httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ralf S. Engelschall" <>
Subject Re: cvs commit: apache-1.3/src/helpers GuessOS PrintPath TestCompile find-dbm-lib fp2rp
Date Sat, 16 Jan 1999 09:36:10 GMT

In article <> you wrote:

> Yes, I realise there were some good changes in jim's patch -- adding ""s
> around things.  Unfortunately I'm not willing to accept the rest of the
> changes.

> Given the variance of shells, I'd half expect there is some shell out
> there that will see:

>    if [ "$opt" != "" ]; 

> and if $opt has a leading - it'll screw up.

Yes, as the fix I'd to apply to yesterday cleary showed.

> Maybe the common practice of using:

>    if [ "x$opt" != "x" ];

> isn't really required, but *it is common practice*, and I don't see any
> reason for us to go around rooting out broken shells just so that we can
> solve some fucking waste of time STYLE DEBATE.  Jesus. 

Yes, and although APACI used a "." instead of a "x" it was used _everywhere_
and this way it was _already_ safe. Instead the recent style changes
introduced new problems because of side-effects.

> Clearly this config stuff is hard enough that you guys have to futz with
> it every revision and make module authors change their tools, and make
> folks doing ports change their scripts and make people have to relearn
> things every new minor revision.  And given that it's that hard I don't
> want one more stupid reason for incompatibilities added to the bunch. 

> flame away. 

Yes, although I see Jim's intentions for coding style cleanups (which are ok
in general, and I've no problem with replacing my "." style with "x" style, of
course) I have to vote +1 for Dean's backout of the changes. There are too
many special cases and at least a review-than-commit cycle sounds reasonable
to me before we do ping-pong commits on the code just because of coding style
                                       Ralf S. Engelschall

View raw message