Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Helmut K. C. Tessarek

On 2024-06-25 20:43, Yann Ylavic wrote:

Should be fixed by
https://github.com/apache/httpd/commit/6937b985ae112de0ad60f12a3cb522b608a4d501.diff


I'll test this tomorrow. Thanks!

--
regards Helmut K. C. Tessarek  KeyID 0x172380A011EF4944
Key fingerprint = 8A55 70C1 BD85 D34E ADBC 386C 1723 80A0 11EF 4944

/*
   Thou shalt not follow the NULL pointer for chaos and madness
   await thee at its end.
*/


OpenPGP_signature.asc
Description: OpenPGP digital signature


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Helmut K. C. Tessarek

On 2024-06-25 19:55, Yann Ylavic wrote:

Done in r1918626. Looks good in my manual testing too, there will be
two calls to ap_proxy_fixup_uds_filename() when proxying from
r->handler, first in proxy_handler() and second in
ap_proxy_pre_request(), but since the first will strip the UDS already
the other call is a noop. We can eventually rework this after the
release.


May I ask what this fixup was supposed to fix? I take it this was 
introduced after 2.4.59, so I guess it must have been added to fix 
something, but broke something else.
If it's too complicated to explain, that's ok too. I am just curious. 
That's all.


--
regards Helmut K. C. Tessarek  KeyID 0x172380A011EF4944
Key fingerprint = 8A55 70C1 BD85 D34E ADBC 386C 1723 80A0 11EF 4944

/*
   Thou shalt not follow the NULL pointer for chaos and madness
   await thee at its end.
*/


OpenPGP_signature.asc
Description: OpenPGP digital signature


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Yann Ylavic
On Wed, Jun 26, 2024 at 12:02 AM Helmut K. C. Tessarek
 wrote:
>
> I tried with rc2, since rc3 is not available yet.
>
> Same issue.

Should be fixed by
https://github.com/apache/httpd/commit/6937b985ae112de0ad60f12a3cb522b608a4d501.diff

Regards;
Yann.


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Yann Ylavic
On Wed, Jun 26, 2024 at 1:17 AM Eric Covener  wrote:
>
> On Tue, Jun 25, 2024 at 7:03 PM Yann Ylavic  wrote:
> >
> > On Wed, Jun 26, 2024 at 12:37 AM Eric Covener  wrote:
> > >
> > > > The attached might work, currently testing but sending early if you
> > > > want to try too.
> > >
> > > looks like proxy: is stripped off after the new call and needs to be
> > > added back in?
> >
> > Yeah, the new call to ap_proxy_fixup_uds_filename() should not take
> > &r->filename as argument.
> > Fixed in this new version.
>
> +1 with basic "fakefpm" and real fpm hello-world.  Please commit to
> trunk when practical.

Done in r1918626. Looks good in my manual testing too, there will be
two calls to ap_proxy_fixup_uds_filename() when proxying from
r->handler, first in proxy_handler() and second in
ap_proxy_pre_request(), but since the first will strip the UDS already
the other call is a noop. We can eventually rework this after the
release.

>
> If we can get another review for 2.4.x I can add it and roll an rc so
> it can be more easily picked up for testing.

+1


Regards;
Yann.


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Eric Covener
On Tue, Jun 25, 2024 at 7:03 PM Yann Ylavic  wrote:
>
> On Wed, Jun 26, 2024 at 12:37 AM Eric Covener  wrote:
> >
> > > The attached might work, currently testing but sending early if you
> > > want to try too.
> >
> > looks like proxy: is stripped off after the new call and needs to be
> > added back in?
>
> Yeah, the new call to ap_proxy_fixup_uds_filename() should not take
> &r->filename as argument.
> Fixed in this new version.

+1 with basic "fakefpm" and real fpm hello-world.  Please commit to
trunk when practical.

If we can get another review for 2.4.x I can add it and roll an rc so
it can be more easily picked up for testing.

-- 
Eric Covener
cove...@gmail.com


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Yann Ylavic
On Wed, Jun 26, 2024 at 12:37 AM Eric Covener  wrote:
>
> > The attached might work, currently testing but sending early if you
> > want to try too.
>
> looks like proxy: is stripped off after the new call and needs to be
> added back in?

Yeah, the new call to ap_proxy_fixup_uds_filename() should not take
&r->filename as argument.
Fixed in this new version.
Index: include/ap_mmn.h
===
--- include/ap_mmn.h	(revision 1918625)
+++ include/ap_mmn.h	(working copy)
@@ -601,6 +601,7 @@
  * 20120211.131 (2.4.59-dev) Add DAV_WALKTYPE_TOLERANT
  * 20120211.131 (2.4.60-dev) Add ap_set_content_type_ex(), ap_filepath_merge(),
  *   and AP_REQUEST_TRUSTED_CT BNOTE.
+ * 20120211.133 (2.4.60-dev) Add ap_proxy_fixup_uds_filename()
  */
 
 #define MODULE_MAGIC_COOKIE 0x41503234UL /* "AP24" */
@@ -608,7 +609,7 @@
 #ifndef MODULE_MAGIC_NUMBER_MAJOR
 #define MODULE_MAGIC_NUMBER_MAJOR 20120211
 #endif
-#define MODULE_MAGIC_NUMBER_MINOR 131 /* 0...n */
+#define MODULE_MAGIC_NUMBER_MINOR 133 /* 0...n */
 
 /**
  * Determine if the server's current MODULE_MAGIC_NUMBER is at least a
Index: modules/proxy/mod_proxy.c
===
--- modules/proxy/mod_proxy.c	(revision 1918625)
+++ modules/proxy/mod_proxy.c	(working copy)
@@ -1298,7 +1298,7 @@ static int proxy_handler(request_rec *r)
 ap_get_module_config(sconf, &proxy_module);
 apr_array_header_t *proxies = conf->proxies;
 struct proxy_remote *ents = (struct proxy_remote *) proxies->elts;
-int i, rc, access_status;
+int rc = DECLINED, access_status, i;
 int direct_connect = 0;
 const char *str;
 apr_int64_t maxfwd;
@@ -1314,22 +1314,33 @@ static int proxy_handler(request_rec *r)
 }
 
 if (!r->proxyreq) {
-rc = DECLINED;
 /* We may have forced the proxy handler via config or .htaccess */
 if (r->handler &&
 strncmp(r->handler, "proxy:", 6) == 0 &&
 strncmp(r->filename, "proxy:", 6) != 0) {
+char *old_filename = r->filename;
+
 r->proxyreq = PROXYREQ_REVERSE;
 r->filename = apr_pstrcat(r->pool, r->handler, r->filename, NULL);
+
 /* Still need to fixup/canonicalize r->filename */
-rc = proxy_fixup(r);
+uri = r->filename + 6;
+rc = ap_proxy_fixup_uds_filename(r, &uri);
+if (rc <= OK) {
+rc = proxy_fixup(r);
+}
+if (rc != OK) {
+r->filename = old_filename;
+r->proxyreq = 0;
+}
 }
-if (rc != OK) {
-return rc;
-}
-} else if (strncmp(r->filename, "proxy:", 6) != 0) {
-return DECLINED;
 }
+else if (strncmp(r->filename, "proxy:", 6) == 0) {
+rc = OK;
+}
+if (rc != OK) {
+return rc;
+}
 
 /* handle max-forwards / OPTIONS / TRACE */
 if ((str = apr_table_get(r->headers_in, "Max-Forwards"))) {
Index: modules/proxy/mod_proxy.h
===
--- modules/proxy/mod_proxy.h	(revision 1918625)
+++ modules/proxy/mod_proxy.h	(working copy)
@@ -1003,6 +1003,16 @@ PROXY_DECLARE(proxy_balancer_shared *) ap_proxy_fi
  proxy_balancer *balancer,
  unsigned int *index);
 
+/*
+ * In the case of the reverse proxy, we need to see if we
+ * were passed a UDS url (eg: from mod_proxy) and adjust uds_path
+ * as required.  
+ * @param rcurrent request
+ * @param url  request url to be fixed
+ * @return OK if fixed up, DECLINED if not UDS, or an HTTP_XXX error
+ */
+PROXY_DECLARE(int) ap_proxy_fixup_uds_filename(request_rec *r, char **url);
+
 /**
  * Get the most suitable worker and/or balancer for the request
  * @param worker   worker used for processing request
Index: modules/proxy/proxy_util.c
===
--- modules/proxy/proxy_util.c	(revision 1918625)
+++ modules/proxy/proxy_util.c	(working copy)
@@ -2429,7 +2429,7 @@ static int ap_proxy_retry_worker(const char *proxy
  * were passed a UDS url (eg: from mod_proxy) and adjust uds_path
  * as required.  
  */
-static int fix_uds_filename(request_rec *r, char **url) 
+PROXY_DECLARE(int) ap_proxy_fixup_uds_filename(request_rec *r, char **url) 
 {
 char *uds_url = r->filename + 6, *origin_url;
 
@@ -2452,7 +2452,7 @@ static int ap_proxy_retry_worker(const char *proxy
 if (!uds_path) {
 ap_log_rerror(APLOG_MARK, APLOG_ERR, 0, r, APLOGNO(10292)
 "Invalid proxy UDS filename (%s)", r->filename);
-return 0;
+return HTTP_BAD_REQUEST;
 }
 apr_table_setn(r->notes, "uds_path", uds_path);
 
@@ -2464,8 +2464,1

Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Helmut K. C. Tessarek



On 2024-06-25 18:34, Eric Covener wrote:

In both GH (read only mirror) and SVN, there is a 2.4.x branch and a tag per rc.


Thank you.


OpenPGP_signature.asc
Description: OpenPGP digital signature


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Eric Covener
> The attached might work, currently testing but sending early if you
> want to try too.

looks like proxy: is stripped off after the new call and needs to be
added back in?

-- 
Eric Covener
cove...@gmail.com


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Eric Covener
On Tue, Jun 25, 2024 at 6:34 PM Eric Covener  wrote:
>
> On Tue, Jun 25, 2024 at 6:32 PM Helmut K. C. Tessarek
>  wrote:
> >
> > On 2024-06-25 18:25, Eric Covener wrote:
> > > As discussed in this thread earlier, the commit that jumped to mind
> > > was not in the rc's (and was hence obviously not the culprit)
> >
> > I missed that, sorry.
> >
> > My question still stands. What is the barnch from which the rcs are cut?
>
> In both GH (read only mirror) and SVN, there is a 2.4.x branch and a tag per 
> rc.

Whoops,e.g. https://github.com/apache/httpd/tree/2.4.60-rc2-candidate


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Eric Covener
On Tue, Jun 25, 2024 at 6:32 PM Helmut K. C. Tessarek
 wrote:
>
> On 2024-06-25 18:25, Eric Covener wrote:
> > As discussed in this thread earlier, the commit that jumped to mind
> > was not in the rc's (and was hence obviously not the culprit)
>
> I missed that, sorry.
>
> My question still stands. What is the barnch from which the rcs are cut?

In both GH (read only mirror) and SVN, there is a 2.4.x branch and a tag per rc.



-- 
Eric Covener
cove...@gmail.com


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Helmut K. C. Tessarek

On 2024-06-25 18:25, Eric Covener wrote:

As discussed in this thread earlier, the commit that jumped to mind
was not in the rc's (and was hence obviously not the culprit)


I missed that, sorry.

My question still stands. What is the barnch from which the rcs are cut?


OpenPGP_signature.asc
Description: OpenPGP digital signature


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Eric Covener
On Tue, Jun 25, 2024 at 6:02 PM Helmut K. C. Tessarek
 wrote:
>
> I tried with rc2, since rc3 is not available yet.
>
> Same issue.
>
> On 2024-06-25 08:44, Eric Covener wrote:
> >   - loglevel trace8 from this request
>
> Yeah, I am not putting that online for public consumption.
> I'll send it to your email address.
>
> >   - whether backing out
> > https://github.com/apache/httpd/commit/9494aa8d52e3c263bc0413b77ac8a73b0d524388.diff
> > from the candidate helps?
>
> Ok. I tried to find this code in rc2, but didn't. I seriously have no
> clue how development works with this project. Do you have a branch that
> I can use and just revert a commit? Trunk is not compatible with the
> 2.4.59 or 2.4.60 code. Referencing random commits from trunk is useless.

As discussed in this thread earlier, the commit that jumped to mind
was not in the rc's (and was hence obviously not the culprit)


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Yann Ylavic
On Tue, Jun 25, 2024 at 11:35 PM Eric Covener  wrote:
>
> On Tue, Jun 25, 2024 at 5:22 PM Eric Covener  wrote:
> >
> > On Tue, Jun 25, 2024 at 5:06 PM Eric Covener  wrote:
> > >
> > > On Tue, Jun 25, 2024 at 4:35 PM Helmut K. C. Tessarek
> > >  wrote:
> > > >
> > > > On 2024-06-25 02:53, Ruediger Pluem wrote:
> > > > > Can you provide more details on your configuration how you forward 
> > > > > stuff to fcgi and what request you made?
> > > >
> > > > Very simple:
> > > >
> > > > 
> > > >  SetHandler  "proxy:unix:/run/php82-fpm/xxx.sock|fcgi://xxx"
> > > > 
> > >
> > > I think this new path needs to account for UDS.  The URL is e.g.
> > > "unix:/tmp/fake.sock|fcgi://xxx..." during proxy_fcgi_canon so it
> > > declines, then the core handler reports the error on the
> > > pseudo-filename.
> >
> > Or for now, skip the newly added proxy_fixup() and escape just
> > r->filename before it gets the handler prefixed to it?
>
> These all seem hairy. We can do something more targetted from the
> start to decrease the fallout?

The attached might work, currently testing but sending early if you
want to try too.
Index: include/ap_mmn.h
===
--- include/ap_mmn.h	(revision 1918625)
+++ include/ap_mmn.h	(working copy)
@@ -601,6 +601,7 @@
  * 20120211.131 (2.4.59-dev) Add DAV_WALKTYPE_TOLERANT
  * 20120211.131 (2.4.60-dev) Add ap_set_content_type_ex(), ap_filepath_merge(),
  *   and AP_REQUEST_TRUSTED_CT BNOTE.
+ * 20120211.133 (2.4.60-dev) Add ap_proxy_fixup_uds_filename()
  */
 
 #define MODULE_MAGIC_COOKIE 0x41503234UL /* "AP24" */
@@ -608,7 +609,7 @@
 #ifndef MODULE_MAGIC_NUMBER_MAJOR
 #define MODULE_MAGIC_NUMBER_MAJOR 20120211
 #endif
-#define MODULE_MAGIC_NUMBER_MINOR 131 /* 0...n */
+#define MODULE_MAGIC_NUMBER_MINOR 133 /* 0...n */
 
 /**
  * Determine if the server's current MODULE_MAGIC_NUMBER is at least a
Index: modules/proxy/mod_proxy.c
===
--- modules/proxy/mod_proxy.c	(revision 1918625)
+++ modules/proxy/mod_proxy.c	(working copy)
@@ -1298,7 +1298,7 @@ static int proxy_handler(request_rec *r)
 ap_get_module_config(sconf, &proxy_module);
 apr_array_header_t *proxies = conf->proxies;
 struct proxy_remote *ents = (struct proxy_remote *) proxies->elts;
-int i, rc, access_status;
+int rc = DECLINED, access_status, i;
 int direct_connect = 0;
 const char *str;
 apr_int64_t maxfwd;
@@ -1314,22 +1314,30 @@ static int proxy_handler(request_rec *r)
 }
 
 if (!r->proxyreq) {
-rc = DECLINED;
 /* We may have forced the proxy handler via config or .htaccess */
 if (r->handler &&
 strncmp(r->handler, "proxy:", 6) == 0 &&
 strncmp(r->filename, "proxy:", 6) != 0) {
+char *old_filename = r->filename;
+/* Still need to fixup/canonicalize r->filename */
 r->proxyreq = PROXYREQ_REVERSE;
 r->filename = apr_pstrcat(r->pool, r->handler, r->filename, NULL);
-/* Still need to fixup/canonicalize r->filename */
-rc = proxy_fixup(r);
+rc = ap_proxy_fixup_uds_filename(r, &r->filename);
+if (rc <= OK) {
+rc = proxy_fixup(r);
+}
+if (rc != OK) {
+r->filename = old_filename;
+r->proxyreq = 0;
+}
 }
-if (rc != OK) {
-return rc;
-}
-} else if (strncmp(r->filename, "proxy:", 6) != 0) {
-return DECLINED;
 }
+else if (strncmp(r->filename, "proxy:", 6) == 0) {
+rc = OK;
+}
+if (rc != OK) {
+return rc;
+}
 
 /* handle max-forwards / OPTIONS / TRACE */
 if ((str = apr_table_get(r->headers_in, "Max-Forwards"))) {
Index: modules/proxy/mod_proxy.h
===
--- modules/proxy/mod_proxy.h	(revision 1918625)
+++ modules/proxy/mod_proxy.h	(working copy)
@@ -1003,6 +1003,16 @@ PROXY_DECLARE(proxy_balancer_shared *) ap_proxy_fi
  proxy_balancer *balancer,
  unsigned int *index);
 
+/*
+ * In the case of the reverse proxy, we need to see if we
+ * were passed a UDS url (eg: from mod_proxy) and adjust uds_path
+ * as required.  
+ * @param rcurrent request
+ * @param url  request url to be fixed
+ * @return OK if fixed up, DECLINED if not UDS, or an HTTP_XXX error
+ */
+PROXY_DECLARE(int) ap_proxy_fixup_uds_filename(request_rec *r, char **url);
+
 /**
  * Get the most suitable worker and/or balancer for the request
  * @param worker   worker used for processing request
Index: modules/proxy/proxy_util.c
===
--- modules/proxy/proxy_util.c	(revision 191862

Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Helmut K. C. Tessarek

I tried with rc2, since rc3 is not available yet.

Same issue.

On 2024-06-25 08:44, Eric Covener wrote:

  - loglevel trace8 from this request


Yeah, I am not putting that online for public consumption.
I'll send it to your email address.


  - whether backing out
https://github.com/apache/httpd/commit/9494aa8d52e3c263bc0413b77ac8a73b0d524388.diff
from the candidate helps?


Ok. I tried to find this code in rc2, but didn't. I seriously have no 
clue how development works with this project. Do you have a branch that 
I can use and just revert a commit? Trunk is not compatible with the 
2.4.59 or 2.4.60 code. Referencing random commits from trunk is useless.


How about a branch 2.4.60 or 2.4.60-rcX that can be used to add/revert 
commits? There must be a branch from which the rcs are created.


Cheers,
  K. C.

--
regards Helmut K. C. Tessarek  KeyID 0x172380A011EF4944
Key fingerprint = 8A55 70C1 BD85 D34E ADBC 386C 1723 80A0 11EF 4944

/*
   Thou shalt not follow the NULL pointer for chaos and madness
   await thee at its end.
*/


OpenPGP_signature.asc
Description: OpenPGP digital signature


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Eric Covener
On Tue, Jun 25, 2024 at 5:22 PM Eric Covener  wrote:
>
> On Tue, Jun 25, 2024 at 5:06 PM Eric Covener  wrote:
> >
> > On Tue, Jun 25, 2024 at 4:35 PM Helmut K. C. Tessarek
> >  wrote:
> > >
> > > On 2024-06-25 02:53, Ruediger Pluem wrote:
> > > > Can you provide more details on your configuration how you forward 
> > > > stuff to fcgi and what request you made?
> > >
> > > Very simple:
> > >
> > > 
> > >  SetHandler  "proxy:unix:/run/php82-fpm/xxx.sock|fcgi://xxx"
> > > 
> >
> > I think this new path needs to account for UDS.  The URL is e.g.
> > "unix:/tmp/fake.sock|fcgi://xxx..." during proxy_fcgi_canon so it
> > declines, then the core handler reports the error on the
> > pseudo-filename.
>
> Or for now, skip the newly added proxy_fixup() and escape just
> r->filename before it gets the handler prefixed to it?

These all seem hairy. We can do something more targetted from the
start to decrease the fallout?



-- 
Eric Covener
cove...@gmail.com


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Eric Covener
On Tue, Jun 25, 2024 at 5:06 PM Eric Covener  wrote:
>
> On Tue, Jun 25, 2024 at 4:35 PM Helmut K. C. Tessarek
>  wrote:
> >
> > On 2024-06-25 02:53, Ruediger Pluem wrote:
> > > Can you provide more details on your configuration how you forward stuff 
> > > to fcgi and what request you made?
> >
> > Very simple:
> >
> > 
> >  SetHandler  "proxy:unix:/run/php82-fpm/xxx.sock|fcgi://xxx"
> > 
>
> I think this new path needs to account for UDS.  The URL is e.g.
> "unix:/tmp/fake.sock|fcgi://xxx..." during proxy_fcgi_canon so it
> declines, then the core handler reports the error on the
> pseudo-filename.

Or for now, skip the newly added proxy_fixup() and escape just
r->filename before it gets the handler prefixed to it?


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Eric Covener
On Tue, Jun 25, 2024 at 4:35 PM Helmut K. C. Tessarek
 wrote:
>
> On 2024-06-25 02:53, Ruediger Pluem wrote:
> > Can you provide more details on your configuration how you forward stuff to 
> > fcgi and what request you made?
>
> Very simple:
>
> 
>  SetHandler  "proxy:unix:/run/php82-fpm/xxx.sock|fcgi://xxx"
> 

I think this new path needs to account for UDS.  The URL is e.g.
"unix:/tmp/fake.sock|fcgi://xxx..." during proxy_fcgi_canon so it
declines, then the core handler reports the error on the
pseudo-filename.

static int proxy_fcgi_canon(request_rec *r, char *url)
{
char *host, sport[7];
const char *err;
char *path;
apr_port_t port, def_port;
fcgi_req_config_t *rconf = NULL;
const char *pathinfo_type = NULL;

if (ap_cstr_casecmpn(url, "fcgi:", 5) == 0) {
url += 5;
}
else {
return DECLINED;
}


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Helmut K. C. Tessarek

On 2024-06-25 08:44, Eric Covener wrote:

  - whether you build both from source / the same way


Yes, and I build always exactly the same way (my own ansible script). 
The only thing that differs is the tarball.



  - Any hint as to why there's a space in that parameter?


Which space and which parameter? I said that I believed there should be 
a space in the log output between the fcgi setup and the resource that 
is accessed.



  - loglevel trace8 from this request
  - whether backing out
https://github.com/apache/httpd/commit/9494aa8d52e3c263bc0413b77ac8a73b0d524388.diff
from the candidate helps?


I can't do that. At least not at the moment. This is a production 
machine. (Yes, I was stupid enough to do that on prod.) I'll reproduce 
on a test VM and get you the info. First I'll try rc3 or whateever the 
latest one is.


Cheers,
  K.C.
--
regards Helmut K. C. Tessarek  KeyID 0x172380A011EF4944
Key fingerprint = 8A55 70C1 BD85 D34E ADBC 386C 1723 80A0 11EF 4944

/*
   Thou shalt not follow the NULL pointer for chaos and madness
   await thee at its end.
*/


OpenPGP_signature.asc
Description: OpenPGP digital signature


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Helmut K. C. Tessarek

On 2024-06-25 02:53, Ruediger Pluem wrote:

Can you provide more details on your configuration how you forward stuff to 
fcgi and what request you made?


Very simple:


SetHandler  "proxy:unix:/run/php82-fpm/xxx.sock|fcgi://xxx"


And then I access my website at https://example.com

Please don't forget that I've been using this configuration for years.
I only change the path to the socket every time when I move to a newer 
PHP release.


Cheers,
  K.C.

--
regards Helmut K. C. Tessarek  KeyID 0x172380A011EF4944
Key fingerprint = 8A55 70C1 BD85 D34E ADBC 386C 1723 80A0 11EF 4944

/*
   Thou shalt not follow the NULL pointer for chaos and madness
   await thee at its end.
*/


OpenPGP_signature.asc
Description: OpenPGP digital signature


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Eric Covener
Vote is abandoned, please look for an RC2 thread soon.

On Tue, Jun 25, 2024 at 10:58 AM Eric Covener  wrote:
>
> On Tue, Jun 25, 2024 at 9:38 AM Yann Ylavic  wrote:
> >
> > On Tue, Jun 25, 2024 at 2:50 PM Eric Covener  wrote:
> > >
> > >  - whether backing out
> > > https://github.com/apache/httpd/commit/9494aa8d52e3c263bc0413b77ac8a73b0d524388.diff
> > > from the candidate helps?
> >
> > I'm not sure r1918553 made it to rc1, did it?
>
> Sorry -- I plan to re-spin to pick it up in the next few hours.



-- 
Eric Covener
cove...@gmail.com


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Eric Covener
On Tue, Jun 25, 2024 at 9:38 AM Yann Ylavic  wrote:
>
> On Tue, Jun 25, 2024 at 2:50 PM Eric Covener  wrote:
> >
> >  - whether backing out
> > https://github.com/apache/httpd/commit/9494aa8d52e3c263bc0413b77ac8a73b0d524388.diff
> > from the candidate helps?
>
> I'm not sure r1918553 made it to rc1, did it?

Sorry -- I plan to re-spin to pick it up in the next few hours.


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Ruediger Pluem



On 6/25/24 3:30 PM, Yann Ylavic wrote:
> On Tue, Jun 25, 2024 at 2:50 PM Eric Covener  wrote:
>>
>>  - whether backing out
>> https://github.com/apache/httpd/commit/9494aa8d52e3c263bc0413b77ac8a73b0d524388.diff
>> from the candidate helps?
> 
> I'm not sure r1918553 made it to rc1, did it?
> 

I cannot see it there.

Regards

Rüdiger


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Yann Ylavic
On Tue, Jun 25, 2024 at 2:50 PM Eric Covener  wrote:
>
>  - whether backing out
> https://github.com/apache/httpd/commit/9494aa8d52e3c263bc0413b77ac8a73b0d524388.diff
> from the candidate helps?

I'm not sure r1918553 made it to rc1, did it?


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Eric Covener
On Tue, Jun 25, 2024 at 2:54 AM Ruediger Pluem  wrote:
>
>
>
> On 6/25/24 8:30 AM, Helmut K. C. Tessarek wrote:
> > Hello,
> >
> > On 2024-06-24 14:28, Eric Covener wrote:
> >> Hi all,
> >>
> >> Please find below the proposed release tarball and signatures:
> >>
> >> https://dist.apache.org/repos/dist/dev/httpd/
> >>
> >> I would like to call a VOTE over the next few days to release
> >> this candidate tarball httpd-2.4.60-rc1 as 2.4.60:
> >> [ ] +1: It's not just good, it's good enough!
> >> [ ] +0: Let's have a talk.
> >> [ ] -1: There's trouble in paradise. Here's what's wrong.
> >
> > It's the first time I tested an rc, but since my current one coredumps on 
> > Fedora 40 (a bug is open), I thought I'd give it a try.
> >
> > Right after starting it, I get permission errors all over the place.
> >
> > Here an example:
> >
> > [core:error] [pid 1385877] AH00132: file permissions deny server access:
> > proxy:unix:/run/php82-fpm/xxx.sock|fcgi://xxx/var/www/index.php
> >
> > As you can see there is no space between the fcgi and the path of the 
> > resource.
> >
> > The last part should be: fcgi://xxx /var/www/index.php
> >
> > Or maybe this is correct and normal way the error message is constructed. 
> > Either way, as soon as I install 2.4.59, the errors are
> > gone. Thus, something is off with this version.
>
> Can you provide more details on your configuration how you forward stuff to 
> fcgi and what request you made?

Also curious about:

 - whether you build both from source / the same way
 - Any hint as to why there's a space in that parameter?
 - loglevel trace8 from this request
 - whether backing out
https://github.com/apache/httpd/commit/9494aa8d52e3c263bc0413b77ac8a73b0d524388.diff
from the candidate helps?


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-25 Thread Stefan Eissing via dev



> Am 24.06.2024 um 20:28 schrieb Eric Covener :
> 
> Hi all,
> 
> Please find below the proposed release tarball and signatures:
> 
> https://dist.apache.org/repos/dist/dev/httpd/
> 
> I would like to call a VOTE over the next few days to release
> this candidate tarball httpd-2.4.60-rc1 as 2.4.60:
> [ ] +1: It's not just good, it's good enough!
> [ ] +0: Let's have a talk.
> [ ] -1: There's trouble in paradise. Here's what's wrong.

+1 on macOS 14.5.

Thanks for doing the release work!

- Stefan


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-24 Thread Ruediger Pluem



On 6/25/24 8:30 AM, Helmut K. C. Tessarek wrote:
> Hello,
> 
> On 2024-06-24 14:28, Eric Covener wrote:
>> Hi all,
>>
>> Please find below the proposed release tarball and signatures:
>>
>> https://dist.apache.org/repos/dist/dev/httpd/
>>
>> I would like to call a VOTE over the next few days to release
>> this candidate tarball httpd-2.4.60-rc1 as 2.4.60:
>> [ ] +1: It's not just good, it's good enough!
>> [ ] +0: Let's have a talk.
>> [ ] -1: There's trouble in paradise. Here's what's wrong.
> 
> It's the first time I tested an rc, but since my current one coredumps on 
> Fedora 40 (a bug is open), I thought I'd give it a try.
> 
> Right after starting it, I get permission errors all over the place.
> 
> Here an example:
> 
> [core:error] [pid 1385877] AH00132: file permissions deny server access:
> proxy:unix:/run/php82-fpm/xxx.sock|fcgi://xxx/var/www/index.php
> 
> As you can see there is no space between the fcgi and the path of the 
> resource.
> 
> The last part should be: fcgi://xxx /var/www/index.php
> 
> Or maybe this is correct and normal way the error message is constructed. 
> Either way, as soon as I install 2.4.59, the errors are
> gone. Thus, something is off with this version.

Can you provide more details on your configuration how you forward stuff to 
fcgi and what request you made?

Regards

Rüdiger



Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-24 Thread Helmut K. C. Tessarek

Hello,

On 2024-06-24 14:28, Eric Covener wrote:

Hi all,

Please find below the proposed release tarball and signatures:

https://dist.apache.org/repos/dist/dev/httpd/

I would like to call a VOTE over the next few days to release
this candidate tarball httpd-2.4.60-rc1 as 2.4.60:
[ ] +1: It's not just good, it's good enough!
[ ] +0: Let's have a talk.
[ ] -1: There's trouble in paradise. Here's what's wrong.


It's the first time I tested an rc, but since my current one coredumps 
on Fedora 40 (a bug is open), I thought I'd give it a try.


Right after starting it, I get permission errors all over the place.

Here an example:

[core:error] [pid 1385877] AH00132: file permissions deny server access: 
proxy:unix:/run/php82-fpm/xxx.sock|fcgi://xxx/var/www/index.php


As you can see there is no space between the fcgi and the path of the 
resource.


The last part should be: fcgi://xxx /var/www/index.php

Or maybe this is correct and normal way the error message is 
constructed. Either way, as soon as I install 2.4.59, the errors are 
gone. Thus, something is off with this version.


Cheers,
  K. C.


--
regards Helmut K. C. Tessarek  KeyID 0x172380A011EF4944
Key fingerprint = 8A55 70C1 BD85 D34E ADBC 386C 1723 80A0 11EF 4944

/*
   Thou shalt not follow the NULL pointer for chaos and madness
   await thee at its end.
*/


OpenPGP_signature.asc
Description: OpenPGP digital signature


Re: [VOTE] Release httpd-2.4.60-rc1 as httpd-2.4.60

2024-06-24 Thread Eric Covener
> this candidate tarball httpd-2.4.60-rc1 as 2.4.60:
> [X] +1: It's not just good, it's good enough!
> [ ] +0: Let's have a talk.
> [ ] -1: There's trouble in paradise. Here's what's wrong.

Looks good on AIX/xlc/ppc (only known issues with system perl and t/ssl/proxy.t)