Re: [Haifux] some addendums to today's gdb meeting
Hi, > 6. regarding re-creating a live process from a core file - some people > already thought about this (as usual). here is an example (i didn't try > this yet): > > http://www.geocities.com/asimshankar/checkpointing/ A long time ago, there was a [n almost] standard UNIX command called undump which would re-create an executable out of a core file. I think PERL users might be familiar with that. If I recall correctly this was one of the first ways to "compile" a perl script to an executable. You start PERL, and make it dump core straight at the beginning. Then you would recreate the perl process via undump. I'm really sorry I missed your lecture. I'm sure it was a excellent one, like always. Emil > > i think there was something else we wanted an answer for - does anyone > remember? > > --guy > ___ > Haifux mailing list > Haifux@haifux.org > http://hamakor.org.il/cgi-bin/mailman/listinfo/haifux > ___ Haifux mailing list Haifux@haifux.org http://hamakor.org.il/cgi-bin/mailman/listinfo/haifux
Re: [Haifux] some addendums to today's gdb meeting
Also, if anyone has information on "the quick and easy way" to debug libc functions (such as tzset), I'd much appreciate it. On Mon, May 4, 2009 at 11:34 PM, Hai Zaar wrote: > On Mon, May 4, 2009 at 10:04 PM, guy keren wrote: > > > > i think there was something else we wanted an answer for - does anyone > > remember? > Hi, Guy! > Thank for the real goo lecture. > I remind you address these topic mentioned today but not covered: > * Debugging python C modules > * Share you experience about debugging optimized code > > Thanks again! > -- > Zaar > ___ > Haifux mailing list > Haifux@haifux.org > http://hamakor.org.il/cgi-bin/mailman/listinfo/haifux > -- Man is the only animal that laughs and weeps, for he is the only animal that is struck with the difference between what things are and what they ought to be. - William Hazlitt Ohad Lutzky ___ Haifux mailing list Haifux@haifux.org http://hamakor.org.il/cgi-bin/mailman/listinfo/haifux
Re: [Haifux] some addendums to today's gdb meeting
On Mon, May 4, 2009 at 10:04 PM, guy keren wrote: > > i think there was something else we wanted an answer for - does anyone > remember? Hi, Guy! Thank for the real goo lecture. I remind you address these topic mentioned today but not covered: * Debugging python C modules * Share you experience about debugging optimized code Thanks again! -- Zaar ___ Haifux mailing list Haifux@haifux.org http://hamakor.org.il/cgi-bin/mailman/listinfo/haifux
[Haifux] some addendums to today's gdb meeting
1. clarification: the command 'watch *(char*)0x65476 == 0' puts a hardware watch-point for a single byte at the given address. if you want to find a change to an item with a different size (e.g. 4 bytes), use a different cast: 'watch *(int*)0x65476 == 0' 2. regarding the 'x' command and endianess: you can control the assumed-endianness of gdb using 'set endian big', or 'set endian little' (the default is 'auto'). use 'show endian' to see the current value of this setting. 3. regarding watch points: you can use 'rwatch ' to be notified whenever the value of this expression is _read_, so you can find when someone access a certain memory area (for example). 4. in general 'help ' or 'help ' gives help about this command. 5. regarding checkpoints and fork on windows - indeed it looks like cygwin's fork copies all the memory to the new process - which means that its a very expensive system call. thus, if your process has a very large footprint, or you try to make too many checkpoints - you'll hit swapping or the virtual memory limitation of your machine. 6. regarding re-creating a live process from a core file - some people already thought about this (as usual). here is an example (i didn't try this yet): http://www.geocities.com/asimshankar/checkpointing/ i think there was something else we wanted an answer for - does anyone remember? --guy ___ Haifux mailing list Haifux@haifux.org http://hamakor.org.il/cgi-bin/mailman/listinfo/haifux