CodeLite compiling - c++11, wxW294, std::string

[scottfuz]

Downloaded latest sources from SVN (5928).
Compiling on Debian Wheezy
Output of gcc -v

Code: Select all

Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.7/lto-wrapper
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Debian 4.7.1-7' --with-bugurl=file:///usr/share/doc/gcc-4.7/README.Bugs --enable-languages=c,c++,go,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-4.7 --enable-shared --enable-linker-build-id --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.7 --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-gnu-unique-object --enable-plugin --enable-objc-gc --with-arch-32=i586 --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 4.7.1 (Debian 4.7.1-7) 

I had compiled and install wx294 locally. However, I had used a couple of local changes when compiling wx294 that seemed to have a rather pronounced effect on CodeLite.
1) I had used -std=c++11 in the CXXFLAGS
2) I used "--enable-stl --enable-std_string_conv_in_wxstring"

Getting CodeLite configured and to compile was "entertaining".

The configure went without a hitch. But make would immediately fail with an error message complaining about the missing "-std=" tag that had been used on wx294 compile. Each time make would fail, I would end up changing the configure file to include the "-std=" directive for a particular component of CodeLite (changes in configure were made in gcc, clang, indexer sections).

The compile would also halt on error complaining about missing a reference to wxPrintf and wxFPrintf. I think these missing references are a consequence of the changes upstream in wx294 (inclusion of Unicode - see wxW Unicode Overview) and my enabling std::string. I could find no references to wxPrintf/wxFPrintf in the wx294 documentation (but wxprintf.cpp exists in the source). I ended up changing the CodeLite sources by replacing the wxPrintf/wxFPrintf statement with its C equivalent in a few places to complete the compile:

• codelite/CodeLite/cpptoken.cpp (line 52),
  codelite/CodeLite/file_logger.cpp (line 68),
  codelite/Plugin/macrosdlg.cpp (line175),
  codelite/codelite_clang/clang.cpp (4x line 24,220,251,300)
  codelite/UnitTestCPP/unittestcppoutputparser.cpp (lines 150 - 152)

There was one other issue with the compile that cropped up. In the file:
codelite/sdk/wxshapeframework/src/wxxmlserializer/PropertyIO.cpp (line 165)

Code: Select all

if( wxIsNaN(value) )

g++ complained about something below this call (ambiguity with isnan(double&)).
I ended up changing this line to

Code: Select all

if( std::isnan(value) )

. Could this be an upstream problem?

So far, the compile has proceed with only a few warnings here and there (unused or uninitialized variables).
Let's see what happens when the compile finishes and can it execute?
--- more to follow ---

[scottfuz]

==== SUCCESS ====
Build and Install Completed!

A couple of issues jumped out during the compile...
codelite/DatabaseExplorer/SqlCommandPanel.cpp (line 210)

Code: Select all

m_gridTable->SetCellValue(value ,rows, i-1);

was flagged as an error during compiling. The wx294 documentation shows the declaration for

Code: Select all

void wxGrid::SetCellValue(const wxString& val, int row, int col )

as deprecated.
Line was replaced with new version

Code: Select all

m_gridTable->SetCellValue(rows, i-1,value);

.

codelite/CallGraph/uicallgraphpanel.cpp (line 107)

Code: Select all

if( m_grid->GetRows() ) m_grid->DeleteRows(0, m_grid->GetRows());

was also flagged during compiling. wxGrid::GetRows is not shown in the documentation. Could this instead be a reference to wxGrid::GetNumberRows? If so, the line becomes

Code: Select all

if( m_grid->GetNumberRows() ) m_grid->DeleteRows(0, m_grid->GetNumberRows());

. But if the number of rows is zero-based counting, should the range for DeleteRows be...

Code: Select all

m_grid->DeleteRows(0, (m_grid->GetNumberRows() - 1));

?

One last issue I noticed...
The install did put a desktop file in place but no icon appears for the menu item. It appears from looking at the codelite.desktop file that the $prefix path was not generated properly for the Icon entry. Line 5 of the desktop file:

Code: Select all

Icon=/usr/share/codelite/images/cubes.png

is the "default" install/prefix path. But if the user puts a different $prefix value during configure, then I suspect this is not picked up during the install of the desktop file. Real minor and easy to fix.

Nice stuff!!!

[eranif]

can you please place all your fixes into a patch file and send it to me (or attach it here)

Thanks,
Eran

[scottfuz]

Eran,
Attaching file...

A couple of caveats:
- I got a warning about using the directive compiling a C file. I may have put an extra "-std=" directive where it doesn't need to be.
- The patch file is rather "polluted" because I had a setting enabled on Geany to strip extra spaces on the ends of lines
- includes change for m_grid->DeleteRows(0, m_grid->GetNumberRows() - 1)

So, question...
Why use fprint instead of cout?

[scottfuz]

I'm trying to tracked down the line in the configure file that causes the link to the icon file to be incorrect.
configure (line 1230)

Code: Select all

sed s@%%PREFIX%%@$prefix@g Runtime/codelite.desktop.template >> Runtime/codelite.desktop

If I get this (big maybe - haven't explored sed much), its doing a replacement of a placeholder with the actual value.
But something is going wrong here....the default value of $prefix is being used.

...and if I look up...way, way up....
configure (line 1190)

Code: Select all

prefix="/usr"

overwrites the existing value... and does not correct for users selection

Edit---
if we move the sed lines below the if/else statements, then prefix should be set properly and usable.
Attaching an update patch file...

[scottfuz]

...
I had compiled and install wx294 locally. However, I had used a couple of local changes when compiling wx294 that seemed to have a rather pronounced effect on CodeLite.
1) I had used -std=c++11 in the CXXFLAGS
2) I used "--enable-stl --enable-std_string_conv_in_wxstring"
...
The compile would also halt on error complaining about missing a reference to wxPrintf and wxFPrintf. I think these missing references are a consequence of the changes upstream in wx294 (inclusion of Unicode - see wxW Unicode Overview) and my enabling std::string. I could find no references to wxPrintf/wxFPrintf in the wx294 documentation (but wxprintf.cpp exists in the source). I ended up changing the CodeLite sources by replacing the wxPrintf/wxFPrintf statement with its C equivalent in a few places to complete the compile:

• codelite/CodeLite/cpptoken.cpp (line 52),
  codelite/CodeLite/file_logger.cpp (line 68),
  codelite/Plugin/macrosdlg.cpp (line175),
  codelite/codelite_clang/clang.cpp (4x line 24,220,251,300)
  codelite/UnitTestCPP/unittestcppoutputparser.cpp (lines 150 - 152)

There was one other issue with the compile that cropped up. In the file:
codelite/sdk/wxshapeframework/src/wxxmlserializer/PropertyIO.cpp (line 165)

Code: Select all

if( wxIsNaN(value) )

g++ complained about something below this call (ambiguity with isnan(double&)).
I ended up changing this line to

Code: Select all

if( std::isnan(value) )

. Could this be an upstream problem?

The common threads here are wxString, using the stl compile directive in wxWidgets and this is all local to my machine.
Probably a usage issue on my end. I'll punt this over to wxWidgets Forum and see what the guru's there have to say.

Thanks for the help!