Changes committed to git://oss.sgi.com/markgw/pcp/pcp-gui.git dev
note----------------^^^^^^
Nathan: you'll have to fix the perms in /oss/git/pcp/pcp-gui.git
some of the subdirs below 'objects' are owned by you. In the future
perhaps we should push as the 'pcp' user (add our public keys
to pcp's authorized_keys file first).
This merges fche's "QString toAscii decorruption" fixes from
pcpfans git://sourceware.org/git/pcpfans.git fche/pcp-gui
and also bumps to version 1.5.7-1 (under development).
commit b43414f5b41e7ca7bc7b826e74dbab955de6b014
Author: Mark Goodwin <mgoodwin@xxxxxxxxxx>
Date: Fri Apr 12 15:50:42 2013 +1000
Bump to 1.5.7-1 (under development)
modified: VERSION
modified: build/rpm/fedora.spec
modified: build/rpm/pcp-gui.spec.in
commit df82a4b54038dee6b00e54f38b53b58267ef4cc8
Author: Mark Goodwin <mgoodwin@xxxxxxxxxx>
Date: Thu Apr 11 09:05:34 2013 -0400
pcp-gui qa fix: add missing default-argument#3 to QmcSource::getSource
commit c407d0107b50b94b9d893687110d898f87c935ae
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Sun Apr 7 14:55:23 2013 -0400
QString toAscii decorruption
A recurrent coding pattern in the code aimed to get C "const char*"
pointers out of QString objects, with code that looks like this:
(const char*) object.toAscii()
This is OK, using the QByteArray conversion-operator, as long as the
value is used before the QByteArray temporary is itself deleted.
However, in a few cases, the value is instead stored, or returned,
leading to corruption and deceit. These have been reworked:
* src/chart/chartdialog.cpp (createChartPlot): strdup the nlabel.
* src/chart/colorscheme.cpp (colorSpec): Inline the rgbi string.
* src/chart/exportdialog.cpp (exportFile): Inline the strcmp args.
* src/chart/view.cpp (saveChart): Simplify and inline metric
name/instance strings.
(saveView): strdup incoming file name.
* src/libqmc/qmc_metric.h (*Ascii()): strdup return value.
* src/libqmc/qmc_source.h (*Ascii()): strdup return value.
commit 4e7f2c9c392f78d171fdee596898d5bc92421fa6
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Sun Apr 7 17:22:25 2013 -0400
Give ChartEngine a dummy virtual destructor
That way, the compiler doesn't complain about "delete aChargEnginePtr".
commit fbbf4fd69302f29243b9428dab97f5482c1813a4
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Sun Apr 7 15:44:38 2013 -0400
Revert "QString toAscii decorruption"
This reverts commit b6b7de7ce33eded33f18a6d1d89c33a0d373957a.
We may need to only handle those cases where the returned
(const char*) qstring.toAscii()
value is saved across statements.
commit b6b7de7ce33eded33f18a6d1d89c33a0d373957a
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Sun Apr 7 14:55:23 2013 -0400
QString toAscii decorruption
A recurrent coding pattern in the code aimed to get C "const char*"
pointers out of QString objects, with code that looks like this:
(const char*) object.toAscii()
This is wrong on several levels. In Qt4 and probably earlier,
toAscii() returns a (temporary) QByteArray object, not a C string.
Any casting to hide this leads to direct junk, e.g., failure to save
pmchart view state to requested file names, or with actual readable
metric names. We switch to
object.toAscii().constData()
throughout. This too is imperfect, as the QByteArray object is
temporary, and will be deallocated at the next statement (;) boundary,
which makes the const char*'s contain junk again. So, cases where
such pointers are stored beyond a statement need to be strdup()'d.
commit 9162d8c74e66b330096195ca8ff38d43bd541989
Author: Frank Ch. Eigler <fche@xxxxxxxxxx>
Date: Sun Apr 7 14:55:04 2013 -0400
add emacs .dir-local.el for formatting preferences
|