[
https://issues.apache.org/jira/browse/STDCXX-213?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12481765
]
Martin Sebor commented on STDCXX-213:
-------------------------------------
The policy that was in place up until the change above was for every stdcxx
library header to #include <rw/_defs.h> as the last #include directive.
The new policy esstablished with the change is that a stdcxx library header
#include <rw/_defs.h> if and only if it doesn't #include any other stdcxx
header.
The exception in both cases is the situation when the #inclusion of the first
header depends on a configuration macro. In that case <rw/_defs.h> must be
#included first (unless the dependency can be resolved by moving the #include
directive below an unconditional #include directive for another stdcxx header).
> add header guard for #include <rw/_defs.h>
> ------------------------------------------
>
> Key: STDCXX-213
> URL: https://issues.apache.org/jira/browse/STDCXX-213
> Project: C++ Standard Library
> Issue Type: Improvement
> Components: Build
> Affects Versions: 4.1.3, 4.1.2
> Environment: all
> Reporter: Martin Sebor
> Assigned To: Martin Sebor
> Priority: Minor
> Fix For: 4.2
>
>
> Moved from the Rogue Wave bug tracking database:
> ****Created By: dean @ Jun 02, 2004 09:58:11 AM****
> Hi Kathy,
> Here is the request that I have for additional work for RogueWave to do for
> us. This is a minor request but has some big payback, especially when the
> RogueWave code is checked into ClearCase. In my simple test, I reduced the
> number of include files that were opened from 200 to 100! I haven't run the
> big test yet (the one that opens more than 2,400 header files.
> The request is to add a different kind of header guard (in addition to the
> one that is currently being used). For example, instead of the code:
> #include <rw/_defs.h>
> Use the code:
> #ifndef _RWSTD_DEFS_H_INCLUDED
> #include <rw/_defs.h>
> #endif
> This should be done in all header files and .cc files for all #includes
> that have header guards.
> Let me know if this isn't clear or you have questions.
> Thanks,
> Bert
> **** Entered By: dean @ Wednesday, June 02, 2004 9:58:03 AM ****
> ****Modified By: sebor @ Apr 11, 2005 03:11:29 PM****
> This is doable but it has been assumed that modern preprocessors implement
> this optimization. It would be useful to know what compiler the customer is
> using.
> Deferred for now.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
