Old code never dies.  Unless you personally go and destroy every repository, 
backup, and website that ever mentions it, it will come back to haunt you.  As 
proof-  I offer up my debugging library.

Back in college, I was a cocky, know-it-all programmer who had never worked on 
anything more than a few thousand lines of code  (as differentiated from now, 
where only the first two things are true).  At that time, after having various 
issues with memory allocation, I wrote a quick debugging library in C++, mainly 
around finding and documenting leaks.  It worked, with its limitations.  At the 
time I thought it was clever.  Now-  dear god is it a hackfest.  I didn't 
really know what I was doing when I did it.


Fast forward 8 years.  A month ago, I get an email from Google saying that 
someone wanted to start a project on gforge with the same name as a project of 
mine on sourceforge-  the debug library.  Now I made a very common name for it 
(this was early on in sourceforge history, good names were available), so I 
figured it was just the name.  Not having touched the code in 8 years, I of 
course said sure and gave up the name-  I figured there was no way anyone would 
ever be using my old code.

That was until this morning.  I got an email (isn't it great having the same 
email for a decade) with an indian sender's name and the subject "hi".  I 
assume its spam, but open it anyway just in case.  And much to my surprise-  
its someone talking about programming.  Specifically, my library.  With 
technical questions.  Dear god, he's actually using it?!?!?!

This of course makes me curious.  So I quickly google the library's name, and 
look at the project page.  Its actually at the 74% activity rating.  I guess 
the fact it actually hit version 1.01 (fixing the bug in version 1.0 where I 
skipped every other element in a list) must have made people hopeful.  Despite 
not having any files in CVS (did sf do CVS back then?) there was a source 
tarball.  In 8 years, over 1000 people have downloaded this piece of shit.  
Although 1/3 of them have read the readme, realised it sucked, and never 
downloaded the code.

So here it is-  proof that old code never dies.  Any suggestions on how to deal 
with the support request?




Just for fun, so you can all mock the 20 year old me, here's a list of the 
"features" of my library.

*Special license, because GPL wasn't good enough-  GPL with a disclaimer 
disallowing for profit redistribution.  I assume nobody is listening to that.  
In fact, I assume they probably aren't following the GPL either.  I hearby 
waive the non-profit requirement.  I'll keep it GPL, but likely not run after 
anyone breaking it.
*"Funny" copyright line.  Complete with typo.
*My own linked list class.  Minor WTF, due to the state of the STL in 2008
*Instead of overriding new and delete, I had wrappers.  Minor wtf, as I 
remember having issues overrding new and delete on some compilers back then
*Could log debug messages to a C FILE* or C++ fstream.
*Defined all my internal structures in the public .h.  In fact, i had no 
private .h
*Adds an optional string name to allocations/deallocations, to allow them to be 
tracked by name.  Not smart enough to add __FILE__ and __LINE__ to the mix
*wrappers for malloc and free
*Very, very, very basic debug logging and function entrance/exit logging.  So 
basic as to be useless.

For the life of me I can't understand why anyone would use this code-  a day of 
work would give you something 5x better.

Gabe






_________________________________________________________________
Get Free (PRODUCT) REDâ„¢  Emoticons, Winks and Display Pics.
http://joinred.spaces.live.com?ocid=TXT_HMTG_prodredemoticons_052008
--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Reply via email to