httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roy T. Fielding" <>
Subject vetoing hide.h
Date Tue, 31 Mar 1998 04:27:06 GMT
I almost did this last month, held back because it was defaulted off,
and then almost did it again when the default was switched, but then
held off until I could get a better handle on why hide was needed.
I now have that, so here goes...

I am vetoing hide.h and any form of symbol rewriting within the server
that isn't absolutely necessary for proper operation (as is the signal
symbol rewrite).  The reasons:

   1) it is poor software engineering (makes the code harder to read
      for the sake of saving the one-time task of replacing the current
      code with new symbol names).

   2) it is a maintenance nightmare because it becomes too easy to
      arbitrarily choose which symbol to use.

   3) it is a 3rd-party nightmare because it breaks the binary API
      in a way that is nearly impossible to see unless you happen to
      be on this mailing list.

Yes, some of these are very subjective reasons.  OTOH, at least I can
claim some expertise in the area.

I am at the LA IETF all week.  That means on Saturday I will reverse
the patch, unless someone comes up with a better solution before then,
which in turn means we better decide on the final prefix this week.
I am inclined toward a single "ap_" prefix for all routines, and separate
libraries for the various distinctions people wanted to make among
"ap_", "apapi_", and "appri_".  In my opinion, it is wrong to make
those distinctions within the symbol name, just as it is wrong to
declare private functions in a public header file.


View raw message