-d inlines doesn't work

[octoploid]

~ % echo "int main(){}" | g++ -g -x c++ - && bloaty -d inlines ./a.out
bloaty: internal error, specified same DWARF attribute more than once

[haberman]

Can you attach your a.out? I can't reproduce this. On my system I get:

# uname -a
Linux 2e6f906813a8 4.4.20-moby #1 SMP Thu Sep 15 12:10:20 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
# echo "int main(){}" | g++ -g -x c++ - && ./bloaty -d inlines ./a.out
     VM SIZE                 FILE SIZE
 --------------           --------------
 100.0%  4.00Mi <stdin>:1  1.24Ki  13.7%
   0.0%     996 [None]     7.83Ki  86.3%
 100.0%  4.00Mi TOTAL      9.07Ki 100.0%

Those results are suspect, but it's not the same as the error you got.

[octoploid]

a.out.gz

[froydnj]

I saw the same error after running bloaty for over an hour on Firefox's libxul.so.

[haberman]

Thanks for the reports -- I've been able to reproduce this now, and am working on a fix as we speak.

[octoploid]

Thanks for the fix. However it now produces "bogus" output for me:

 % ./bloaty -d inlines bloaty
     VM SIZE                                    FILE SIZE
 --------------                              --------------
   2.4%   111Ki [None]                        8.45Mi  95.7%
  89.4%  4.01Mi ../sysdeps/x86_64start.S:63   14.8Ki   0.2%
   8.1%   373Ki ../sysdeps/x86_64crti.S:69     373Ki   4.1%
   0.0%       7 ../sysdeps/x86_64start.S:107       7   0.0%
   0.0%       7 ../sysdeps/x86_64start.S:108       7   0.0%
   0.0%       7 ../sysdeps/x86_64start.S:110       7   0.0%
   0.0%       6 ../sysdeps/x86_64start.S:120       6   0.0%
   0.0%       4 ../sysdeps/x86_64crti.S:80         4   0.0%
   0.0%       4 ../sysdeps/x86_64crtn.S:44         4   0.0%
   0.0%       4 ../sysdeps/x86_64start.S:90        4   0.0%
   0.0%       3 ../sysdeps/x86_64start.S:79        3   0.0%
   0.0%       3 ../sysdeps/x86_64start.S:88        3   0.0%
   0.0%       1 ../sysdeps/x86_64crtn.S:45         1   0.0%
   0.0%       1 ../sysdeps/x86_64start.S:122       1   0.0%
   0.0%       1 ../sysdeps/x86_64start.S:85        1   0.0%
   0.0%       1 ../sysdeps/x86_64start.S:93        1   0.0%
   0.0%       1 ../sysdeps/x86_64start.S:97        1   0.0%
 100.0%  4.49Mi TOTAL                         8.83Mi 100.0%

[haberman]

Whoops, clearly I need some better tests for this.

[haberman]

Hmm, I can't reproduce your results. Here's what I get on my system:

$ ./bloaty bloaty -d inlines
     VM SIZE                                                                                        FILE SIZE
 --------------                                                                                  --------------
  27.1%   124Ki [None]                                                                            7.11Mi  95.6%
  55.1%   254Ki [Other]                                                                            254Ki   3.3%
   1.8%  8.34Ki /usr/include/c++/4.8ostream:535                                                   8.34Ki   0.1%
   1.6%  7.57Ki /usr/include/c++/4.8/bitsbasic_ios.h:456                                          7.57Ki   0.1%
   1.4%  6.66Ki /usr/include/c++/4.8/bitsbasic_string.h:539                                       6.66Ki   0.1%
   1.3%  6.19Ki /usr/include/c++/4.8streambuf:466                                                 6.19Ki   0.1%
   1.2%  5.48Ki /usr/bin/../lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/extatomicit  5.48Ki   0.1%
   1.0%  4.64Ki /usr/bin/../lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/extatomicit  4.64Ki   0.1%
   1.0%  4.60Ki /usr/bin/../lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/bitsmove.h:  4.60Ki   0.1%
   0.9%  4.34Ki /usr/bin/../lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/bitsmove.h:  4.34Ki   0.1%
   0.9%  4.28Ki /usr/bin/../lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/bitsbasic_s  4.28Ki   0.1%
   0.8%  3.80Ki /usr/include/c++/4.8/bitsbasic_string.h:249                                       3.80Ki   0.0%
   0.8%  3.58Ki /usr/include/c++/4.8/bitsbasic_ios.h:276                                          3.58Ki   0.0%
   0.7%  3.38Ki /usr/bin/../lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/bitsbasic_s  3.38Ki   0.0%
   0.7%  3.20Ki /usr/bin/../lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/bitsbasic_s  3.20Ki   0.0%
   0.6%  2.92Ki /usr/include/c++/4.8/bitsbasic_string.h:583                                       2.92Ki   0.0%
   0.6%  2.89Ki /usr/include/c++/4.8ostream:385                                                   2.89Ki   0.0%
   0.6%  2.69Ki /usr/include/c++/4.8/bitschar_traits.h:271                                        2.69Ki   0.0%
   0.6%  2.66Ki /usr/include/c++/4.8/bitsbasic_string.h:240                                       2.66Ki   0.0%
   0.6%  2.57Ki /usr/include/c++/4.8ostream:93                                                    2.57Ki   0.0%
   0.5%  2.49Ki /usr/include/c++/4.8sstream:424                                                   2.49Ki   0.0%
 100.0%   461Ki TOTAL                                                                             7.44Mi 100.0%

Can you attach the file that causes this output?

[haberman]

What system are you working on, by the way?

[octoploid]

I'm running Linux. And the output of "-d inlines" is always the same no matter what binary is used.
a.out.gz

[octoploid]

Thanks it work fine now.
One additional nit: There is a missing / in the output, e.g.:

markus@x4 7.0.0 % bloaty -d inlines ./cc1plus
     VM SIZE                                            FILE SIZE
 --------------                                      --------------
  42.8%  12.1Mi [None]                                 181Mi  91.8%
  47.8%  13.6Mi [Other]                               13.6Mi   6.9%
   2.2%   649Ki ../../gcc/gcctree.h:3159               649Ki   0.3%
   1.5%   442Ki ../../gcc/gcctree.h:3158               442Ki   0.2%
   1.2%   353Ki ../../gcc/gcctree.h:3045               353Ki   0.2%
   0.7%   208Ki ../../gcc/gcctree.h:3168               208Ki   0.1%
   0.5%   136Ki ../../gcc/gcctree.h:3167               136Ki   0.1%
   0.5%   130Ki ../../gcc/gcctree.h:3044               130Ki   0.1%
   0.4%   103Ki ../../gcc/gccvec.h:732                 103Ki   0.1%
   0.3%  99.1Ki ../../gcc/gcctree.h:3545              99.1Ki   0.0%
   0.3%  92.2Ki .genrtl.h:29                          92.2Ki   0.0%
   0.3%  80.1Ki ../../gcc/gcctree.h:3245              80.1Ki   0.0%
   0.2%  60.7Ki ../../gcc/gccgimple-match-head.c:780  60.7Ki   0.0%
   0.2%  50.5Ki ../../gcc/gcctree.h:3214              50.5Ki   0.0%
   0.2%  47.6Ki ../../gcc/gcctree.h:3246              47.6Ki   0.0%
   0.2%  47.0Ki ../../gcc/gcctree.h:3248              47.0Ki   0.0%
   0.2%  46.0Ki ../../gcc/gccvec.h:1069               46.0Ki   0.0%
   0.1%  43.2Ki ../../gcc/gcctree.h:3531              43.2Ki   0.0%
   0.1%  39.7Ki ../../gcc/gcctree.h:3244              39.7Ki   0.0%
   0.1%  38.4Ki .genrtl.h:33                          38.4Ki   0.0%
   0.1%  38.3Ki ../../gcc/gccvec.h:288                38.3Ki   0.0%
 100.0%  28.3Mi TOTAL                                  197Mi 100.0%

Should be ../../gcc/gcc/tree.h:3159 not ../../gcc/gcctree.h:3159 .

[haberman]

Ah yes. I've just pushed a fix for that.