Re: [PATCH v3] gitweb: standarize HTTP status codes

[Jakub Narebski <jnareb@gmail.com>]

Lea Wiemann <lewiemann@gmail.com> writes:

> Changes since v2: die_error now adds the reason strings defined by RFC
> 2616 to the HTTP status code; incorporated Jakub's other suggestions.
> Diff to v2 follows.

This address both of my concerns: first, that for someone examining
Mechanize-based gitweb test number like 403, or 500 would be magical
number without explanation (reason phrase) other than 'Error'.

Second, that for casual / accidental gitweb developer who has to add
or modify a bit of code with die_error(...) wouldn't know which of
"magic number" to use, if the case didn't fail into described
situation.  Now it is enough to example die_error(...) in addition to
similar code...
 
> I didn't use the HTTP_NOT_FOUND etc. suggestion because I found it too
> verbose and obtrusive.

I can agree with that.

> Just a friendly reminder, please remember that discussing fairly
> trivial changes in-depth might be not a good use of all participants'
> time [...]

Well, this was what I though was patch revies... :-/

> Anyways, I hope everyone is happy with this version of the patch.

FWIW:

Acked-by: Jakub Narebski <jnareb@gmail.com>

-- 
Jakub Narebski
Poland
ShadeHawk on #git