[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference turris::decladebug

Title:Digital Ladebug debugger
Moderator:TLE::LUCIA
Created:Fri Feb 28 1992
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:969
Total number of notes:3959

904.0. "Question about new C++ template support" by CXXC::MJHANS (Matthew Hanselman, DEC C/C++) Mon Mar 31 1997 11:34

    Re: new support for debugging templates in C++:
    
    Will ladebug eventually hide the mangled source names (e.g.
    "x_0016.hxx~alt~deccxx_3DB94EC1"`f`x as opposed to "x_0016.hxx"`f`x),
    or leave it the way it is now?
    
    I just need to know how to update some of our testing benchmarks.
    
    							- Matt
T.RTitleUserPersonal
Name
DateLines
904.1GEMGRP::MONTELEONEMon Mar 31 1997 13:4010

   I hope that the extra information encoded in the file name will be hidden.

   For one thing, the checksum like sequence at the end of the name will
   be varying consistently, so you will have a tough time of maintaining
   your benchmarks without adding a filter...


   Bob
904.2TLE::BRETTMon Mar 31 1997 22:187
There is no 'spec' for this stuff, but the Ada support has been there for
years, and this stuff is not hidden.

Since the user has to be able to cut/paste this stuff, it is unlikely that
it will be hidden in the future.

/Bevin
904.3Cannot find source fileCXXC::MJHANSMatthew Hanselman, DEC C/C++Wed Apr 02 1997 11:2210
    Okay
    
    But we are seeing error messages like
    
    	Cannot find source file x_0016.hxx~alt~deccxx_XXXXXXXX
    
    I'm assuming that's because we haven't picked up a version of ladebug
    that deals with mangled names, rather than being a bug in ladebug, no?
    
    								- Matt
904.4TLE::BRETTWed Apr 02 1997 13:574
I hope so - Steve Merritt will get you a 4.0 production BL34 today to make sure
that is the problem

/Bevin
904.5It worksCXXC::MJHANSMatthew Hanselman, DEC C/C++Wed Apr 02 1997 15:455
    A 4.0-34 ladebug does fix this problem.
    
    We'll upgrade cosf on Monday, when Charlie & I will both be here.
    
    								- Matt
904.6CXXC::MJHANSMatthew Hanselman, DEC C/C++Tue Apr 08 1997 18:5126
    I was just fooling around with this a little, and the way the user
    has to deal with this is a little unfriendly.
    
    Take the code example in 917, or any code in a non-mergable file.  The
    user can no longer use the command
    
    	stop at "filename":line-number
    
    but instead has to use
    
    	stop at "filename~alt~deccxx_NNNNNNNN":line-number
    
    If you try the 1st, it won't let you.  Nothing is said about the
    non-mergable file, and the user has to find some way of finding out the
    CRC (which changes with each compilation)
    
    As a user, I could argue that using the 1st form should actually go
    through and set the breakpoint at that line in all the non-mergable
    files matching that pattern.  At the very least, I would like a more
    helpful message if I attempt to use the 1st form (even if it's a
    referral to a ladebug help topic explaining the CRC bit somewhat)
                                                              
    To summarize: it took me a little while to figure it out, and I knew
    what was going on.  Somebody else might not be as lucky.
    
    							- Matt
904.7TLE::BRETTTue Apr 08 1997 22:456
agreed -   at least you get bpt buttons in the gui now that work.

the rest of the work is also needed for other languages - especially for
static functions in c.

/Bevin