[PATCH] gitweb: Return or exit after done serving request

[PATCH] gitweb: Return or exit after done serving request

[Jakub Narebski]

Check if there is a caller in top frame of gitweb, and either 'return'
if gitweb code is wrapped in subroutine, or 'exit' if it is not.

This should avoid

  gitweb.cgi: Subroutine git_SOMETHING redefined at gitweb.cgi line NNN

warnings in error_log when running gitweb with mod_perl (using
ModPerl::Registry handler)

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
---
I have noticed bunch of

  [Sun Jun 13 11:58:02 2010] gitweb.cgi: Subroutine git_opml redefined
  at /var/www/perl/gitweb/gitweb.cgi line 6808.

warnings in /var/log/httpd/error_log, after running gitweb from mod_perl.
This fixes it.

 gitweb/gitweb.perl |    9 +++++++++
 1 files changed, 9 insertions(+), 0 deletions(-)

diff --git a/gitweb/gitweb.perl b/gitweb/gitweb.perl
index e108bbc..02f366d 100755
--- a/gitweb/gitweb.perl
+++ b/gitweb/gitweb.perl
@@ -987,7 +987,16 @@ if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
 	die_error(400, "Project needed");
 }
 $actions{$action}->();
+
 DONE_GITWEB:
+if (defined caller) {
+	# wrapped in a subroutine processing requests,
+	# e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
+	return;
+} else {
+	# pure CGI script, serving single request
+	exit;
+}
 1;
 
 ## ======================================================================

Re: [PATCH] gitweb: Return or exit after done serving request

[Junio C Hamano]

Jakub Narebski <jnareb@gmail.com> writes:

> diff --git a/gitweb/gitweb.perl b/gitweb/gitweb.perl
> index e108bbc..02f366d 100755
> --- a/gitweb/gitweb.perl
> +++ b/gitweb/gitweb.perl
> @@ -987,7 +987,16 @@ if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
>  	die_error(400, "Project needed");
>  }
>  $actions{$action}->();
> +
>  DONE_GITWEB:
> +if (defined caller) {
> +	# wrapped in a subroutine processing requests,
> +	# e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
> +	return;
> +} else {
> +	# pure CGI script, serving single request
> +	exit;
> +}
>  1;

Is the last "1;" still needed if we did this?

I am guessing that this new codeblock will go inside "sub run" when
merging with your c2394fe (gitweb: Put all per-connection code in run()
subroutine, 2010-05-07) and Sam's a0446e7 (gitweb: Add support for
FastCGI, using CGI::Fast, 2010-05-07).  If I am mistaken, please advise.

Thanks.

Re: [PATCH] gitweb: Return or exit after done serving request

[Jakub Narebski]

On Mon, 14 June 2010, Junio C Hamano wrote:
> Jakub Narebski <jnareb@gmail.com> writes:
> 
> > diff --git a/gitweb/gitweb.perl b/gitweb/gitweb.perl
> > index e108bbc..02f366d 100755
> > --- a/gitweb/gitweb.perl
> > +++ b/gitweb/gitweb.perl
> > @@ -987,7 +987,16 @@ if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
> >  	die_error(400, "Project needed");
> >  }
> >  $actions{$action}->();
> > +
> >  DONE_GITWEB:
> > +if (defined caller) {
> > +	# wrapped in a subroutine processing requests,
> > +	# e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
> > +	return;
> > +} else {
> > +	# pure CGI script, serving single request
> > +	exit;
> > +}
> >  1;
> 
> Is the last "1;" still needed if we did this?

Probably not, and probably we never need this.

> 
> I am guessing that this new codeblock will go inside "sub run" when
> merging with your c2394fe (gitweb: Put all per-connection code in run()
> subroutine, 2010-05-07) and Sam's a0446e7 (gitweb: Add support for
> FastCGI, using CGI::Fast, 2010-05-07).  If I am mistaken, please advise.

No, the code is meant to be after last code to be run in gitweb, 
in *top level* scope, otherwise it would always return.

So the patch merged with commits mentioned above ('jn/gitweb-fastcgi',
old or new) would look like the following... but I am not sure if it
is needed, as I tried to reproduce error / warning that I got to ensure
that the patch really fixes it (as in new post "sub run" code there
are subroutine definitions before end of 'main' code, i.e. run(); in
new post "sub run" code, $actions{$action}->(); in pre "sub run" code)
and I couldn't get again this:

  /var/log/httpd/error_log:
  ...
  [Sun Jun 13 11:58:02 2010] gitweb.cgi: Subroutine git_atom redefined at /var/www/perl/gitweb/gitweb.cgi line 6804.
  [Sun Jun 13 11:58:02 2010] gitweb.cgi: Subroutine git_opml redefined at /var/www/perl/gitweb/gitweb.cgi line 6808.
  ...gi-bin

Lets table it now.

-- >8 --
Subject: [RFC/PATCH] gitweb: Return or exit after done serving request

Check if there is a caller in top frame of gitweb, and either 'return'
if gitweb code is wrapped in subroutine, or 'exit' if it is not.

This should avoid

  gitweb.cgi: Subroutine git_SOMETHING redefined at gitweb.cgi line NNN

warnings in error_log when running gitweb with mod_perl (using
ModPerl::Registry handler)

Signed-off-by: Jakub Narebski <jnareb@gmail.com>
---
 gitweb/gitweb.perl |   10 ++++++++++
 1 files changed, 10 insertions(+), 0 deletions(-)

diff --git a/gitweb/gitweb.perl b/gitweb/gitweb.perl
index 14ef50c..43589b5 100755
--- a/gitweb/gitweb.perl
+++ b/gitweb/gitweb.perl
@@ -1111,6 +1111,16 @@ sub run {
 
 run();
 
+if (defined caller) {
+	# wrapped in a subroutine processing requests,
+	# e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
+	return;
+} else {
+	# pure CGI script, serving single request
+	# or 'exit' hijacked appropriately
+	exit;
+}
+
 ## ======================================================================
 ## action links
 
-- 
1.7.0.1

Re: [PATCH] gitweb: Return or exit after done serving request

[Junio C Hamano]

Jakub Narebski <jnareb@gmail.com> writes:

> On Mon, 14 June 2010, Junio C Hamano wrote:
> ...
>> Is the last "1;" still needed if we did this?
>
> Probably not, and probably we never need this.
>> 
>> I am guessing that this new codeblock will go inside "sub run" when
>> merging with your c2394fe (gitweb: Put all per-connection code in run()
>> subroutine, 2010-05-07) and Sam's a0446e7 (gitweb: Add support for
>> FastCGI, using CGI::Fast, 2010-05-07).  If I am mistaken, please advise.
>
> No, the code is meant to be after last code to be run in gitweb, 
> in *top level* scope, otherwise it would always return.

Ok, tentatively I fixed the mismerge of 'pu' from the last night and
pushed it out.