. It should be pointed out that System Requirements say: "Pentium-compatible PC (Pentium III, Athlon or more-recent system recommended);" The P3 was an non-SSE2 processor. I wonder how many people assume LO should work because of this, and then when it doesn't they just move on to a different program.
Maybe the real bug is in the System Requirements then?
(In reply to drmmedia from comment #38) > I've just spent the better part of a day trying to install LibreOffice on an > old PC with a clean Windows 7 installation and I couldn't figure out why it > wouldn't work. > My other Athlon XP CPU is running v4.2 without issue. > > Anyway, Athlon XP 2800+ (no SSE2 of course), Windows 7 x86: > Loading .XLS documents generates Read Error: Internal Import Error. > Loading .PPS hangs the program. > I'm sure there are other issues, but it was enough for me to uninstall LO. Tried today on a Pentium III processor. Could not reproduce hang with this .pps: http://www.reading.ac.uk/web/files/www_File_Library/example.pps Please attach a .PPS that hangs in your setup. Could reproduce .xls read error. Debugging will continue tomorrow.
Created attachment 114883 [details] Backtrace from WinXP, Libo 4.4.2.2, opening .xls file In Windbg I did Debug - Event filters - enabled Illegal instruction. Then it breaked out when opening the .xls threw an illegal instruction and I could analyze. Thanks to Michael Stahl for tips.
Michael Stahl committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=c4a267215c10647ff281e9e629faa0a8a201f4b4 tdf#82430: gcc-wrappers: command line *must* start with a space It will be available in 5.0.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
thanks Beluga, very useful stack trace, found the issue hopefully this should fix everything in 4.3 that isn't experimetal
Michael Stahl committed a patch related to this issue. It has been pushed to "libreoffice-4-4": http://cgit.freedesktop.org/libreoffice/core/commit/?id=0dac335acbcd382fff422e50058d8aa21646a364&h=libreoffice-4-4 tdf#82430: gcc-wrappers: command line *must* start with a space It will be available in 4.4.4. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Michael Stahl committed a patch related to this issue. It has been pushed to "libreoffice-4-3": http://cgit.freedesktop.org/libreoffice/core/commit/?id=2c3f7421d550b6ae8c6bfdcde709dfba57aa1d3a&h=libreoffice-4-3 tdf#82430: gcc-wrappers: command line *must* start with a space It will be available in 4.3.8. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Michael Stahl committed a patch related to this issue. It has been pushed to "libreoffice-4-3-7": http://cgit.freedesktop.org/libreoffice/core/commit/?id=b2c9319c508b340694017b1458ab5feb4d948585&h=libreoffice-4-3-7 tdf#82430: gcc-wrappers: command line *must* start with a space It will be available in 4.3.7. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Michael Stahl committed a patch related to this issue. It has been pushed to "libreoffice-4-4-3": http://cgit.freedesktop.org/libreoffice/core/commit/?id=eccbf9df2e08d78b96b01b891e801fc8ef9b1f88&h=libreoffice-4-4-3 tdf#82430: gcc-wrappers: command line *must* start with a space It will be available in 4.4.3. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Thanks to Michael Stahl download the file : libo-44~2015-04-21_07.17.34_LibreOfficeDev_4.4.4.0.0_Win_x86.msi Install it : OK Run it and LBO run well on XP SP3 processeur AMD Athlon XP1800 SSE Test open files *.xls, *.doc, *.ppt, odt, *.ods, *svg export as PDF is OK also I will test with version 4.3.8 libo-43~2015-04-21xxx (not available today) So now need to tests with other testers
No problem opening .xls with 4.3.8. 5.0 from 21.4. did not launch, first complained about missing MSVCR120.dll and after I added it, said "The procedure entry point InetNtopW could not be located in the dynamic link library WS2_32.dll". Version: 4.3.8.0.0+ Build ID: 589f7cbe3789224a8836f3cd4705f067be6789c5 TinderBox: Win-x86@42, Branch:libreoffice-4-3, Time: 2015-04-20_12:02:51
guess i should mention this: 5.0 will be built with a different toolchain, and if it's configured with the 8.x SDK it will flat out not run on Windows XP. if you have any issue with 5.0 releases please file a new bug, don't re-open this one.
(In reply to Michael Stahl from comment #52) > guess i should mention this: 5.0 will be built with a different toolchain, > and if it's configured with the 8.x SDK it will flat out not run on Windows > XP. > > if you have any issue with 5.0 releases please file a new bug, don't re-open > this one. Cloph told me my issue with 5.0 was because I used SI-GUI to do a parallel install :) I'm not at the machine so can't test, but here is a workaround: https://wiki.documentfoundation.org/Development/msvc-x86_64#MS_Visual_Studio_C.2B.2B_2013_runtimes I'll set this to VERIFIED.
download the file : 4.3.8 libo-43~2015-04-20_12.02.51_LibreOfficeDev_4.3.8.0.0_Win_x86.msi Install it : OK Run it and LBO run well on XP SP3 processeur AMD Athlon XP1800 SSE Test open files *.xls, *.doc, *.ppt, odt, *.ods, *svg export as PDF is OK also So now I need to test the last full version testing LibreOffice_4.3.x.x_Win_x86.msi
Test with LibreOffice_4.3.7.2_Win_x86.msi Version: 4.3.7.2 Build ID: 8a35821d8636a03b8bf4e15b48f59794652c68ba All is OK on XP SP3 processeur AMD Athlon XP1800 SSE
Test on version 4.5.5.2 All is OK non problem on XP SP3 processeur AMD Athlon XP1800 SSE So delete on https://wiki.documentfoundation.org/ReleaseNotes/4.4 "Platform compatibility / Windows Known Issue: there are problems with running the Windows release on older CPUs that lack support for SSE2 instructions.."
Has the performance drop been evaluated ? (with or without sse2 optimisation) Is there any stat on how many computers are still running with such an old processor ? IMHO, it's quite crazy to hardly work to improve LO perf (using OOpenCL, OpenGL, and other perf opt), and at the same time dropping SSE2 optimisation. Maybe someone can consider it's time to forget old processors. Why not starting with LO5 Branch ? Of course, if 20 % of LO users are using non sse2 processor, it's a good thing, but what if it is 0,0000 1 % ? Surely, this number decreases months after months
Created attachment 119977 [details] Crash after closing file 1
Created attachment 119978 [details] crash after closin file 2
Created attachment 119979 [details] crash after closing file 3
Created attachment 119980 [details] crash after closing file 4
Created attachment 119981 [details] crash after closin file 5
Created attachment 119982 [details] crash after closing file 6 all files are opened with a double-click and closed again without doing anything else.
old computers in use: PC1: Win98SE / IDT WinChip C6 200 MHz (MMX) PC2: XPproSP3 / Intel Celeron 600 MHz (MMX; SSE) PC3: Win98SE / AMD Athlon 800 MHz (MMX+; 3DNow!+) PC4: XPSP3 / Intel Pentioum III E (MMX; SSE) PC5: Win7 / AMD Athlon XP2600+ (MMX; 3DNow!+; SSE) PC6: XPproSP3 / AMD Athlon XP3000+ (MMX+; 3DNow!+; SSE)
Pleas, do not change the version number which is intended to reflect to oldest version in which the issue has been seen. Best regards. JBF
Sorry, I have some problems with the bugzilla comment system. Can anybody correct my mistakes and bring the messages to the proper location? Thank you !