[Bug debug/67163] g2 generates incorrect decl_line

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67163

--- Comment #5 from Andrew Pinski  ---
*** Bug 67162 has been marked as a duplicate of this bug. ***

[Bug debug/67163] g2 generates incorrect decl_line

[dehao at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67163

--- Comment #1 from dehao at gcc dot gnu.org ---
Created attachment 36157
  -- https://gcc.gnu.org/bugzilla/attachment.cgi?id=36157action=edit
preprocessed file

[Bug debug/67163] g2 generates incorrect decl_line

[dehao at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67163

--- Comment #2 from dehao at gcc dot gnu.org ---
Created attachment 36158
  -- https://gcc.gnu.org/bugzilla/attachment.cgi?id=36158action=edit
preprocessed file

[Bug debug/67163] g2 generates incorrect decl_line

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67163

--- Comment #3 from Andrew Pinski pinskia at gcc dot gnu.org ---
 g2 generates incorrect decl_line (150 is the decl_line for another function: 
 addValue)

You did paste the full debug info as the file numbers are way different between
the two.

[Bug debug/67163] g2 generates incorrect decl_line

[dehao at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67163

--- Comment #4 from dehao at gcc dot gnu.org ---
Right, in -g2 binary debug info, the decl_file is pointing to the header file
in which the function is declared. However, if the definition is available, it
makes more sense to point decl_file and decl_line where it's actually defined,
right? Additionally, it's better to keep g1 and g2 debug info consistent.