Hi,

following a remark of Eric Covener about 'indent', I started to play with it. I have committed a few patches in trunk on modules that don't have any differences between 2.4 and trunk.
    debugging/mod_firehose    (only in trunk)
    debugging/mod_dumpio
    debugging/mod_bucketeer
    echo/mod_echo
    aaa/mod_allowmethods

My first goal was to see if many things "had" to be fixed.

Well, depending on the files, these can lead to update a relatively large amount of lines.


1) This kind of patch should have no impact on generated code.
I md5sum the .o files before and after modification. Some difference exist.
I guess that it is due to debugging information embedded in the .o file.

What is the best way to check that these changes have no side effects?
Should disabling the debug in my makefile be enough?


2) As it involve sometimes a lot of changes, I was wondering if it worthed it. Even if binary comparison can be a good starting point, reviews required by the voting system could be very time consuming for others. Moreover, in case of #if, not all paths would be checked with binary comparison. This could be bug prone.

On the other side, if housekeeping is never done, will never have the benefit of it...


3) These kind of changes can easily conflict with other patches. Backporting things would then become a real nightmare for everyone. So, I think that no such style issue should be fixed on files that differ between trunk and 2.4 (or only 1 line changes or after the last difference)
However, the question remains in case of 2.2 backport.


So, do you think that such "clean up" worth the effort or that things should be left as-is ?
Thanks for feed back.

CJ

Reply via email to