[ 
https://issues.apache.org/jira/browse/TS-130?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12831063#action_12831063
 ] 

John Plevyak commented on TS-130:
---------------------------------

I am still getting the hang of git and I seem to have mixed up some other 
changes into this
this patch which needs to be cleaned up to include only the config.h changes.

I'll rebuild the patch.

... DDebug is short for DebugDebug since we decided to keep Debug diags in
the release version and there are things in the cache which really for use on 
DEBUG builds they
would be DebugDebug (or perhaps ReallyDebug ? :) . Several subsystems turn off 
the some
Diags for release builds because the engineers who wrote never intended them 
for release builds
and so there are several incompatible ways they did it.   The saddest 
commentary on this battle
is in logging where the author #define NO_DIAGS and then someone else explicity
#undef NO_DIAGS in the Diags.h header in response.  

I would like to propose that we use Diag("xxx") for diags that are available 
all the time and
Debug("xx") for diags that are only available on DEBUG builds and standardize 
the way to
specify which builds a particular diagnostic message will be available in.


> we should generate a config.h file instead of putting all the -Ds in the 
> CFLAGS
> -------------------------------------------------------------------------------
>
>                 Key: TS-130
>                 URL: https://issues.apache.org/jira/browse/TS-130
>             Project: Traffic Server
>          Issue Type: Improvement
>            Reporter: John Plevyak
>            Assignee: John Plevyak
>         Attachments: config.h-jp-v1.patch, ts-config.h-jp-v2.patch
>
>
> We should use a config.h instead of a huge pile of -Ds

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to