|
From: Yao Qi <qiy...@cn...> - 2005-11-29 09:20:35
Attachments:
writing-tools.patch
|
I updated the "Debugging C funtions" in docs/writing-tools.xml per README_DEVELOPERS, and I also modified "Regression Tests". Here is a patch in attachment, any comments? -- Regards, Yao ------------ Yao Qi |
|
From: Tom H. <to...@co...> - 2005-11-29 10:02:34
|
In message <200...@cn...>
Yao Qi <qiy...@cn...> wrote:
> I updated the "Debugging C funtions" in docs/writing-tools.xml per
> README_DEVELOPERS, and I also modified "Regression Tests". Here is a
> patch in attachment, any comments?
Applied (with a few tweaks). Hopefully the XML is right but as it
seems to be virtually impossible to convert the XML to anything
readable on my systems I can't check it.
One thing I didn't apply is the core->vgcore change as I don't think
it is right - it all depends what "C functions used by your tool" is
taken to mean.
If it means the main tool code then it is core that you need to debug
not vgcore - only functions in the preload library (and possibly helper
functions called from generated code?) can be debugged from a vgcore
image.
Given that the next section uses the same phrase and is clearly
talking about debugging the main tool code I left it as core rather
than vgcore.
Tom
--
Tom Hughes (to...@co...)
http://www.compton.nu/
|
|
From: Tom H. <to...@co...> - 2005-11-29 11:07:28
|
In message <200...@cn...>
Yao Qi <qiy...@cn...> wrote:
> I could convert XML to html by 'make all-docs' in docs/. Maybe I could
> send writing-tools.html to you and you could have look at it. If there
> is something wrong, I would like to improve docs/xml/writing-tools.xml.
The problem is that I use Fedora and they ship a version of xmltex
(20020625) which contains a version of pdfxmltex (3.141592-1.21a-2.2)
which doesn't work with the stylesheets etc that the valgrind XML docs
use.
Actually it turns out that the HTML generation does work so I can
work from that.
> "core" and "vgcore" would be distinguished in some document, since they are
> a little confusing, at least to me ,:-). Any opinions?
Basically core files are generated by the OS and represent the process
as seen by the kernel, so are a dump of valgrind itself.
The vgcore files are generated by valgrind and represent the virtual
process it is emulating just as if the program valgrind is running was
running outside valgrind and the kernel generated a core dump for it.
So code running on the virtual CPU will produce a vgcore file if it
dies with a fatal signal while code running on the real CPU will
produce a core file.
Tom
--
Tom Hughes (to...@co...)
http://www.compton.nu/
|
|
From: Nicholas N. <nj...@cs...> - 2005-11-29 16:11:57
|
On Tue, 29 Nov 2005, Tom Hughes wrote: > Basically core files are generated by the OS and represent the process > as seen by the kernel, so are a dump of valgrind itself. > > The vgcore files are generated by valgrind and represent the virtual > process it is emulating just as if the program valgrind is running was > running outside valgrind and the kernel generated a core dump for it. > > So code running on the virtual CPU will produce a vgcore file if it > dies with a fatal signal while code running on the real CPU will > produce a core file. I wonder if this distinction is more trouble than it's worth. Perhaps Valgrind could produce core files with the same name that the OS would give. The same corefile name would be used regardless of whether the client or Valgrind crashed. Would that less confusing than having to explain about "vgcore"? Nick |
|
From: Yao Qi <qiy...@cn...> - 2005-11-29 10:57:49
|
On Tue, Nov 29, 2005 at 10:02:23AM +0000, Tom Hughes wrote: > In message <200...@cn...> > Yao Qi <qiy...@cn...> wrote: > > > I updated the "Debugging C funtions" in docs/writing-tools.xml per > > README_DEVELOPERS, and I also modified "Regression Tests". Here is a > > patch in attachment, any comments? > > Applied (with a few tweaks). Hopefully the XML is right but as it > seems to be virtually impossible to convert the XML to anything > readable on my systems I can't check it. Thanks for your comment in advance! I could convert XML to html by 'make all-docs' in docs/. Maybe I could send writing-tools.html to you and you could have look at it. If there is something wrong, I would like to improve docs/xml/writing-tools.xml. > > One thing I didn't apply is the core->vgcore change as I don't think > it is right - it all depends what "C functions used by your tool" is > taken to mean. > > If it means the main tool code then it is core that you need to debug > not vgcore - only functions in the preload library (and possibly helper > functions called from generated code?) can be debugged from a vgcore > image. > > Given that the next section uses the same phrase and is clearly > talking about debugging the main tool code I left it as core rather > than vgcore. Well, It seems that I misunderstand "C functions" in that part. "core" and "vgcore" would be distinguished in some document, since they are a little confusing, at least to me ,:-). Any opinions? > > Tom > > -- > Tom Hughes (to...@co...) > http://www.compton.nu/ > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click > _______________________________________________ > Valgrind-developers mailing list > Val...@li... > https://lists.sourceforge.net/lists/listinfo/valgrind-developers > -- Regards, Yao ------------ Yao Qi |