valgrind-valkyrie-dev — Discussion list for developers of Valkyrie, a GUI for Valgrind

[Valgrind-valkyrie-dev] valkyrie: r422 - trunk/valkyrie/tool_views

[<sv...@va...>]

Author: cerion
Date: 2008-06-07 16:44:20 +0100 (Sat, 07 Jun 2008)
New Revision: 422

Log:
Don't prevent horizontal scrolling for log window
Caught by Anton Dudarenko.



Modified:
   trunk/valkyrie/tool_views/memcheck_view.cpp


Modified: trunk/valkyrie/tool_views/memcheck_view.cpp
===================================================================
--- trunk/valkyrie/tool_views/memcheck_view.cpp	2008-06-07 14:48:43 UTC (rev 421)
+++ trunk/valkyrie/tool_views/memcheck_view.cpp	2008-06-07 15:44:20 UTC (rev 422)
@@ -79,7 +79,6 @@
    lView->setSorting( -1 );
    lView->setMargin( 5 );
    lView->addColumn( "" );
-   lView->header()->setStretchEnabled( true, 0 );
    lView->header()->hide();
 
    savelogButton->setEnabled( false );

[Valgrind-valkyrie-dev] valkyrie: r421 - trunk/valkyrie/core

[<sv...@va...>]

Author: cerion
Date: 2008-06-07 15:48:43 +0100 (Sat, 07 Jun 2008)
New Revision: 421

Log:
Fix null ptr defef, happens if memcheck finished before expected while
reading logfile.

Thanks to Gert Steenssens for tracking this one down.



Modified:
   trunk/valkyrie/core/memcheck_object.cpp


Modified: trunk/valkyrie/core/memcheck_object.cpp
===================================================================
--- trunk/valkyrie/core/memcheck_object.cpp	2008-06-07 14:17:59 UTC (rev 420)
+++ trunk/valkyrie/core/memcheck_object.cpp	2008-06-07 14:48:43 UTC (rev 421)
@@ -629,9 +629,9 @@
    qApp->processEvents();
 
    /* Try reading some more data */
-   if ( !m_vgreader->parseContinue()) {
+   if ( m_vgreader && !m_vgreader->parseContinue()) {
       /* Parsing failed: stop m_vgproc, if running */
-      if (m_vgproc->isRunning())
+      if ( m_vgproc && m_vgproc->isRunning())
          m_vgproc->stop();  /* signal -> processDone() */
    }
 }

[Valgrind-valkyrie-dev] valkyrie: r420 - in trunk: valkyrie valkyrie/core valkyrie/help valkyrie/options valkyrie/tool_utils vk_logmerge

[<sv...@va...>]

Author: cerion
Date: 2008-06-07 15:17:59 +0100 (Sat, 07 Jun 2008)
New Revision: 420

Log:
Fix compile errors for Slackware 10.1
 - #included <stdlib.h> in affected source files.

Thanks to Adam Nielson.




Modified:
   trunk/valkyrie/core/valkyrie_object.cpp
   trunk/valkyrie/help/vk_messages.cpp
   trunk/valkyrie/main.cpp
   trunk/valkyrie/options/vk_option.cpp
   trunk/valkyrie/tool_utils/async_process.cpp
   trunk/valkyrie/tool_utils/vk_process.h
   trunk/vk_logmerge/vklm_main.cpp


Modified: trunk/valkyrie/core/valkyrie_object.cpp
===================================================================
--- trunk/valkyrie/core/valkyrie_object.cpp	2008-06-07 13:50:59 UTC (rev 419)
+++ trunk/valkyrie/core/valkyrie_object.cpp	2008-06-07 14:17:59 UTC (rev 420)
@@ -19,6 +19,8 @@
 #include "vk_option.h"         // PERROR* and friends 
 
 #include <stdio.h>
+#include <stdlib.h>            /* exit */
+
 #include <qapplication.h>
 
 

Modified: trunk/valkyrie/help/vk_messages.cpp
===================================================================
--- trunk/valkyrie/help/vk_messages.cpp	2008-06-07 13:50:59 UTC (rev 419)
+++ trunk/valkyrie/help/vk_messages.cpp	2008-06-07 14:17:59 UTC (rev 420)
@@ -14,6 +14,7 @@
 #include "msgbox_icons.h"
 
 #include <stdarg.h>           /* va_start, va_end */
+#include <stdlib.h>           /* exit */
 
 #include <qapplication.h>
 #include <qobjectlist.h>

Modified: trunk/valkyrie/main.cpp
===================================================================
--- trunk/valkyrie/main.cpp	2008-06-07 13:50:59 UTC (rev 419)
+++ trunk/valkyrie/main.cpp	2008-06-07 14:17:59 UTC (rev 420)
@@ -10,6 +10,7 @@
 #include <qapplication.h>
 #include <qdir.h>
 #include <signal.h>    /* signal name macros, and the signal() prototype */
+#include <stdlib.h>    /* exit */
 
 #include "vk_utils.h"              // vk_assert(), vkPrint() etc
 #include "parse_cmd_args.h"        // parseCmdArgs()

Modified: trunk/valkyrie/options/vk_option.cpp
===================================================================
--- trunk/valkyrie/options/vk_option.cpp	2008-06-07 13:50:59 UTC (rev 419)
+++ trunk/valkyrie/options/vk_option.cpp	2008-06-07 14:17:59 UTC (rev 420)
@@ -13,6 +13,7 @@
 #include <ctype.h>
 #include <math.h>
 #include <malloc.h>
+#include <stdlib.h>         /* exit */
 
 #include <qdir.h>
 

Modified: trunk/valkyrie/tool_utils/async_process.cpp
===================================================================
--- trunk/valkyrie/tool_utils/async_process.cpp	2008-06-07 13:50:59 UTC (rev 419)
+++ trunk/valkyrie/tool_utils/async_process.cpp	2008-06-07 14:17:59 UTC (rev 420)
@@ -19,12 +19,13 @@
  * See the file LICENSE.GPL for the full license details.
  */
 
-#include <unistd.h>                /* dup2(), _exit(),  */
+#include <errno.h>                 /* Fedora Core needs this */
+#include <fcntl.h>                 /* open(), O_RDONLY, O_WRONLY */
+#include <signal.h>
+#include <stdlib.h>                /* exit */
 #include <sys/types.h>             /* waitpid()         */
 #include <sys/wait.h>              /* waitpid()         */
-#include <fcntl.h>                 /* open(), O_RDONLY, O_WRONLY */
-#include <signal.h>
-#include <errno.h>                 /* Fedora Core needs this */
+#include <unistd.h>                /* dup2(), _exit(),  */
 
 #include "async_process.h"
 #include "vk_utils.h"

Modified: trunk/valkyrie/tool_utils/vk_process.h
===================================================================
--- trunk/valkyrie/tool_utils/vk_process.h	2008-06-07 13:50:59 UTC (rev 419)
+++ trunk/valkyrie/tool_utils/vk_process.h	2008-06-07 14:17:59 UTC (rev 420)
@@ -28,6 +28,7 @@
 #include <qsocketnotifier.h>
 
 #include <signal.h>
+#include <stdlib.h>            // exit
 
 
 class VKMembuf

Modified: trunk/vk_logmerge/vklm_main.cpp
===================================================================
--- trunk/vk_logmerge/vklm_main.cpp	2008-06-07 13:50:59 UTC (rev 419)
+++ trunk/vk_logmerge/vklm_main.cpp	2008-06-07 14:17:59 UTC (rev 420)
@@ -9,6 +9,7 @@
 
 #include <sys/types.h>   // getpid
 #include <unistd.h>      // getpid
+#include <stdlib.h>      // exit
 
 #include <qfileinfo.h>
 #include <qdom.h>

[Valgrind-valkyrie-dev] valkyrie: r419 - trunk/valkyrie/options

[<sv...@va...>]

Author: cerion
Date: 2008-06-07 14:50:59 +0100 (Sat, 07 Jun 2008)
New Revision: 419

Log:
Fix crash:
Clicking valgrind options item first time would lead to a
crash -> init() executes updates to the page, which tried to update the page
before it had been setup completely.

Thanks to Pengcheng Zou for the patch.




Modified:
   trunk/valkyrie/options/options_window.cpp


Modified: trunk/valkyrie/options/options_window.cpp
===================================================================
--- trunk/valkyrie/options/options_window.cpp	2008-06-07 13:34:53 UTC (rev 418)
+++ trunk/valkyrie/options/options_window.cpp	2008-06-07 13:50:59 UTC (rev 419)
@@ -218,8 +218,8 @@
 
       /* make sure the item seletion is sync'd */
       m_categories->setSelected( cit->catId(), true );
+      m_wStack->raiseWidget( cit->page() );
       cit->page()->init();
-      m_wStack->raiseWidget( cit->page() );
    }
 
 }

[Valgrind-valkyrie-dev] valkyrie: r418 - trunk

[<sv...@va...>]

Author: cerion
Date: 2008-06-07 14:34:53 +0100 (Sat, 07 Jun 2008)
New Revision: 418

Log:
Better editor support in configure: gvim, gview, kate
Thanks to Pim Nijdamd, Anton Dudarenko



Modified:
   trunk/configure.in


Modified: trunk/configure.in
===================================================================
--- trunk/configure.in	2008-06-07 13:29:10 UTC (rev 417)
+++ trunk/configure.in	2008-06-07 13:34:53 UTC (rev 418)
@@ -69,7 +69,7 @@
 
 # Check for source editor
 AC_ARG_WITH([editor],
-   [  --with-editor=PROGRAM   Source-code editor [[PROGRAM=$EDITOR|emacs|nedit]] ],
+   [  --with-editor=PROGRAM   Source-code editor [[PROGRAM=$EDITOR|emacs|nedit|gvim|gview|kate]] ],
    [ # given
       # TODO: proper way of handling both --xyz=/path/to/foo, and --xyz=foo ?
       if test -x $withval; then
@@ -82,10 +82,14 @@
          AC_MSG_ERROR([Editor '$withval' not found])
       fi
 
-      # if !emacs && !nedit, give warning
+      # if unknown editor, give warning
       VK_ED_NAME=`echo $VK_EDITOR | sed "s/.*\///" | sed "s/-.*//"`
-      if test "$VK_ED_NAME" != "emacs" && test "$VK_ED_NAME" != "nedit"; then
-         AC_MSG_WARN([Unknown editor '$withval'.  If possible, setup an editor flag to go to the target line-number, where %n will be replaced with the line-number.])
+      if test "$VK_ED_NAME" != "emacs" && \
+         test "$VK_ED_NAME" != "nedit" && \
+         test "$VK_ED_NAME" != "gvim" && \
+         test "$VK_ED_NAME" != "gview" && \
+         test "$VK_ED_NAME" != "kate"; then
+         AC_MSG_WARN([Unknown editor '$withval'.  If possible, set editor options in valkyrie to go to the target line-number, e.g. 'gvim +%n', where %n will be replaced with the line-number.])
       fi
    ],
    [ # not given
@@ -100,9 +104,9 @@
             AC_PATH_PROG([VK_EDITOR], [$EDITOR])
 	 fi
       fi
-      #if $EDITOR wasn't set, or failed check, try emacs|nedit
+      #if $EDITOR wasn't set, or failed check, try likely editors
       if test x"$VK_EDITOR" = x ; then
-        AC_PATH_PROGS([VK_EDITOR], [emacs nedit])
+        AC_PATH_PROGS([VK_EDITOR], [emacs nedit gvim kate])
         if test x"$VK_EDITOR" = x ; then
            # Editor not found
            AC_MSG_WARN([No suitable editor found: may be configured at runtime.])
@@ -110,10 +114,17 @@
       fi
    ])
 
-#if emacs|nedit: add 'replacement' flags for line
+#Setup 'line' flag for known editors
 VK_ED_NAME=`echo $VK_EDITOR | sed "s/.*\///" | sed "s/-.*//"`
-if test "$VK_ED_NAME" = "emacs" || test "$VK_ED_NAME" == "nedit"; then
+if test "$VK_ED_NAME" == "emacs" || \
+   test "$VK_ED_NAME" == "nedit" || \
+   test "$VK_ED_NAME" == "gvim"  || \
+   test "$VK_ED_NAME" == "gview"; then
    VK_EDITOR="$VK_EDITOR +%n"
+else
+   if test "$VK_ED_NAME" == "kate"; then
+      VK_EDITOR="$VK_EDITOR --line %n"
+   fi
 fi
 
 

[Valgrind-valkyrie-dev] valkyrie: r417 - in trunk/valkyrie: . tool_utils

[<sv...@va...>]

Author: cerion
Date: 2008-06-07 14:29:10 +0100 (Sat, 07 Jun 2008)
New Revision: 417

Log:
Cleanup compiler warnings (patch from Diego Petten?\195?\178)



Modified:
   trunk/valkyrie/tool_utils/vglog.cpp
   trunk/valkyrie/vk_file_utils.cpp


Modified: trunk/valkyrie/tool_utils/vglog.cpp
===================================================================
--- trunk/valkyrie/tool_utils/vglog.cpp	2006-08-15 15:50:51 UTC (rev 416)
+++ trunk/valkyrie/tool_utils/vglog.cpp	2008-06-07 13:29:10 UTC (rev 417)
@@ -863,7 +863,7 @@
 {
    VgErrCounts mErrCounts = errorcounts();
 
-   vklmPrint( 2, "--- update matches (n=%d) --- ", sErrors.count());
+   vklmPrint( 2, "--- update matches (n=%zu) --- ", sErrors.count());
 
    /* --- find matches: update master err, delete slave err ---  */
 
@@ -927,7 +927,7 @@
       vklmPrint( 2, " ");
    }
 
-   vklmPrint( 2, "--- append non-matches (n=%d) --- ", sErrors.count() );
+   vklmPrint( 2, "--- append non-matches (n=%zu) --- ", sErrors.count() );
 
    /* if no errcounts, and sErrors > 0, create empty errcounts */
    if ( mErrCounts.isNull() && sErrors.count() > 0 ) {
@@ -965,7 +965,7 @@
 
 bool VgLog::mergeLeakErrors( VgErrorList sLeakErrors )
 {
-   vklmPrint( 2, "--- update matches (n=%d) ---", sLeakErrors.count() );
+   vklmPrint( 2, "--- update matches (n=%zu) ---", sLeakErrors.count() );
 
    /* --- for each leak in master ---  */
    VgErrorList errs = leaks();
@@ -1001,7 +1001,7 @@
       vklmPrint( 2, " ");
    }
 
-   vklmPrint( 2, "--- append non-matches (n=%d) ---", sLeakErrors.count() );
+   vklmPrint( 2, "--- append non-matches (n=%zu) ---", sLeakErrors.count() );
 
    /* --- append remaining slave leaks to master --- */
    VgErrorList::Iterator sIt;

Modified: trunk/valkyrie/vk_file_utils.cpp
===================================================================
--- trunk/valkyrie/vk_file_utils.cpp	2006-08-15 15:50:51 UTC (rev 416)
+++ trunk/valkyrie/vk_file_utils.cpp	2008-06-07 13:29:10 UTC (rev 417)
@@ -3021,7 +3021,7 @@
       if ( d->paths->hasFocus() )
          d->ignoreNextKeyPress = true;
 
-		vkWarn( this, "Error", op->protocolDetail().latin1() );
+		vkWarn( this, "Error", "%s", op->protocolDetail().latin1() );
       switch ( op->errorCode() ) {
       case QNetworkProtocol::ErrParse:
       case QNetworkProtocol::ErrValid:

[Valgrind-valkyrie-dev] valkyrie: r416 - trunk/valkyrie

[<sv...@va...>]

Author: cerion
Date: 2006-08-15 16:50:51 +0100 (Tue, 15 Aug 2006)
New Revision: 416

Log:
This one has slipped through for ages!
build using config test result of QT_DIR, not env QTDIR!



Modified:
   trunk/valkyrie/Makefile.am


Modified: trunk/valkyrie/Makefile.am
===================================================================
--- trunk/valkyrie/Makefile.am	2006-07-15 10:14:39 UTC (rev 415)
+++ trunk/valkyrie/Makefile.am	2006-08-15 15:50:51 UTC (rev 416)
@@ -142,7 +142,7 @@
 # Note: if need UIC stuff, see autoqt docs
 
 # Compile flags
-valkyrie_CXXFLAGS = -pipe -O0 -g -W -Wall -Wno-non-virtual-dtor -Wundef -Wredundant-decls -Wno-unreachable-code -fno-inline -D_REENTRANT -I$(srcdir)/help -I$(srcdir)/core -I$(srcdir)/options -I$(srcdir)/tool_views  -I$(srcdir)/tool_utils -I$(srcdir)/../icons -I/usr/X11R6/include -I$(QTDIR)/mkspecs/linux-g++ $(QT_CXXFLAGS)
+valkyrie_CXXFLAGS = -pipe -O0 -g -W -Wall -Wno-non-virtual-dtor -Wundef -Wredundant-decls -Wno-unreachable-code -fno-inline -D_REENTRANT -I$(srcdir)/help -I$(srcdir)/core -I$(srcdir)/options -I$(srcdir)/tool_views  -I$(srcdir)/tool_utils -I$(srcdir)/../icons -I/usr/X11R6/include -I$(QT_DIR)/mkspecs/linux-g++ $(QT_CXXFLAGS)
 
 # Link flags
 valkyrie_LDADD := $(QT_LIBS) -lfreetype -lm -lpthread

[Valgrind-valkyrie-dev] valkyrie: r415 - trunk

[<sv...@va...>]

Author: cerion
Date: 2006-07-15 11:14:39 +0100 (Sat, 15 Jul 2006)
New Revision: 415

Log:
vk req's update



Modified:
   trunk/README


Modified: trunk/README
===================================================================
--- trunk/README	2006-07-13 10:53:17 UTC (rev 414)
+++ trunk/README	2006-07-15 10:14:39 UTC (rev 415)
@@ -13,15 +13,19 @@
   svn co svn://svn.valgrind.org/valkyrie/trunk valkyrie
 
 
+Requirements
+~~~~~~~~~~~~
+Qt (v3.2.0 to v3.3.6) <http://www.trolltech.com>
+Valgrind (>=v3.0.0) <http://www.valgrind.org>
+
+
 Executive Summary
 ~~~~~~~~~~~~~~~~~
-Valkyrie is a GPL'd graphical front-end to the Valgrind suite of tools
-for debugging and profiling x86-Linux programs.  
+Valkyrie is a GPL'd graphical front-end to the Valgrind suite
+of tools for debugging and profiling programs.
 
-Valkyrie uses the Qt widget library <http://www.trolltech.com>, and
-(should) run on any platform that 
-  (a) Qt (>= v3.2.0) runs on, and
-  (b) Valgrind (>= v3.0.0) runs on.
+Valkyrie uses the Qt widget library, and (should) run on any platform
+that both Qt and Valgrind run on.
 
 To get the most out of Valkyrie, you must read the Valgrind
 documentation: <http://www.valgrind.org/docs/>

[Valgrind-valkyrie-dev] valkyrie: r414 - trunk/valkyrie/tool_views

[<sv...@va...>]

Author: cerion
Date: 2006-07-13 11:53:17 +0100 (Thu, 13 Jul 2006)
New Revision: 414

Log:
add check for reading empty <errorcounts> tags
more backward compatibility for qt pre v3.3.0



Modified:
   trunk/valkyrie/tool_views/memcheck_view.cpp
   trunk/valkyrie/tool_views/memcheck_view.h
   trunk/valkyrie/tool_views/vglogview.cpp


Modified: trunk/valkyrie/tool_views/memcheck_view.cpp
===================================================================
--- trunk/valkyrie/tool_views/memcheck_view.cpp	2006-07-13 10:15:47 UTC (rev 413)
+++ trunk/valkyrie/tool_views/memcheck_view.cpp	2006-07-13 10:53:17 UTC (rev 414)
@@ -91,8 +91,13 @@
    connect( lView, SIGNAL(selectionChanged()),
             this,  SLOT(itemSelected()) );
    /* launch editor with src file loaded */
+#if (QT_VERSION-0 >= 0x030200)
    connect( lView, SIGNAL(doubleClicked(QListViewItem*, const QPoint&, int)),
-            this,  SLOT(launchEditor(QListViewItem*, const QPoint&, int)) );
+            this,  SLOT(launchEditor(QListViewItem*)) );
+#else // QT_VERSION < 3.2
+   connect( lView, SIGNAL(doubleClicked(QListViewItem*)),
+            this,  SLOT(launchEditor(QListViewItem*)) );
+#endif
 
 //zz   QFont clientout_fnt( "Adobe Courier", 9, QFont::Normal, false );
 //zz   clientout_fnt.setStyleHint( QFont::TypeWriter );
@@ -302,7 +307,7 @@
 
 /* checks if itemType() == SRC.  if true, and item isReadable or
    isWriteable, launches an editor with the source file loaded */
-void MemcheckView::launchEditor( QListViewItem* lv_item, const QPoint&, int )
+void MemcheckView::launchEditor( QListViewItem* lv_item )
 {
    if ( !lv_item || !lv_item->parent() ) return;
 

Modified: trunk/valkyrie/tool_views/memcheck_view.h
===================================================================
--- trunk/valkyrie/tool_views/memcheck_view.h	2006-07-13 10:15:47 UTC (rev 413)
+++ trunk/valkyrie/tool_views/memcheck_view.h	2006-07-13 10:53:17 UTC (rev 414)
@@ -53,7 +53,7 @@
    void openAllItems(bool);
    void openOneItem();
    void showSrcPath();
-   void launchEditor(  QListViewItem*, const QPoint&, int );
+   void launchEditor(  QListViewItem* );
 
 private:
    QListView* lView;

Modified: trunk/valkyrie/tool_views/vglogview.cpp
===================================================================
--- trunk/valkyrie/tool_views/vglogview.cpp	2006-07-13 10:15:47 UTC (rev 413)
+++ trunk/valkyrie/tool_views/vglogview.cpp	2006-07-13 10:53:17 UTC (rev 414)
@@ -780,6 +780,8 @@
       break;
    }
    case VgElement::ERRORCOUNTS: {
+      if (elem.childNodes().count() == 0)  // ignore empty errorcounts
+         break;
       VgErrCounts ec = (VgErrCounts&)elem;
       /* update topStatus */
       topStatus->updateErrorCounts( ec );

[Valgrind-valkyrie-dev] valkyrie: r413 - trunk/valkyrie/help

[<sv...@va...>]

Author: cerion
Date: 2006-07-13 11:15:47 +0100 (Thu, 13 Jul 2006)
New Revision: 413

Log:
change busy cursor type to stay compatible with qt pre v3.3.0



Modified:
   trunk/valkyrie/help/hand_book.cpp


Modified: trunk/valkyrie/help/hand_book.cpp
===================================================================
--- trunk/valkyrie/help/hand_book.cpp	2006-06-07 12:53:10 UTC (rev 412)
+++ trunk/valkyrie/help/hand_book.cpp	2006-07-13 10:15:47 UTC (rev 413)
@@ -86,7 +86,7 @@
 bool VkTextBrowser::launch_browser(const QString& link)
 {
    bool ok = false;
-   QApplication::setOverrideCursor(Qt::BusyCursor);
+   QApplication::setOverrideCursor(Qt::WaitCursor);
    
    /* try config::vk::browser */
    QString rc_browser = vkConfig->rdEntry( "browser", "valkyrie" );

[Valgrind-valkyrie-dev] valkyrie: r412 - trunk/m4

[<sv...@va...>]

Author: sewardj
Date: 2006-06-07 13:53:10 +0100 (Wed, 07 Jun 2006)
New Revision: 412

Log:
Add license info.


Modified:
   trunk/m4/bnv_have_qt.m4


Modified: trunk/m4/bnv_have_qt.m4
===================================================================
--- trunk/m4/bnv_have_qt.m4	2006-06-02 16:47:32 UTC (rev 411)
+++ trunk/m4/bnv_have_qt.m4	2006-06-07 12:53:10 UTC (rev 412)
@@ -80,6 +80,11 @@
 dnl @version 2006-03-12
 dnl @license AllPermissive
 
+dnl Copyright (C) 2006 Bastiaan Veelo <Bas...@Ve...> 
+dnl Copying and distribution of this file, with or without modification,
+dnl are permitted in any medium without royalty provided the copyright 
+dnl notice and this notice are preserved.
+
 dnl Copyright (C) 2001, 2002, 2003, 2005, 2006 Bastiaan Veelo
 
 dnl THANKS! This code includes bug fixes and contributions made by:

Re: [Valgrind-valkyrie-dev] [Valgrind-users] XML output for Helgrind?

[Cerion Armour-B. <ce...@ke...>]

Hi Diego,

Sorry for the slow responses. We're finally pushing Valkyrie forward though, 
and expect a beta minor release next week. This will support the latest 
Valgrind release.

We're also starting to think more about the next stage for Valkyrie, and 
people are prodding at porting it to QT4, and wondering whether this implies 
a re-write...

I'm wondering if you made any progress with your Ruby idea?  I don't know much 
about Ruby, but I've heard lots of praise for it.

Something important is what the next generation tool will need to be able to 
support: namely, the other valgind tools: cachegrind, helgrind etc., and what 
kind of functionality it therefore must be able to support: graphics, 
whatever. Is Ruby a better choice than QT? Are you for Ruby because you know 
it better than QT, or does it offer particular advantages?

Cheers,
Cerion


On Monday 17 Dec 2007, Diego 'Flameeyes' Pettenò wrote:
> Hi,
>
> I started trying to work on a Ruby frontend for Valgrind, hoping that
> the most difficult task was to parse the XML properly, but it seems like
> only memcheck supports XML output.
>
> Now the simplest way is to write a parser of the helgrind output, but
> I'd see as more "proper" way to actually implement XML output in
> helgrind, if that's acceptable.
>
> So the question is, if I spend some time to implement XML output in
> Helgrind, would it be acccepted and maintained afterward (for instance
> if new output is added)?
>
> I can't promise anything, but as I consider valgrind an important tool
> in xine's development, I'd gladly take some time off xine to work on it.

Re: [Valgrind-valkyrie-dev] sugesstions...

[Cerion Armour-B. <ce...@ke...>]

Thanks for the feedback Anton,

Very clear now what's going on with 3i,iii, cheers - I should be able to sort 
those out.
Thanks for the playback - good in principle, but I'm afraid I'd rather 
discourage this... screenshots and text should be enough. I don't like 
opening strange attachments, and the size makes it unsuitable for the dev 
list.

2. is somewhat tougher... I don't see a nice general way to solve this one. 
I'll ponder some more.

Cheers,
Cerion


On Monday 09 Jun 2008, Anton Dudarenko wrote:
> 1. Thanks!!!
> 2. Our sources sometimes became unavailable because sometimes servers which
> keep this files are turning off )))
> And in this case I would like to use my local version of code - so I have
> to use the path not from debug info but from my point of view. And so I
> want to still be able to open errors messages even from inaccessible files
> )) 3i Maybe you have a latest development build - not final available
> 1.2.0?? But in any way - I've created a small avi and a log-file - all is
> 7zipped - see attach.
> In this movie)) you can see all "bugs".
>
> In the beginning of movie I am showing to you 3i bug behavior.
> I am pressing but. "'Open|Close errors'" and only SCP and UNC was
> unfolded))
>
> After that I am showing to you behavior of but. "'Open|Close the selected
> item'" - bug 3iii - It becomes available ONLY if I was opening 1 tree node
> by mouse and selected some subitems(function`s call stack) in this tree
> node..
>
> After that I am again showing to you 3i bug BUT with some new unexpected
> behavior - I've closed all nodes by mouse but button still is still visible
> as pushed. Even when I opened/closed node by "arrow key" buttons from the
> keyboard.
>
> After that bug 3iii again was shown. Now I was used "arrow key" buttons
> from the keyboard. And a mouse again))
>
> Hope this will help)))
>
> -----Original Message-----
> From: Cerion Armour-Brown [mailto:ce...@ke...]
> Sent: Saturday, June 07, 2008 8:33 PM
> To: val...@li...
> Cc: Anton Dudarenko
> Subject: Re: [Valgrind-valkyrie-dev] sugesstions...
>
> Thanks for the feedback Anton.
>
> We're busy now fixing some bugs and supporting the latest valgrind
> release...
>
> 1. Fixed
> 2. I don't really understand why this is necessary: do you move your
> sources
>
> after building?
> 3i. ?! I can't reproduce this - can you send me the logfile pls?
> 3ii. Well spotted :-) I'll add it to the todo list
> 3iii. I don't understand what you mean, can you clarify?
> 4. I've added more options for the configure script - adding the
> functionality
> in the valkyrie options is too much work for now: maybe in the future :-)
>
> Regards,
> Cerion
>
> On Tuesday 03 Jun 2008, Anton Dudarenko wrote:
> > Your tool is simply excellent, but I would like to notice something in
> > its behaviour and to suggest improving it.
> >
> > 1. To include horizontal scrolling for a LOG window. Otherwise it is
> > impossible to see the full text of the memory check`s messages of an
>
> error.
>
> > 2. To include possibility to open on double click a file even if it is
> > certain by you as inaccessible. The obvious decision - simply to display
> > a inquiry window with the text "On the given way a file it is not
> > accessible - whether YOU wish to specify its place independently?" And if
> > the user
>
> has
>
> > specified it - to use the received way further and as consequence -
>
> already
>
> > at once to open an editor window with the specified file on the specified
> > line at double click...
> >
> > 3. To "cure" behaviour of buttons 'Open|Close errors' and 'Open|Close the
> > selected item'. In the current version they work not correctly.
> >
> > 3i. The first
> > command opens a tree only for SysCallParam and UnInitCondition nodes. The
> > others remain folded.
> >
> > 3ii. And it remains turned ON (the button visually is
> > pressed) even if I have folded after that all nodes by means of "arrow
>
> key"
>
> > buttons on the keyboard.
> >
> > 3iii.
> > The second command to become accessible to the chosen tree node of an
>
> error
>
> > messages only if this node has been preliminary opened and will fold
> > again by mouse or by means of "arrow key" buttons on the keyboard.
> >
> > 4. In a window of program properties at a field of editor choice well to
> > provide auto filling of a command line for 2-4 widespread editors
> >
> > gvim +%n
> >
> > vi +%n
> >
> > vimx +%n
> >
> > kate --line %n
> >
> > etc....
> >
> > And for the KATE editor it is necessary on one to reduce number of a line
> > on which it will be necessary to move in the opened editor. It is an
> > error of this editor itself that at the jump from a command line for the
> > line 5, for example, actually it passes on 6.

Re: [Valgrind-valkyrie-dev] Valkyrie status

[Cerion Armour-B. <ce...@ke...>]

Amit,

I forgot it was you who was looking into the port to QT4.
How's that going?

I believe it may be a fair bit of work, and we may well want to take advantage 
of lessons learnt from the first version, and take it a step further...

We really need to think about requirements and useability with respect to 
Cachegrind, Helgrind, as well as Memcheck, and have all that clear before we 
just simply port it...

We need to hear from the other users out there what they love/hate/would like 
to see, etc...

Valgrind developers may be interested in adding xml support for other tools, 
and/or other functionality to valgrind to make the interface simpler...

We also need version control on a central repo. Do you already have something 
setup, or shall we set this up on the current svn server? 

What's the current state of play?
Regards,
Cerion


On Tuesday 10 Jun 2008, Cerion Armour-Brown wrote:
> Hi Amit,
>
> Completely fair question - we've had a tough time finding time for
> Valkyrie, and not keeping up with Valgrind is inexcusable :-)
>
> I'm not sure what the future holds for Valkyrie, but for the moment we're
> finally pushing it forward. Beta release should be next week, supporting
> the latest Valgrind, and fixing a whole bunch of bugs.
>
> I think for the time being, we'll try to run it along similar lines to
> Valgrind, so only core developers have repo write priveliges. To become
> part of the core dev team people will need to first send patches and earn
> the trust/respect of the other core dev members.
>
> So, to sum up, the interest is there, but time is (as always) tight, so
> comments, suggestions, fixes, and best of all: patches, are very welcome!
>
> Cheers,
> Cerion
>
> On Wednesday 05 Mar 2008, Amit Perelman wrote:
> > Hello,
> >
> > I am open source developer from Israel. I am interesting in Valkyrie.
> > It seen that currently Valkyrie support only memcheck from valgrind 3.2.3
> > and that the latest release was in 2006. My question is does the
> > development discontinued ? I would like to make Valkyrie build against
> > valgrind 3.3.0 and add support to Cachegrind and Massif.
> > How can we make this happen ? If you lost interest in the project,
> > I am willing to take charge on it.
> >
> > Amit Perelman

Re: [Valgrind-valkyrie-dev] Valkyrie status

[Cerion Armour-B. <ce...@ke...>]

Hi Amit,

Completely fair question - we've had a tough time finding time for Valkyrie, 
and not keeping up with Valgrind is inexcusable :-)

I'm not sure what the future holds for Valkyrie, but for the moment we're 
finally pushing it forward. Beta release should be next week, supporting the 
latest Valgrind, and fixing a whole bunch of bugs.

I think for the time being, we'll try to run it along similar lines to 
Valgrind, so only core developers have repo write priveliges. To become part 
of the core dev team people will need to first send patches and earn the 
trust/respect of the other core dev members.

So, to sum up, the interest is there, but time is (as always) tight, so 
comments, suggestions, fixes, and best of all: patches, are very welcome!

Cheers,
Cerion


On Wednesday 05 Mar 2008, Amit Perelman wrote:
> Hello,
>
> I am open source developer from Israel. I am interesting in Valkyrie.
> It seen that currently Valkyrie support only memcheck from valgrind 3.2.3
> and that the latest release was in 2006. My question is does the
> development discontinued ? I would like to make Valkyrie build against
> valgrind 3.3.0 and add support to Cachegrind and Massif.
> How can we make this happen ? If you lost interest in the project,
> I am willing to take charge on it.
>
> Amit Perelman

Re: [Valgrind-valkyrie-dev] valkyrie-1.20

[Cerion Armour-B. <ce...@ke...>]

Hi Enrique,

Sorry it's taken so long to get back to this. We're finally pushing Valkyrie 
forward - it should be ready for a beta release coming monday.

Couple of questions about your findings...

1. configure does give options for setting the document install directory:
--docdir, --htmldir
But it seems these don't work! I'll try to get them working.
Assuming I can fix this, is this sufficient?

2. Can you still reproduce this? If so, please run from the commandline, and 
send me:
 - anything interesting on stdout/err
 - a copy of the relevant temporary generated xml file in ~/.valkyrie/logs/
   (this is auto-deleted when you close valkyrie, so make a copy of the latest 
file before closing).

3. Save Log is disabled? You mean when the error of (2) is given? I've changed 
the message slightly to say 'close this dialog, and then save log'. Does this 
solve this issue, or do I not understand?

4. Very strange... If there's an error such as dir not writeable, you're 
always given the chance to try again in a different place...
The text "Save failed" doesn't exist in the code. Can you reproduce this and 
tell me _exactly_ what happens?

5. Fixed.

Thanks for your feedback!
Cerion


On Tuesday 25 Dec 2007, you wrote:
> Hi,
>
> I packaged Valkyrie-1.2.0 for Slackware 12/DroplineGnome 2.20.  The
> following configuration was applied:
> prefix = /usr
> sysconfdir = /etc
> localstatedir = /var
>
> Documentation has been moved to /usr/doc/valkyrie-1.2.0/html.
>
> These are my observations and errors encountered thus far:
>
> 1 The path for the Help contents are hard-coded to /usr/doc. Since
> documentation has been moved, the
>   docs are not visible unless the Helpviewer is pointed to the right
> files. There is no way also to set paths.
>
> 2 Valkyrie's Valgrind fails to run on any binary I pointed it to but not
> when I used Valgrind alone. Error
>   messages are:
>
>   XML Parse error
>   Valgrind XML output is incomplete
>
>   Run error
>   Process exited with return value 1. This is likely to simply be the
> client program return value. If, however,
>   you suspect Valgrind itself may have crashed, please 'Save Log' and
> examine for details.
>
> 3 "Save Log" is disabled.
>
> 4 On exit, the Valkyrie prompts to save the output or not. When I click
> "Save", a file dialog opens. I entered
>   the name of the binary file I was testing then another error says
> "Save failed". So I discarded the output.
>
> 5 I also changed the  path on where to save the output but the save
> functionality seems to be hard-coded also
>   to /home/<acct>/.valkyrie/logs.
>
> I am a heavy Valgrind user as the reason I am attracted to Valkyrie. I
> was hoping this app will save me some
> time keeping track of Valgrind's output. I also hope that the feedbacks
> may help you in improving the software.
>
> Merry Christmas!

Re: [Valgrind-valkyrie-dev] sugesstions...

[Cerion Armour-B. <ce...@ke...>]

Thanks for the feedback Anton.

We're busy now fixing some bugs and supporting the latest valgrind release...

1. Fixed
2. I don't really understand why this is necessary: do you move your sources 
after building?
3i. ?! I can't reproduce this - can you send me the logfile pls?
3ii. Well spotted :-) I'll add it to the todo list
3iii. I don't understand what you mean, can you clarify?
4. I've added more options for the configure script - adding the functionality 
in the valkyrie options is too much work for now: maybe in the future :-)

Regards,
Cerion

On Tuesday 03 Jun 2008, Anton Dudarenko wrote:
> Your tool is simply excellent, but I would like to notice something in its
> behaviour and to suggest improving it.
>
> 1. To include horizontal scrolling for a LOG window. Otherwise it is
> impossible to see the full text of the memory check`s messages of an error.
>
> 2. To include possibility to open on double click a file even if it is
> certain by you as inaccessible. The obvious decision - simply to display a
> inquiry window with the text "On the given way a file it is not accessible
> - whether YOU wish to specify its place independently?" And if the user has
> specified it - to use the received way further and as consequence - already
> at once to open an editor window with the specified file on the specified
> line at double click...
>
> 3. To "cure" behaviour of buttons 'Open|Close errors' and 'Open|Close the
> selected item'. In the current version they work not correctly.
> 
> 3i. The first 
> command opens a tree only for SysCallParam and UnInitCondition nodes. The
> others remain folded.
> 
> 3ii. And it remains turned ON (the button visually is 
> pressed) even if I have folded after that all nodes by means of "arrow key"
> buttons on the keyboard.
>
> 3iii.
> The second command to become accessible to the chosen tree node of an error 
> messages only if this node has been preliminary opened and will fold again
> by mouse or by means of "arrow key" buttons on the keyboard.
>
> 4. In a window of program properties at a field of editor choice well to
> provide auto filling of a command line for 2-4 widespread editors
>
> gvim +%n
>
> vi +%n
>
> vimx +%n
>
> kate --line %n
>
> etc....
>
> And for the KATE editor it is necessary on one to reduce number of a line
> on which it will be necessary to move in the opened editor. It is an error
> of this editor itself that at the jump from a command line for the line 5,
> for example, actually it passes on 6.

[Valgrind-valkyrie-dev] sugesstions...

[Anton D. <ad...@na...>]

Your tool is simply excellent, but I would like to notice something in its
behaviour and to suggest improving it.

1. To include horizontal scrolling for a LOG window. Otherwise it is
impossible to see the full text of the memory check`s messages of an error.

2. To include possibility to open on double click a file even if it is
certain by you as inaccessible. The obvious decision - simply to display a
inquiry window with the text "On the given way a file it is not accessible -
whether YOU wish to specify its place independently?" And if the user has
specified it - to use the received way further and as consequence - already
at once to open an editor window with the specified file on the specified
line at double click...

3. To "cure" behaviour of buttons 'Open|Close errors' and 'Open|Close the
selected item'. In the current version they work not correctly. The first
command opens a tree only for SysCallParam and UnInitCondition nodes. The
others remain folded. And it remains turned ON (the button visually is
pressed) even if I have folded after that all nodes by means of "arrow key"
buttons on the keyboard. 

The second command to become accessible to the chosen tree node of an error
messages only if this node has been preliminary opened and will fold again
by mouse or by means of "arrow key" buttons on the keyboard. 

4. In a window of program properties at a field of editor choice well to
provide auto filling of a command line for 2-4 widespread editors

gvim +%n

vi +%n

vimx +%n

kate --line %n

etc....

And for the KATE editor it is necessary on one to reduce number of a line on
which it will be necessary to move in the opened editor. It is an error of
this editor itself that at the jump from a command line for the line 5, for
example, actually it passes on 6.

 

 

Re: [Valgrind-valkyrie-dev] porting Valkyrie to QT4

[Cerion Armour-B. <ce...@ke...>]

On Friday 11 Apr 2008, Amit Perelman wrote:
> Hi All,
>
> In the very near future I want to start porting Valkyrie to QT4.
> If anyone has useful infomation regarding this task I will glad to hear it.


Hi Amit,
"Dinosaur" had a go at this back in 2006, but didn't manage to finish it. Take 
a look at our correspondance below.
I have his code from 2006, but haven't used it yet. I'll send this along in an 
off-list email (is large), hopefully this will be of some use!
Note - the codebase hasn't changed much for quite a while now, so his code 
will still be diff'able with the trunk.

Regards,
Cerion


----------  Forwarded Message  ----------

Subject: Re: Porting Valkyrie to Qt 4
Date: Monday 07 Aug 2006
From: Cerion Armour-Brown <ce...@op...>
To: Dinosaur <din...@us...>

On Saturday 05 August 2006 10:27, Dinosaur wrote:
> Hello!
>
> I got trouble while trying to port Valkyrie to Qt 4 - there's no analog or
> replacement to QCleanupHandler and docs doesn't contain any information
> about this class (particularily, "Porting from Qt 3 to Qt 4" chapter).
>
> The situation is worse than I expected... :(

Ah, you're falling over an experiment: 'FileDialog', taken from Qt3's 
QFileDialog.
Don't worry - you can rip that out.  It's not actually used anywhere yet, and 
we can rewrite it from Qt4's QFileDialog if needed.

If you come across other stuff like this (i.e. code taken from Qt3 that was 
either buggy or didn't have enough functionality), it'd probably be a good 
idea to start from Qt4's version of the same.

Cheers,
Cerion
-------------------------------------------------------



----------  Forwarded Message  ----------

Subject: Re: Valkyrie and Qt4 - IDs
Date: Monday 18 Sep 2006
From: Dinosaur <din...@us...>
To: "Cerion Armour-Brown" <ce...@op...>

> On Sunday 17 September 2006 14:13, you wrote:
>> Hello!
>>
>> Well, I think Valkyrie GUI needs major re-writing to be ported to Qt4 as
>> there're no IDs for menus nor widget stacks. Qt4 menus work directly with
>> actions (QAction) and that causes LOTS of changes in the code (primarily
>> main_window.cpp that seems to need more than 50% to be re-written).
>>
>> I can send the partially ported code if you wish.
>>
>> Thank you.

> Hi Dinosaur,
>
> Great that you've been working on it!
> If you've had enough, sure, send us what you have, and we'll try to get to 
it when we can.
Here it is. Of course, I don't know if it's ported correctly as the 
compilation fails.
There's a new switch in configure script, look at that and, probably, edit it 
as I'm not very familiar with autotools.

> If you're up for a pushing it a bit further, however, here's some 
thoughts...
Yes, I'll continue the work.

> I just had a quick look at QStackedWidget, and I 
see 'currentIndex()', 'indexOf()' etc.  At first glance these look fine to 
replace the old QWidgetStack usage, or have I missed something?
Unfortunatelly, I've had very little expierence with Qt3 so the primary source 
for the information is "Porting from Qt 3 to Qt 4" page of Qt4 docs.

> As for menus, yes, hmm, looks like a bit more work...
> For the tools, looks like we need a QActionGroup with the exclusive property 
set, create the actions for those tools, and add those actions to the menu 
and the lower statusFrame tool buttons...  Plus more messing around to 
disable file->run etc. if there is no current tool window...
>
> (Assuming we can still retain stack ids...)
> Perhaps have a map of toolView_id -> tool_action, so can enable/disable the 
right tool actions in setToggles() ?
> And replacing FILE_RUN etc enums with QActions shouldn't be a problem.
>
> Hth,
> Cerion
While trying to port existing code to the new version of Qt, I still think 
that it'd be better to create new Valkyrie 2 branch made for Qt4 only. All 
those workarounds will make the code (primarily in relatively complex files 
like main_window.cpp with lots of uncompatible code) look terrible and very 
difficult to support. There were no serious code changes for a while, anyway.

-- 
Dinosaur
-------------------------------------------------------

[Valgrind-valkyrie-dev] valkyrie 1.20 crash/bugfix

[Donna R. <do...@va...>]

Am posting this to the list so is visible to all.
The relevant bit of code referred to is in /core/memcheck_object.cpp
and the crash is on line "if ( !m_vgreader->parseContinue()) {"

void Memcheck::readVgLog()
{
  // vkPrint("Memcheck::readVgLog()");
  vk_assert( view !- 0 );
  vk_assert( m_vgreader != 0 );
  vk_assert( m_vgproc != 0 );

  qApp-> processEvents();

  /* Try reading some more data */
  if ( !m_vgreader->parseContinue()) {
    /* Parsing failed: stop m_vgproc, if running */
    if (m_vgproc->isRunning())
      m_vgproc->stop(); /* signal -> rpocessDone() */
  }
}


--------------------------------------------------

valkyrie 1.20 crash/bugfix
 Date: Yesterday 10:23:23
 From: "Gert Steenssens" <gst...@gm...>
 To: su...@op...
 
Hi,

I've been using your great GUI valkyrie for checking some memory leaks with
valgrind,
and have noticed a strange crash when the program I was profiling exited.

valkyrie is parsing valgrind output and all of a sudden crashes.
The exact location of the crash you can see on the image attached when I ran
it with KDBG.
Turns out it's due to a NULL pointer deref...

To resolve it I just added a check to see if  'vgreader' was not a NULL
pointer before it is referenced.
which resolved my problem though I haven't searched for the exact reason of
the crash...

Just thought you might like to know... ;)

and thanks for valkyrie!

greetings,
Gert

[Valgrind-valkyrie-dev] porting Valkyrie to QT4

[Amit P. <ami...@gm...>]

Hi All,

In the very near future I want to start porting Valkyrie to QT4.
If anyone has useful infomation regarding this task I will glad to hear it.

-- 
Amit Perelman

Wovon man nicht sprechen kann, 
darüber muß man schweigen.

[Valgrind-valkyrie-dev] 2nd test to list as subscriber

[Donna R. <do...@ke...>]

This is test #2

[Valgrind-valkyrie-dev] testing valkyrie mailing list

[Donna R. <do...@va...>]

xxx

-- 
No trees were destroyed in the sending 
of this contaminant-free message.
However, we do concede that a significant number 
of electrons may have been inconvenienced.