Little memdump

[Jeff McGlynn]

On 2007-03-15 02:03:44 -0700, Bill Baxter <dnewsgroup at billbaxter.com> said:

> Jeff McGlynn wrote:
>> On 2007-03-12 17:40:16 -0700, Tom <tom at nospam.com> said:
>> 
>>> Hi people,
>>> 
>>>    I just wanted to share with you a *very-humble* memdump procedure 
>>> that has been very helpful to me in my tedious D-debugging nights.
>>> 
>>> Kind regards,
>> 
>> Here is the code that I use:
>> 
>> import std.stdio;
>> import std.ctype;
>> 
>> void debug_dump(ubyte[] data) {
>>     uint len = data.length;
>>     writefln("=== GOT %d BYTES:", len);
>>         for(uint offset = 0; offset < len; offset += 20) {
>>         uint end = offset + 20;
>>         if (end > len) end = len;
>>                for(uint ix = offset; ix < end; ++ix) {
>>             writef("%02X ", data[ix]);
>>         }
>>                if (offset + 20 > len) {
>>             char[] space = new char[3 * (offset + 20 - len)];
>>             space[] = ' ';
>>                        writef(space);
>>         }
>>                writef("| ");
>>                for(uint ix = offset; ix < end; ++ix) {
>>             if (isprint(data[ix])) {
>>                 writef("%s", cast(char) data[ix]);
>>             } else {
>>                 writef(".");
>>             }
>>         }
>>                writefln();
>>     }
>> 
>>     writefln();
>> }
>> 
> 
> 
> Just curious -- what are you guys doing that you have to dump memory so often?
> 
> --bb

I recently wrote a FastCGI server, which uses a binary protocol.  Using 
this allowed me to debug protocol errors without adding new code each 
time.

-- 

Jeff McGlynn