Project

General

Profile

Actions

Feature #2067

closed

base.h and http_req.h type names

Added by hazzadous over 14 years ago. Updated almost 14 years ago.

Status:
Wontfix
Priority:
Normal
Category:
-
Target version:
ASK QUESTIONS IN Forums:

Description

There seem to be several variable names in these files that share their names with some of their instances. Is there any chance we can append _t to specify types?

  • type http_req -> type http_req_t
  • type request -> type request_t
  • type response -> type response_t
  • type physical -> type physical_t
  • type stat_cache -> type stat_cache_t

This doesn't effect the C world, but its a little annoying if your trying to compile C++ code against these headers.


Files

Issue2067Trunk.diff (9.82 KB) Issue2067Trunk.diff Issue 2067 Patch for Trunk (1.5.0) jborgohain, 2009-12-17 02:55
issue2067Trunk.dff (11.4 KB) issue2067Trunk.dff jborgohain, 2010-01-23 21:13

Related issues 1 (0 open1 closed)

Related to Feature #1944: Bindings for C++ modulesWontfix2009-03-22Actions
Actions #1

Updated by jborgohain over 14 years ago

Here is a patch that includes the changes requested in this issue. This patch is for the trunk (1.5.0).

Actions #2

Updated by stbuehler over 14 years ago

  • % Done changed from 100 to 0
  • Missing in 1.5.x set to No

*_t is reserved for the implementation.

Actions #3

Updated by jborgohain over 14 years ago

Should I change it from "*_t" to something like "*_type"? Or how should this be tackled?

Actions #4

Updated by Olaf-van-der-Spek over 14 years ago

stbuehler wrote:

*_t is reserved for the implementation.

What do you mean?

Actions #5

Updated by icy over 14 years ago

_t is reserved for POSIX implementations. http://www.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html
Many people ignore it and it would probably not be enforceable but well...

Actions #6

Updated by jborgohain about 14 years ago

Resubmitting patch with new changes. Replaced suffix _t with _type.

Actions #7

Updated by Olaf-van-der-Spek about 14 years ago

icy wrote:

_t is reserved for POSIX implementations. http://www.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html

That doesn't sound right, although it's indeed what that doc says.

Many people ignore it and it would probably not be enforceable but well...

I think you should ignore it too.

Actions #8

Updated by jborgohain about 14 years ago

Olaf-van-der-Spek wrote:

I think you should ignore it too.

Well, I have submitted two patches, one puts "_t" and the new one puts "_type". I can't commit, so... :)

Actions #9

Updated by stbuehler about 14 years ago

  • We will obviously not change the type names in 1.4.x (stable version, ...)
  • As we are working on a new version (2.0), i don't want to put too much effort in 1.5. This means that i don't want to get more differences between 1.4 and 1.5, so that at least some patches are easy to port between them.

So i don't think this patch will go upstream. But i think the type/member names in 2.0 will be ok for c++ (and we prefixed all names, so hopefully no namespace conflicts with stupid lua libraries...).

Actions #10

Updated by stbuehler almost 14 years ago

  • Status changed from New to Wontfix
Actions #11

Updated by gstrauss about 2 months ago

Actions

Also available in: Atom