Re: [PATCH v5 00/11] add performance tracing facility

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Jun 25, 2014 at 9:49 PM, Karsten Blees <karsten.blees@xxxxxxxxx> wrote:
> Am 25.06.2014 16:28, schrieb Duy Nguyen:
>> On Wed, Jun 11, 2014 at 2:55 PM, Karsten Blees <karsten.blees@xxxxxxxxx> wrote:
>>> Here's v5 of the performance tracing patch series, now including a bunch of cleanups and adding timestamp, file and line to all trace output.
>>>
>>> I'm particularly interested in feedback for the output format. As file names have different lengths, printing file:line as prefix results in unaligned output:
>>>
>>>  > GIT_TRACE=1 git stash list
>>>  00:12:10.544266 git.c:512 trace: exec: 'git-stash' 'list'
>>>  00:12:10.544266 run-command.c:337 trace: run_command: 'git-stash' 'list'
>>>  00:12:10.649779 git.c:312 trace: built-in: git 'rev-parse' '--git-dir'
>>
>> Can I have an (build-time) option to show <function>:<line> instead of
>> <file>:<line>? I know it's not supported by all compilers, which may
>> make support a bit cumbersome..
>>
>
> Is this really useful? <file>:<line> is unique, but <function>:<line> is not. E.g. in case of "hash_name:47" you'd have to guess if its the one in attr.c or name-hash.c...

It depends on your view. If I'm tracing a certain operation, function
names let me know roughly what's going on without looking at the code
because all (or many of) the names and their purposes are already in
my mind. But for publishing the traces, then I agree <file>:<line> is
better.
-- 
Duy
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]