Re: LyX crash when mathmod cdots is used
On 12-Apr-2001 John Levon wrote: perhaps you altered the size of InsetText at some point ? I don't know maybe! Surely I altered the size of Inset/UpdatableInset! I got gcc 2.96-80 working (me being silly), and the bug doesn't occur. So I think the patch below or similar should go into CVS. Is the 2.96-80 from rawhide? I would let Lars decide if he likes to integrate the patch. If he gives the OK I'll do it! Jrgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jrgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ If you mess with a thing long enough, it'll break. -- Schmidt
Re: LyX crash when mathmod cdots is used
On Fri, 13 Apr 2001, Juergen Vigna wrote: Is the 2.96-80 from rawhide? yes I would let Lars decide if he likes to integrate the patch. If he gives the OK I'll do it! sure john -- "To be fair i do look quite like a monkey ..." - Peter Reid
Re: LyX crash when mathmod cdots is used
On 12-Apr-2001 John Levon wrote: > perhaps you altered the size of InsetText at some point ? I don't know maybe! Surely I altered the size of Inset/UpdatableInset! > I got gcc 2.96-80 working (me being silly), and the bug doesn't > occur. So I think the patch below or similar should go into > CVS. Is the 2.96-80 from rawhide? I would let Lars decide if he likes to integrate the patch. If he gives the OK I'll do it! Jürgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jürgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ If you mess with a thing long enough, it'll break. -- Schmidt
Re: LyX crash when mathmod cdots is used
On Fri, 13 Apr 2001, Juergen Vigna wrote: > Is the 2.96-80 from rawhide? yes > I would let Lars decide if he likes to integrate the patch. If he gives > the OK I'll do it! sure john -- "To be fair i do look quite like a monkey ..." - Peter Reid
Re: LyX crash when mathmod cdots is used
On 11-Apr-2001 John Levon wrote: currently on Clone(), (when it works properly) we are resetting each cell's InsetText three times: once in Init, where they are constructed, once in operator=, where they shouldbe/are copied, and then again in Clone(). The second one causes the problem, so if we change the copy c-tor to not use operator= itself, then we can avoid the second part of this. A small optimisation of course ... Well you're right we do an assignment too much! I commented out the asignment in Clone() as it is already done in the new() call! Thanks for pointing me to this! I also tried to optimize the new() call for clone a bit. Jrgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jrgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Gnagloot, n.: A person who leaves all his ski passes on his jacket just to impress people. -- Rich Hall, "Sniglets"
Re: LyX crash when mathmod cdots is used
Well you're right we do an assignment too much! I commented out the asignment in Clone() as it is already done in the new() call! Thanks for pointing me to this! I also tried to optimize the new() call for clone a bit. Just my two cents: Clone() _always_ should look like SomeThing * SomeThing::Clone() const { return new SomeThing(*this); } and everything else should be handled by the copy constructor. Andre' -- Andr Pnitz . [EMAIL PROTECTED] C++-Programmierer gesucht ... Naeheres unter http://mathematik.htwm.de/job
Re: LyX crash when mathmod cdots is used
On 12-Apr-2001 Andre Poenitz wrote: Well you're right we do an assignment too much! I commented out the asignment in Clone() as it is already done in the new() call! Thanks for pointing me to this! I also tried to optimize the new() call for clone a bit. Just my two cents: Clone() _always_ should look like SomeThing * SomeThing::Clone() const { return new SomeThing(*this); } and everything else should be handled by the copy constructor. Well it does now look like that, thought I don't know what's the copy constructor (there was a warning from Lars about this cc but I asked him for explications and didn't understand them so I just left it asis :) Jrgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jrgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Mistakes are often the stepping stones to utter failure.
Re: LyX crash when mathmod cdots is used
Well it does now look like that, thought I don't know what's the copy constructor (there was a warning from Lars about this cc but I asked him for explications and didn't understand them so I just left it asis :) The copy constructor is something like Something::Something(Something const t) { // make sure we are an exact copy of t } If it is not defined by yourself, the compiler will provide one automatically by using the copy constructors of all Something::members. This is often exactly what is necessary, but it breaks with "naked owning pointers", since it copies the only _pointer_ itself and does not create a copy of the item pointed to. (The same holds true for operator=() btw) This necessity to manually handle copy construction and assignment (and destruction) when using "naked owning pointers" is one of the major reasons why we should used some kind of smart pointer (like boost::shared_pointer) as class members instead. Andre' -- Andr Pnitz . [EMAIL PROTECTED] C++-Programmierer gesucht ... Naeheres unter http://mathematik.htwm.de/job
Re: LyX crash when mathmod cdots is used
On Thu, 12 Apr 2001, Juergen Vigna wrote: in Clone() as it is already done in the new() call! Thanks for pointing me to this! I also tried to optimize the new() call for clone a bit. great. however please don't commit this until the optimisation problem is resolved one way or the other. I tried to compile with gcc-2.96-80 from rawhide, but *that* wouldn't compile either :/ john -- "What, so the microchips are in the MCU ?" - Sam Burton learning about the PC architecture
Re: LyX crash when mathmod cdots is used
On 12-Apr-2001 John Levon wrote: On Thu, 12 Apr 2001, Juergen Vigna wrote: in Clone() as it is already done in the new() call! Thanks for pointing me to this! I also tried to optimize the new() call for clone a bit. great. however please don't commit this until the optimisation problem is resolved one way or the other. Well don't worry I've still some work to do and won't commit anytime soon :) I tried to compile with gcc-2.96-80 from rawhide, but *that* wouldn't compile either :/ To tell you the truth I compiled with optimization on and don't get the segfault anymore, but don't ask me at what time this happend, in all the changes, backchanges tries, remove tries, other go, ... ;) Jrgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jrgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Coming together is a beginning; keeping together is progress; working together is success.
Re: LyX crash when mathmod cdots is used
On Thu, 12 Apr 2001, Juergen Vigna wrote: To tell you the truth I compiled with optimization on and don't get the segfault anymore, but don't ask me at what time this happend, in all the changes, backchanges tries, remove tries, other go, ... ;) perhaps you altered the size of InsetText at some point ? I got gcc 2.96-80 working (me being silly), and the bug doesn't occur. So I think the patch below or similar should go into CVS. Also, I've got a little further with my gcc 3.0 troubles. It seems c_str() + friends are NOT being inlined, but are being included verbatim (i.e. T in the nm symbol output). This obviously leads to the multiply-defined problems. However, the pre-processed output seems normal. It is being caused by the use of stringstream.str().c_str() in e.g. tostr() template. Removing the c_str() stops the problem. Now obviously we can't do that because of lyxstring, ... so is this a compiler bug or our problem ?[1] thanks john [1] perhaps I'm now compiler-bug happy ;) Index: INSTALL === RCS file: /home/moz/cvs/lyx-devel/INSTALL,v retrieving revision 1.1.1.1 diff -u -r1.1.1.1 INSTALL --- INSTALL 2001/03/19 22:51:19 1.1.1.1 +++ INSTALL 2001/04/12 15:51:26 @@ -325,6 +325,13 @@ If you your version of gcc-c++ is 2.96-69, you may not need to do this. + o If you are using RedHat 7.0, optimisation of the lyx binary is +disabled by default due to a compiler bug in their version of gcc. +If you are using the gcc rpm version of 2.96-80 or above, the bug +has been fixed and optimisation can be re-enabled. Note that LyX +will not compile correctly with a gcc rpm version of less than +2.96-69. + o It is possible to compile lyx with Tru64 Unix cxx compiler version 6.2, provided one uses CXX='cxx -std strict_ansi' Index: config/lyxinclude.m4 === RCS file: /home/moz/cvs/lyx-devel/config/lyxinclude.m4,v retrieving revision 1.2 diff -u -r1.2 lyxinclude.m4 --- config/lyxinclude.m42001/04/05 20:03:02 1.2 +++ config/lyxinclude.m42001/04/12 15:56:53 @@ -206,7 +206,10 @@ 2.95.1) CXXFLAGS="-g $lyx_opt -fpermissive -fno-rtti -fno-exceptions";; 2.95.2) CXXFLAGS="-g $lyx_opt -fno-rtti -fno-exceptions";; 2.95.*) CXXFLAGS="-g $lyx_opt -fno-exceptions";; - 2.96*) CXXFLAGS="-g $lyx_opt -fno-exceptions";; +dnl we can't optimise on 2.96 due to RH's buggy compiler. +dnl 2.96-80 is OK. It is probably worth checking __VERSION__ +dnl so RH7.1 and above can compile with optimisation + 2.96*) CXXFLAGS="-g -fno-exceptions";; 3.0*)CXXFLAGS="-g $lyx_opt";; *2.91.*) CXXFLAGS="-g $lyx_opt -fno-rtti -fno-exceptions";; *) CXXFLAGS="-g $lyx_opt -fno-rtti -fno-exceptions";;
Re: LyX crash when mathmod cdots is used
On 11-Apr-2001 John Levon wrote: > currently on Clone(), (when it works properly) we are resetting each cell's InsetText > three times: > once in Init, where they are constructed, once in operator=, where they shouldbe/are > copied, > and then again in Clone(). > > The second one causes the problem, so if we change the copy c-tor to not use >operator= > itself, > then we can avoid the second part of this. A small optimisation of course ... Well you're right we do an assignment too much! I commented out the asignment in Clone() as it is already done in the new() call! Thanks for pointing me to this! I also tried to optimize the new() call for clone a bit. Jürgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jürgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Gnagloot, n.: A person who leaves all his ski passes on his jacket just to impress people. -- Rich Hall, "Sniglets"
Re: LyX crash when mathmod cdots is used
> Well you're right we do an assignment too much! I commented out the asignment > in Clone() as it is already done in the new() call! Thanks for pointing me to > this! I also tried to optimize the new() call for clone a bit. Just my two cents: Clone() _always_ should look like SomeThing * SomeThing::Clone() const { return new SomeThing(*this); } and everything else should be handled by the copy constructor. Andre' -- André Pönitz . [EMAIL PROTECTED] C++-Programmierer gesucht ... Naeheres unter http://mathematik.htwm.de/job
Re: LyX crash when mathmod cdots is used
On 12-Apr-2001 Andre Poenitz wrote: >> Well you're right we do an assignment too much! I commented out the asignment >> in Clone() as it is already done in the new() call! Thanks for pointing me to >> this! I also tried to optimize the new() call for clone a bit. > > Just my two cents: > > Clone() _always_ should look like > > SomeThing * SomeThing::Clone() const > { > return new SomeThing(*this); > } > > and everything else should be handled by the copy constructor. Well it does now look like that, thought I don't know what's the copy constructor (there was a warning from Lars about this cc but I asked him for explications and didn't understand them so I just left it asis :) Jürgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jürgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Mistakes are often the stepping stones to utter failure.
Re: LyX crash when mathmod cdots is used
> Well it does now look like that, thought I don't know what's the copy > constructor (there was a warning from Lars about this cc but I asked him > for explications and didn't understand them so I just left it asis :) The copy constructor is something like Something::Something(Something const & t) { // make sure we are an exact copy of t } If it is not defined by yourself, the compiler will provide one automatically by using the copy constructors of all Something::members. This is often exactly what is necessary, but it breaks with "naked owning pointers", since it copies the only _pointer_ itself and does not create a copy of the item pointed to. (The same holds true for operator=() btw) This necessity to manually handle copy construction and assignment (and destruction) when using "naked owning pointers" is one of the major reasons why we should used some kind of smart pointer (like boost::shared_pointer) as class members instead. Andre' -- André Pönitz . [EMAIL PROTECTED] C++-Programmierer gesucht ... Naeheres unter http://mathematik.htwm.de/job
Re: LyX crash when mathmod cdots is used
On Thu, 12 Apr 2001, Juergen Vigna wrote: > in Clone() as it is already done in the new() call! Thanks for pointing me to > this! I also tried to optimize the new() call for clone a bit. great. however please don't commit this until the optimisation problem is resolved one way or the other. I tried to compile with gcc-2.96-80 from rawhide, but *that* wouldn't compile either :/ john -- "What, so the microchips are in the MCU ?" - Sam Burton learning about the PC architecture
Re: LyX crash when mathmod cdots is used
On 12-Apr-2001 John Levon wrote: > On Thu, 12 Apr 2001, Juergen Vigna wrote: > >> in Clone() as it is already done in the new() call! Thanks for pointing me to >> this! I also tried to optimize the new() call for clone a bit. > > great. however please don't commit this until the optimisation problem is resolved > one way or the other. Well don't worry I've still some work to do and won't commit anytime soon :) > I tried to compile with gcc-2.96-80 from rawhide, but *that* wouldn't compile either >:/ To tell you the truth I compiled with optimization on and don't get the segfault anymore, but don't ask me at what time this happend, in all the changes, backchanges tries, remove tries, other go, ... ;) Jürgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jürgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Coming together is a beginning; keeping together is progress; working together is success.
Re: LyX crash when mathmod cdots is used
On Thu, 12 Apr 2001, Juergen Vigna wrote: > To tell you the truth I compiled with optimization on and don't get the segfault > anymore, but don't ask me at what time this happend, in all the changes, backchanges > tries, remove tries, other go, ... ;) > perhaps you altered the size of InsetText at some point ? I got gcc 2.96-80 working (me being silly), and the bug doesn't occur. So I think the patch below or similar should go into CVS. Also, I've got a little further with my gcc 3.0 troubles. It seems c_str() + friends are NOT being inlined, but are being included verbatim (i.e. T in the nm symbol output). This obviously leads to the multiply-defined problems. However, the pre-processed output seems normal. It is being caused by the use of stringstream.str().c_str() in e.g. tostr() template. Removing the c_str() stops the problem. Now obviously we can't do that because of lyxstring, ... so is this a compiler bug or our problem ?[1] thanks john [1] perhaps I'm now compiler-bug happy ;) Index: INSTALL === RCS file: /home/moz/cvs/lyx-devel/INSTALL,v retrieving revision 1.1.1.1 diff -u -r1.1.1.1 INSTALL --- INSTALL 2001/03/19 22:51:19 1.1.1.1 +++ INSTALL 2001/04/12 15:51:26 @@ -325,6 +325,13 @@ If you your version of gcc-c++ is > 2.96-69, you may not need to do this. + o If you are using RedHat 7.0, optimisation of the lyx binary is +disabled by default due to a compiler bug in their version of gcc. +If you are using the gcc rpm version of 2.96-80 or above, the bug +has been fixed and optimisation can be re-enabled. Note that LyX +will not compile correctly with a gcc rpm version of less than +2.96-69. + o It is possible to compile lyx with Tru64 Unix cxx compiler version 6.2, provided one uses CXX='cxx -std strict_ansi' Index: config/lyxinclude.m4 === RCS file: /home/moz/cvs/lyx-devel/config/lyxinclude.m4,v retrieving revision 1.2 diff -u -r1.2 lyxinclude.m4 --- config/lyxinclude.m42001/04/05 20:03:02 1.2 +++ config/lyxinclude.m42001/04/12 15:56:53 @@ -206,7 +206,10 @@ 2.95.1) CXXFLAGS="-g $lyx_opt -fpermissive -fno-rtti -fno-exceptions";; 2.95.2) CXXFLAGS="-g $lyx_opt -fno-rtti -fno-exceptions";; 2.95.*) CXXFLAGS="-g $lyx_opt -fno-exceptions";; - 2.96*) CXXFLAGS="-g $lyx_opt -fno-exceptions";; +dnl we can't optimise on 2.96 due to RH's buggy compiler. +dnl 2.96-80 is OK. It is probably worth checking __VERSION__ +dnl so RH7.1 and above can compile with optimisation + 2.96*) CXXFLAGS="-g -fno-exceptions";; 3.0*)CXXFLAGS="-g $lyx_opt";; *2.91.*) CXXFLAGS="-g $lyx_opt -fno-rtti -fno-exceptions";; *) CXXFLAGS="-g $lyx_opt -fno-rtti -fno-exceptions";;
Re: LyX crash when mathmod cdots is used
On 10-Apr-2001 John Levon wrote: But I know, you are right (that's why I spent forever tring to find broken assembly). Unfortunately I cannot seem to compile LyX with gcc 3.0. The string stuff is multiply-defined at link time. config.h looks sensible and seems to be included in the files it is complaining about. Do you use LyXString Class or the stringclass provided by the STL? we can easily work round the tabular optimisation problem, it will also slightly speed up tabular stuff ;) ??? Jrgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jrgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ "Beware of programmers carrying screwdrivers." -- Chip Salzenberg
Re: LyX crash when mathmod cdots is used
On Wed, 11 Apr 2001, Juergen Vigna wrote: On 10-Apr-2001 John Levon wrote: But I know, you are right (that's why I spent forever tring to find broken assembly). Unfortunately I cannot seem to compile LyX with gcc 3.0. The string stuff is multiply-defined at link time. config.h looks sensible and seems to be included in the files it is complaining about. Do you use LyXString Class or the stringclass provided by the STL? Well obviously since gcc 3.0 string is good, I let configure discover this and use std::string from gcc ! When that didn't work I tried --with-included-string, which also broke :/ I'm really confused by this. Maybe this is wrong : export CC=/usr/gcc3.0/bin/gcc export CXX=/usr/gcc3.0/bin/g++ ./configure make I verified lyxstring.h never got included btw ! we can easily work round the tabular optimisation problem, it will also slightly speed up tabular stuff ;) ??? currently on Clone(), (when it works properly) we are resetting each cell's InsetText three times: once in Init, where they are constructed, once in operator=, where they shouldbe/are copied, and then again in Clone(). The second one causes the problem, so if we change the copy c-tor to not use operator= itself, then we can avoid the second part of this. A small optimisation of course ... john -- "I will be sad if I either look up or down after my death and don't see my son fast asleep on the same benches on which I have slept." - Lord Onslow, opposing hereditary peer reforms
Re: LyX crash when mathmod cdots is used
On 10-Apr-2001 John Levon wrote: > But I know, you are right (that's why I spent forever tring to find broken > assembly). Unfortunately I cannot seem to compile LyX with gcc 3.0. The > string stuff is multiply-defined at link time. config.h looks sensible and seems > to be included in the files it is complaining about. Do you use LyXString Class or the stringclass provided by the STL? > we can easily work round the tabular optimisation problem, it will also slightly > speed up tabular stuff ;) ??? Jürgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jürgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ "Beware of programmers carrying screwdrivers." -- Chip Salzenberg
Re: LyX crash when mathmod cdots is used
On Wed, 11 Apr 2001, Juergen Vigna wrote: > > On 10-Apr-2001 John Levon wrote: > > > But I know, you are right (that's why I spent forever tring to find broken > > assembly). Unfortunately I cannot seem to compile LyX with gcc 3.0. The > > string stuff is multiply-defined at link time. config.h looks sensible and seems > > to be included in the files it is complaining about. > > Do you use LyXString Class or the stringclass provided by the STL? Well obviously since gcc 3.0 string is good, I let configure discover this and use std::string from gcc ! When that didn't work I tried --with-included-string, which also broke :/ I'm really confused by this. Maybe this is wrong : export CC=/usr/gcc3.0/bin/gcc export CXX=/usr/gcc3.0/bin/g++ ./configure make I verified lyxstring.h never got included btw ! > > we can easily work round the tabular optimisation problem, it will also slightly > > speed up tabular stuff ;) > > ??? currently on Clone(), (when it works properly) we are resetting each cell's InsetText three times: once in Init, where they are constructed, once in operator=, where they shouldbe/are copied, and then again in Clone(). The second one causes the problem, so if we change the copy c-tor to not use operator= itself, then we can avoid the second part of this. A small optimisation of course ... john -- "I will be sad if I either look up or down after my death and don't see my son fast asleep on the same benches on which I have slept." - Lord Onslow, opposing hereditary peer reforms
Re: LyX crash when mathmod cdots is used
"Dekel" == Dekel Tsur [EMAIL PROTECTED] writes: Dekel On Mon, Apr 09, 2001 at 09:56:22AM -0700, Michael K Cheng Dekel wrote: I got some feedback and it apprears with redhat 7, you'll have to compile without the optimization flag "-O2" so set CXXFLAGS='-g' and recompile worked!! Dekel So perhaps we should disable optimizations for gcc 2.96 in the Dekel configure script. It depends. Does it work with latest rpm revision or not? JMarc
Re: LyX crash when mathmod cdots is used
"Michael" == Michael K Cheng [EMAIL PROTECTED] writes: Michael I got some feedback and it apprears with redhat 7, you'll Michael have to compile without the optimization flag "-O2" so set Michael CXXFLAGS='-g' and recompile worked!! What revision of the gcc rpm do you have? JMarc
Re: LyX crash when mathmod cdots is used
On 10-Apr-2001 Jean-Marc Lasgouttes wrote: It depends. Does it work with latest rpm revision or not? I have the latest and it doesn't work! Jrgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jrgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ A rolling stone gathers no moss. -- Publilius Syrus
Re: LyX crash when mathmod cdots is used
"Juergen" == Juergen Vigna [EMAIL PROTECTED] writes: Juergen On 10-Apr-2001 Jean-Marc Lasgouttes wrote: It depends. Does it work with latest rpm revision or not? Juergen I have the latest and it doesn't work! You mean you have a crash on outputing \cdots? JMarc
Re: LyX crash when mathmod cdots is used
On 10-Apr-2001 Jean-Marc Lasgouttes wrote: Juergen I have the latest and it doesn't work! You mean you have a crash on outputing \cdots? Well no I don't use Mathed, but I crash when using tabulars and compiling with optimization activated! Is that enough? Jrgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jrgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ "Get back to your stations!" "We're beaming down to the planet, sir." -- Kirk and Mr. Leslie, "This Side of Paradise", stardate 3417.3
Re: LyX crash when mathmod cdots is used
"Juergen" == Juergen Vigna [EMAIL PROTECTED] writes: Juergen Well no I don't use Mathed, but I crash when using tabulars Juergen and compiling with optimization activated! Is that enough? I was interested in the particular \cdots bug. We have to make a difference between bugs that are fixed by recent gcc-2.97 updates and those which are not... Concerning the tabular bug, we might get redhat to fix it if we had a small test example, but I understand this is difficult. JMarc
Re: LyX crash when mathmod cdots is used
On 10-Apr-2001 Jean-Marc Lasgouttes wrote: Concerning the tabular bug, we might get redhat to fix it if we had a small test example, but I understand this is difficult. I guess it's not difficult at all. Just compile the source with optimization create a tabular and try to type inside it. Here it crashes every time! Jrgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jrgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ "And remember: Evil will always prevail, because Good is dumb." -- Spaceballs
Re: LyX crash when mathmod cdots is used
"Juergen" == Juergen Vigna [EMAIL PROTECTED] writes: Juergen I guess it's not difficult at all. Just compile the source Juergen with optimization create a tabular and try to type inside it. Juergen Here it crashes every time! I mean provide a source code with less than 100k lines! JMarc PS: anyway I don't care, it does not crash here :)
Re: LyX crash when mathmod cdots is used
On 10 Apr 2001, Jean-Marc Lasgouttes wrote: "Juergen" == Juergen Vigna [EMAIL PROTECTED] writes: Juergen I guess it's not difficult at all. Just compile the source Juergen with optimization create a tabular and try to type inside it. Juergen Here it crashes every time! I mean provide a source code with less than 100k lines! Well as I said it's narrowed down to one line behaving really oddly, so ... But I know, you are right (that's why I spent forever tring to find broken assembly). Unfortunately I cannot seem to compile LyX with gcc 3.0. The string stuff is multiply-defined at link time. config.h looks sensible and seems to be included in the files it is complaining about. I'm totally stuck on this one too, I'm really pissed off with thing at the moment ... we can easily work round the tabular optimisation problem, it will also slightly speed up tabular stuff ;) PS: anyway I don't care, it does not crash here :) ;) thanks john -- "Before we had central heating, we used to put our socks and pants on the radiator to warm them up." - Dr. Fox
Re: LyX crash when mathmod cdots is used
> "Dekel" == Dekel Tsur <[EMAIL PROTECTED]> writes: Dekel> On Mon, Apr 09, 2001 at 09:56:22AM -0700, Michael K Cheng Dekel> wrote: >> I got some feedback and it apprears with redhat 7, you'll have to >> compile without the optimization flag "-O2" so set CXXFLAGS='-g' >> and recompile worked!! Dekel> So perhaps we should disable optimizations for gcc 2.96 in the Dekel> configure script. It depends. Does it work with latest rpm revision or not? JMarc
Re: LyX crash when mathmod cdots is used
> "Michael" == Michael K Cheng <[EMAIL PROTECTED]> writes: Michael> I got some feedback and it apprears with redhat 7, you'll Michael> have to compile without the optimization flag "-O2" so set Michael> CXXFLAGS='-g' and recompile worked!! What revision of the gcc rpm do you have? JMarc
Re: LyX crash when mathmod cdots is used
On 10-Apr-2001 Jean-Marc Lasgouttes wrote: > It depends. Does it work with latest rpm revision or not? I have the latest and it doesn't work! Jürgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jürgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ A rolling stone gathers no moss. -- Publilius Syrus
Re: LyX crash when mathmod cdots is used
> "Juergen" == Juergen Vigna <[EMAIL PROTECTED]> writes: Juergen> On 10-Apr-2001 Jean-Marc Lasgouttes wrote: >> It depends. Does it work with latest rpm revision or not? Juergen> I have the latest and it doesn't work! You mean you have a crash on outputing \cdots? JMarc
Re: LyX crash when mathmod cdots is used
On 10-Apr-2001 Jean-Marc Lasgouttes wrote: > Juergen> I have the latest and it doesn't work! > > You mean you have a crash on outputing \cdots? Well no I don't use Mathed, but I crash when using tabulars and compiling with optimization activated! Is that enough? Jürgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jürgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ "Get back to your stations!" "We're beaming down to the planet, sir." -- Kirk and Mr. Leslie, "This Side of Paradise", stardate 3417.3
Re: LyX crash when mathmod cdots is used
> "Juergen" == Juergen Vigna <[EMAIL PROTECTED]> writes: Juergen> Well no I don't use Mathed, but I crash when using tabulars Juergen> and compiling with optimization activated! Is that enough? I was interested in the particular \cdots bug. We have to make a difference between bugs that are fixed by recent gcc-2.97 updates and those which are not... Concerning the tabular bug, we might get redhat to fix it if we had a small test example, but I understand this is difficult. JMarc
Re: LyX crash when mathmod cdots is used
On 10-Apr-2001 Jean-Marc Lasgouttes wrote: > Concerning the tabular bug, we might get redhat to fix it if we had a > small test example, but I understand this is difficult. I guess it's not difficult at all. Just compile the source with optimization create a tabular and try to type inside it. Here it crashes every time! Jürgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jürgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ "And remember: Evil will always prevail, because Good is dumb." -- Spaceballs
Re: LyX crash when mathmod cdots is used
> "Juergen" == Juergen Vigna <[EMAIL PROTECTED]> writes: Juergen> I guess it's not difficult at all. Just compile the source Juergen> with optimization create a tabular and try to type inside it. Juergen> Here it crashes every time! I mean provide a source code with less than 100k lines! JMarc PS: anyway I don't care, it does not crash here :)
Re: LyX crash when mathmod cdots is used
On 10 Apr 2001, Jean-Marc Lasgouttes wrote: > > "Juergen" == Juergen Vigna <[EMAIL PROTECTED]> writes: > > Juergen> I guess it's not difficult at all. Just compile the source > Juergen> with optimization create a tabular and try to type inside it. > Juergen> Here it crashes every time! > > I mean provide a source code with less than 100k lines! > Well as I said it's narrowed down to one line behaving really oddly, so ... But I know, you are right (that's why I spent forever tring to find broken assembly). Unfortunately I cannot seem to compile LyX with gcc 3.0. The string stuff is multiply-defined at link time. config.h looks sensible and seems to be included in the files it is complaining about. I'm totally stuck on this one too, I'm really pissed off with thing at the moment ... we can easily work round the tabular optimisation problem, it will also slightly speed up tabular stuff ;) > PS: anyway I don't care, it does not crash here :) ;) thanks john -- "Before we had central heating, we used to put our socks and pants on the radiator to warm them up." - Dr. Fox
Re: LyX crash when mathmod cdots is used
"Michael" == Michael K Cheng [EMAIL PROTECTED] writes: Michael Content-Type: TEXT/PLAIN; charset=US-ASCII LyX crashes Michael whenever I save or dvi a document that contains 'cdots' in Michael math equations. Are you using redhat 7? If you do, did you upgrade gcc to at least -69 rpm revision? JMarc
Re: LyX crash when mathmod cdots is used
I got some feedback and it apprears with redhat 7, you'll have to compile without the optimization flag "-O2" so set CXXFLAGS='-g' and recompile worked!! Thanks, Mike On 9 Apr 2001, Jean-Marc Lasgouttes wrote: "Michael" == Michael K Cheng [EMAIL PROTECTED] writes: Michael Content-Type: TEXT/PLAIN; charset=US-ASCII LyX crashes Michael whenever I save or dvi a document that contains 'cdots' in Michael math equations. Are you using redhat 7? If you do, did you upgrade gcc to at least -69 rpm revision? JMarc -- - |Michael K. Chenghttp://ieng9.ucsd.edu/~kcheng/ | - |Signal Transmission and Recording (STAR) Lab, UCSD | |Tel: (858)822-0436 http://www-cwc.ucsd.edu/star/ | -
Re: LyX crash when mathmod cdots is used
On Mon, Apr 09, 2001 at 09:56:22AM -0700, Michael K Cheng wrote: I got some feedback and it apprears with redhat 7, you'll have to compile without the optimization flag "-O2" so set CXXFLAGS='-g' and recompile worked!! So perhaps we should disable optimizations for gcc 2.96 in the configure script.
Re: LyX crash when mathmod cdots is used
> "Michael" == Michael K Cheng <[EMAIL PROTECTED]> writes: Michael> Content-Type: TEXT/PLAIN; charset=US-ASCII LyX crashes Michael> whenever I save or dvi a document that contains 'cdots' in Michael> math equations. Are you using redhat 7? If you do, did you upgrade gcc to at least -69 rpm revision? JMarc
Re: LyX crash when mathmod cdots is used
I got some feedback and it apprears with redhat 7, you'll have to compile without the optimization flag "-O2" so set CXXFLAGS='-g' and recompile worked!! Thanks, Mike On 9 Apr 2001, Jean-Marc Lasgouttes wrote: > > "Michael" == Michael K Cheng <[EMAIL PROTECTED]> writes: > > Michael> Content-Type: TEXT/PLAIN; charset=US-ASCII LyX crashes > Michael> whenever I save or dvi a document that contains 'cdots' in > Michael> math equations. > > Are you using redhat 7? If you do, did you upgrade gcc to at least -69 > rpm revision? > > JMarc > -- - |Michael K. Chenghttp://ieng9.ucsd.edu/~kcheng/ | - |Signal Transmission and Recording (STAR) Lab, UCSD | |Tel: (858)822-0436 http://www-cwc.ucsd.edu/star/ | -
Re: LyX crash when mathmod cdots is used
On Mon, Apr 09, 2001 at 09:56:22AM -0700, Michael K Cheng wrote: > I got some feedback and it apprears with redhat 7, you'll have to compile > without the optimization flag "-O2" so set CXXFLAGS='-g' and recompile > worked!! So perhaps we should disable optimizations for gcc 2.96 in the configure script.
Re: LyX crash when mathmod cdots is used
On 04-Apr-2001 Dekel Tsur wrote: On Wed, Apr 04, 2001 at 10:23:05AM -0700, Michael K Cheng wrote: I'm not sure what that os pointer is, but I've compiled LyX on redhat 7.0 yesterday. I can crash it by opening a new document and inserting a equation [M-m d] and type "s=0,\cdots,d-1" (for s=0,...,d-1), then save or dvi will cause it to crash. Try opening and dvi(ing) or saving the attached document. g++ version: Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/2.96/specs gcc version 2.96 2731 (Red Hat Linux 7.0) This seems so much the same problem I have with that compiler! Compile LyX without optimization (CXXFLAGS=-g) this should fix it. If I compile with optimization activated I get the same problems as you (spurius crashes when working with a lot of paragraphs at the same time like you do if you have a tabular!) Jrgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jrgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ QOTD: "It's been real and it's been fun, but it hasn't been real fun."
Re: LyX crash when mathmod cdots is used
"Juergen" == Juergen Vigna [EMAIL PROTECTED] writes: Juergen On 04-Apr-2001 Dekel Tsur wrote: On Wed, Apr 04, 2001 at 10:23:05AM -0700, Michael K Cheng wrote: I'm not sure what that os pointer is, but I've compiled LyX on redhat 7.0 yesterday. I can crash it by opening a new document and inserting a equation [M-m d] and type "s=0,\cdots,d-1" (for s=0,...,d-1), then save or dvi will cause it to crash. Try opening and dvi(ing) or saving the attached document. g++ version: Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/2.96/specs gcc version 2.96 2731 (Red Hat Linux 7.0) Juergen This seems so much the same problem I have with that Juergen compiler! Compile LyX without optimization (CXXFLAGS=-g) this Juergen should fix it. If I compile with optimization activated I get Juergen the same problems as you (spurius crashes when working with a Juergen lot of paragraphs at the same time like you do if you have a Juergen tabular!) And check that you have the latest revision of the gcc rpms (at least -69, from what I understand). JMarc
Re: LyX crash when mathmod cdots is used
On 04-Apr-2001 Dekel Tsur wrote: > On Wed, Apr 04, 2001 at 10:23:05AM -0700, Michael K Cheng wrote: >> I'm not sure what that os pointer is, but I've compiled LyX on redhat >> 7.0 yesterday. I can crash it by opening a new document and inserting a >> equation [M-m d] and type "s=0,\cdots,d-1" (for s=0,...,d-1), then save or >> dvi will cause it to crash. Try opening and dvi(ing) or saving the >> attached document. >> >> g++ version: >> Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/2.96/specs >> gcc version 2.96 2731 (Red Hat Linux 7.0) This seems so much the same problem I have with that compiler! Compile LyX without optimization (CXXFLAGS=-g) this should fix it. If I compile with optimization activated I get the same problems as you (spurius crashes when working with a lot of paragraphs at the same time like you do if you have a tabular!) Jürgen -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jürgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ QOTD: "It's been real and it's been fun, but it hasn't been real fun."
Re: LyX crash when mathmod cdots is used
> "Juergen" == Juergen Vigna <[EMAIL PROTECTED]> writes: Juergen> On 04-Apr-2001 Dekel Tsur wrote: >> On Wed, Apr 04, 2001 at 10:23:05AM -0700, Michael K Cheng wrote: >>> I'm not sure what that os pointer is, but I've compiled LyX on >>> redhat 7.0 yesterday. I can crash it by opening a new document and >>> inserting a equation [M-m d] and type "s=0,\cdots,d-1" (for >>> s=0,...,d-1), then save or dvi will cause it to crash. Try opening >>> and dvi(ing) or saving the attached document. >>> >>> g++ version: Reading specs from >>> /usr/lib/gcc-lib/i386-redhat-linux/2.96/specs gcc version 2.96 >>> 2731 (Red Hat Linux 7.0) Juergen> This seems so much the same problem I have with that Juergen> compiler! Compile LyX without optimization (CXXFLAGS=-g) this Juergen> should fix it. If I compile with optimization activated I get Juergen> the same problems as you (spurius crashes when working with a Juergen> lot of paragraphs at the same time like you do if you have a Juergen> tabular!) And check that you have the latest revision of the gcc rpms (at least -69, from what I understand). JMarc
LyX crash when mathmod cdots is used
LyX crashes whenever I save or dvi a document that contains 'cdots' in math equations. Thanks. (gdb) run Starting program: /extra/install/lyx-1.1.6fix1/src/lyx [New Thread 1024 (LWP 5905)] Warning: Setting language `default' to `english'. Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 1024 (LWP 5905)] 0x40151530 in ostream::operator () from /usr/lib/libstdc++-libc6.2-2.so.3 Current language: auto; currently c (gdb) bt #0 0x40151530 in ostream::operator () from /usr/lib/libstdc++-libc6.2-2.so.3 #1 0x81aefbf in MathDotsInset::Write (this=0x8413670, os=@0x20) at math_write.C:59 #2 0x81af897 in MathParInset::Write (this=0x8413300, os=@0xbfffcc18, fragile=false) at math_write.C:237 #3 0x81afd74 in mathed_write (p=0x8413300, os=@0xbfffcc18, newlines=0xbfffc94c, fragile=false, label=@0x83d2ff4) at math_write.C:350 #4 0x8197d19 in InsetFormula::Latex (this=0x83d2fd0, os=@0xbfffcc18, fragile=false) at formula.C:331 #5 0x8197ce0 in InsetFormula::Write (this=0x83d2fd0, buf=0x8413348, os=@0xbfffcc18) at formula.C:322 #6 0x814813a in LyXParagraph::writeFile (this=0x8414020, buf=0x8413348, os=@0xbfffcc18, params=@0x8413358, footflag=0 '\000', dth=0 '\000') at paragraph.C:280 #7 0x8096cee in Buffer::writeFile (this=0x8413348, fname=@0x84134d8, flag=false) at buffer.C:1306 #8 0x809653e in Buffer::save (this=0x8413348) at buffer.C:1193 #9 0x80ee34c in MenuWrite (bv=0x83cd420, buffer=0x8413348) at lyx_cb.C:168 #10 0x80e in MenuWriteAs (bv=0x83cd420, buffer=0x8413348) at lyx_cb.C:260 #11 0x8111e9a in LyXFunc::Dispatch (this=0x83cec48, ac=6, do_not_use_this_arg=@0xb42c) at lyxfunc.C:811 #12 0x825d9e2 in Menubar::Pimpl::MenuCallback (ob=0x83c6c20, button=1) at /usr/include/g++-3/std/bastring.h:176 #13 0x8289d4c in fl_object_qread () at ../sigc++/slot.h:111 #14 0x82888b5 in fl_check_forms () at ../sigc++/slot.h:111 #15 0x8207160 in GUIRunTime::runTime () at GUIRunTime.C:79 #16 0x80f7dc8 in LyXGUI::runTime (this=0x83a6398) at lyx_gui.C:419 #17 0x80f9f74 in LyX::LyX (this=0x0, argc=0xb6d0, argv=0xb734) at ../src/lyx_main.C:168 #18 0x8145326 in main (argc=1, argv=0xb734) at ../src/main.C:40 #19 0x401b3e51 in __libc_start_main (main=0x8145190 main, argc=1, ubp_av=0xb734, init=0x804e0ec _init, fini=0x83258e4 _fini, rtld_fini=0x4000e264 _dl_fini, stack_end=0xb72c) at ../sysdeps/generic/libc-start.c:111 (gdb) (gdb) info locals No symbol table info available. (gdb) up #1 0x81aefbf in MathDotsInset::Write (this=0x8413670, os=@0x20) at math_write.C:59 59 os '\\' name ' '; Current language: auto; currently c++ (gdb)
Re: LyX crash when mathmod cdots is used
On Wed, Apr 04, 2001 at 01:08:44AM -0700, Michael K Cheng wrote: LyX crashes whenever I save or dvi a document that contains 'cdots' in math equations. I can't recreate the problem. Can you give details on your system (compiler, libraries etc.). #0 0x40151530 in ostream::operator () from /usr/lib/libstdc++-libc6.2-2.so.3 #1 0x81aefbf in MathDotsInset::Write (this=0x8413670, os=@0x20) ^ How did this happen ? at math_write.C:59 #2 0x81af897 in MathParInset::Write (this=0x8413300, os=@0xbfffcc18, fragile=false) at math_write.C:237
Re: LyX crash when mathmod cdots is used
On Wed, Apr 04, 2001 at 10:23:05AM -0700, Michael K Cheng wrote: I'm not sure what that os pointer is, but I've compiled LyX on redhat 7.0 yesterday. I can crash it by opening a new document and inserting a equation [M-m d] and type "s=0,\cdots,d-1" (for s=0,...,d-1), then save or dvi will cause it to crash. Try opening and dvi(ing) or saving the attached document. g++ version: Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/2.96/specs gcc version 2.96 2731 (Red Hat Linux 7.0) Also the emergency save does'nt work. All the document.lyx.emergency files are of 0 bytes!! and it also truncates the original document.lyx file to 0 bytes. Very dangerous!! I was able to recover with document.lyx~ backup, otherwise a crash would wipe out my doc entirely. It seems that there is something wrong with your LyX installation, and I don't know the reason for it. Maybe your compiler ???
LyX crash when mathmod cdots is used
LyX crashes whenever I save or dvi a document that contains 'cdots' in math equations. Thanks. (gdb) run Starting program: /extra/install/lyx-1.1.6fix1/src/lyx [New Thread 1024 (LWP 5905)] Warning: Setting language `default' to `english'. Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 1024 (LWP 5905)] 0x40151530 in ostream::operator<< () from /usr/lib/libstdc++-libc6.2-2.so.3 Current language: auto; currently c (gdb) bt #0 0x40151530 in ostream::operator<< () from /usr/lib/libstdc++-libc6.2-2.so.3 #1 0x81aefbf in MathDotsInset::Write (this=0x8413670, os=@0x20) at math_write.C:59 #2 0x81af897 in MathParInset::Write (this=0x8413300, os=@0xbfffcc18, fragile=false) at math_write.C:237 #3 0x81afd74 in mathed_write (p=0x8413300, os=@0xbfffcc18, newlines=0xbfffc94c, fragile=false, label=@0x83d2ff4) at math_write.C:350 #4 0x8197d19 in InsetFormula::Latex (this=0x83d2fd0, os=@0xbfffcc18, fragile=false) at formula.C:331 #5 0x8197ce0 in InsetFormula::Write (this=0x83d2fd0, buf=0x8413348, os=@0xbfffcc18) at formula.C:322 #6 0x814813a in LyXParagraph::writeFile (this=0x8414020, buf=0x8413348, os=@0xbfffcc18, params=@0x8413358, footflag=0 '\000', dth=0 '\000') at paragraph.C:280 #7 0x8096cee in Buffer::writeFile (this=0x8413348, fname=@0x84134d8, flag=false) at buffer.C:1306 #8 0x809653e in Buffer::save (this=0x8413348) at buffer.C:1193 #9 0x80ee34c in MenuWrite (bv=0x83cd420, buffer=0x8413348) at lyx_cb.C:168 #10 0x80e in MenuWriteAs (bv=0x83cd420, buffer=0x8413348) at lyx_cb.C:260 #11 0x8111e9a in LyXFunc::Dispatch (this=0x83cec48, ac=6, do_not_use_this_arg=@0xb42c) at lyxfunc.C:811 #12 0x825d9e2 in Menubar::Pimpl::MenuCallback (ob=0x83c6c20, button=1) at /usr/include/g++-3/std/bastring.h:176 #13 0x8289d4c in fl_object_qread () at ../sigc++/slot.h:111 #14 0x82888b5 in fl_check_forms () at ../sigc++/slot.h:111 #15 0x8207160 in GUIRunTime::runTime () at GUIRunTime.C:79 #16 0x80f7dc8 in LyXGUI::runTime (this=0x83a6398) at lyx_gui.C:419 #17 0x80f9f74 in LyX::LyX (this=0x0, argc=0xb6d0, argv=0xb734) at ../src/lyx_main.C:168 #18 0x8145326 in main (argc=1, argv=0xb734) at ../src/main.C:40 #19 0x401b3e51 in __libc_start_main (main=0x8145190 , argc=1, ubp_av=0xb734, init=0x804e0ec <_init>, fini=0x83258e4 <_fini>, rtld_fini=0x4000e264 <_dl_fini>, stack_end=0xb72c) at ../sysdeps/generic/libc-start.c:111 (gdb) (gdb) info locals No symbol table info available. (gdb) up #1 0x81aefbf in MathDotsInset::Write (this=0x8413670, os=@0x20) at math_write.C:59 59 os << '\\' << name << ' '; Current language: auto; currently c++ (gdb)
Re: LyX crash when mathmod cdots is used
On Wed, Apr 04, 2001 at 01:08:44AM -0700, Michael K Cheng wrote: > LyX crashes whenever I save or dvi a document that contains > 'cdots' in math equations. I can't recreate the problem. Can you give details on your system (compiler, libraries etc.). > #0 0x40151530 in ostream::operator<< () from /usr/lib/libstdc++-libc6.2-2.so.3 > #1 0x81aefbf in MathDotsInset::Write (this=0x8413670, os=@0x20) ^ How did this happen ? > at math_write.C:59 > #2 0x81af897 in MathParInset::Write (this=0x8413300, os=@0xbfffcc18, > fragile=false) at math_write.C:237
Re: LyX crash when mathmod cdots is used
On Wed, Apr 04, 2001 at 10:23:05AM -0700, Michael K Cheng wrote: > I'm not sure what that os pointer is, but I've compiled LyX on redhat > 7.0 yesterday. I can crash it by opening a new document and inserting a > equation [M-m d] and type "s=0,\cdots,d-1" (for s=0,...,d-1), then save or > dvi will cause it to crash. Try opening and dvi(ing) or saving the > attached document. > > g++ version: > Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/2.96/specs > gcc version 2.96 2731 (Red Hat Linux 7.0) > > Also the emergency save does'nt work. All the .lyx.emergency > files are of 0 bytes!! and it also truncates the original .lyx > file to 0 bytes. Very dangerous!! I was able to recover with > .lyx~ backup, otherwise a crash would wipe out my doc entirely. It seems that there is something wrong with your LyX installation, and I don't know the reason for it. Maybe your compiler ???