News:

The new Release 25.03 is out! You can download binaries for Windows and many major Linux distros here .

Main Menu

[Ubuntu] "instanciated from here" build message not clickable

Started by Kalith, July 27, 2011, 02:39:47 PM

Previous topic - Next topic

Kalith

Hi,

In the build messages panel, whenever I get "instantiated from here" messages, for example (taken from the build log) :
In file included from ../Frost/Utils/include/frost_utils_types.h:109:0,
                 from ../Frost/Utils/include/frost_utils.h:18,
                 from include/blockchunk.h:4,
                 from /home/kalith/Programmation/Mycraft/src/blockchunk.cpp:1:
../Frost/Utils/include/frost_utils_ptr.h: In member function 'Frost::s_ptr<T>::operator Frost::s_ptr<N>() [with N = Frost::Block, T = const Frost::Block]':
/home/kalith/Programmation/Mycraft/src/blockchunk.cpp:572:75:   instantiated from here
../Frost/Utils/include/frost_utils_ptr.h:317:36: error: invalid conversion from 'const Frost::Block*' to 'Frost::Block*'
../Frost/Utils/include/frost_utils_ptr.h:317:36: error:   initializing argument 1 of 'Frost::s_ptr<T>::s_ptr(T*) [with T = Frost::Block]'


... I can only click on the last two error messages (the actual error), and not on the one just before, the "instantiated from here" one.
The actual error location is of no use to me, as it points to a smart pointer header. I'm interested in the root of the problem (/home/kalith/Programmation/Mycraft/src/blockchunk.cpp:572), but the build messages panel doesn't allow me to reach it.

I have to go there manually and, most of the time, to do so I have to expand the "file" tab because the file path is too long.

Can it be fixed somehow ?

oBFusCATed

(most of the time I ignore long posts)
[strangers don't send me private messages, I'll ignore them; post a topic in the forum, but first read the rules!]

Kalith

If I understand correctly, there are two variants of this message, with and without the character position ?
If so, I know very little of regex, but the following seems to work correctly for both cases :
([][{}() \t#%$~A-Za-z0-9_:+/\.-]+):([0-9]+):[0-9\:][ \t]+([iI]nstantiated from .*)

oBFusCATed

Kalith:
I'll try it later today. The thing is that newer compilers can generate the :x:y message,
the older ones generate only the :x message, so we should support both variants.
(most of the time I ignore long posts)
[strangers don't send me private messages, I'll ignore them; post a topic in the forum, but first read the rules!]

Kalith


oBFusCATed

No, I've forgotten about it. Adding to the todo, so I can look at it next week.
(most of the time I ignore long posts)
[strangers don't send me private messages, I'll ignore them; post a topic in the forum, but first read the rules!]

oBFusCATed

I've found some time to test the reg expression and it doesn't work.
Have you tried it with the regexp testbed plugin?

I've an expression which works, but it requires changes to the code, because it returns the line number at two different positions for the two cases. :(
(most of the time I ignore long posts)
[strangers don't send me private messages, I'll ignore them; post a topic in the forum, but first read the rules!]

Jenna

Can you please test these two regexes:
Both have to be included before (!) the existing Compiler warning and before the existing Compiler error.
Only tested on linux with gcc4.4 to 4.6 (only 4.5 and 4.6 lead to  the issue.

Compiler warning (2):
([][{}() \t#%$~[:alnum:]&_:+/\.-]+):([0-9]+):[0-9]+:[ \t]([Ww]arning:[ \t].*)

Compiler error (2)
([][{}() \t#%$~[:alnum:]&_:+/\.-]+):([0-9]+):[0-9]+:[ \t](.*)

Line number etc. are the same as for the existing regexes.