Re: svn commit: r1451633 - in /httpd/httpd/trunk: include/ap_mmn.h modules/proxy/mod_proxy.h modules/proxy/proxy_util.c
Is there any way you could add the #ifdef stuff? Since I lack a Windows and/or NetWare system, it would be better, I think, if someone who did actually fixed this instead of us simply removing it, and the person who fixed it was able to test the fix :) Thx! On Mar 1, 2013, at 9:10 PM, Guenter Knauf fua...@apache.org wrote: Hi Jim, Am 01.03.2013 17:21, schrieb j...@apache.org: Author: jim Date: Fri Mar 1 16:21:49 2013 New Revision: 1451633 URL: http://svn.apache.org/r1451633 Log: Add in rough uds support (Bugx 54101) from Blaise Tarrblaise.t...@gmail.com Modified: httpd/httpd/trunk/include/ap_mmn.h httpd/httpd/trunk/modules/proxy/mod_proxy.h httpd/httpd/trunk/modules/proxy/proxy_util.c please revert this - it breaks winsock platforms which lack support for unix sockets and sockaddr_un (Windows as well as NetWare when build with winsock). If we need this then we must either move that stuff into a separate unix-only file or ifdef everything related to unix sockets with #if APR_HAVE_SYS_UN_H ... Gün.
Re: svn commit: r1451633 - in /httpd/httpd/trunk: include/ap_mmn.h modules/proxy/mod_proxy.h modules/proxy/proxy_util.c
Hopefully r1451905 addressed this, but not having a test platform, I'm only guessing... On Mar 2, 2013, at 9:12 AM, Jim Jagielski j...@jagunet.com wrote: Is there any way you could add the #ifdef stuff? Since I lack a Windows and/or NetWare system, it would be better, I think, if someone who did actually fixed this instead of us simply removing it, and the person who fixed it was able to test the fix :) Thx! On Mar 1, 2013, at 9:10 PM, Guenter Knauf fua...@apache.org wrote: Hi Jim, Am 01.03.2013 17:21, schrieb j...@apache.org: Author: jim Date: Fri Mar 1 16:21:49 2013 New Revision: 1451633 URL: http://svn.apache.org/r1451633 Log: Add in rough uds support (Bugx 54101) from Blaise Tarrblaise.t...@gmail.com Modified: httpd/httpd/trunk/include/ap_mmn.h httpd/httpd/trunk/modules/proxy/mod_proxy.h httpd/httpd/trunk/modules/proxy/proxy_util.c please revert this - it breaks winsock platforms which lack support for unix sockets and sockaddr_un (Windows as well as NetWare when build with winsock). If we need this then we must either move that stuff into a separate unix-only file or ifdef everything related to unix sockets with #if APR_HAVE_SYS_UN_H ... Gün.
[off-topic] ApacheCon is over - thanks for the people I met and the lessons I learned.
This is perhaps a bit off-topic, but since we don't really have a defocus@ mailing list, I thought I'd just send my email to this list. So, ApacheCon North America 2013 is finally over, and what a great conference it was. I want to thank the people I met for being so awesome (Rainer, Mike, Rich, Igor, Jim, Roy, Jeff, Bill, Chris, Daniel, the list goes on...) and for doing some very impressive talks that showed that the httpd project is still very much alive and kicking. I was especially impressed by Jim and Rich's talks, and I hope that in time, I can become, if not as awesome then just a hint more awesome at doing presentations. Furthermore, it was a great pleasure to meet some of the people that don't necessarily contribute, but simply use httpd, and get some feedback from them on what we can improve and what we're doing right. I did see quite a few small fixes during the conference, which comes to show that interaction between committers and users do matter, even when it's more informal as the various barcamps and in-the-hallway meetings were. There were also interesting questions being asked at the talks, which will in time lead to some commits to fix or enhance various bits and pieces of the httpd code. Personally, I'll be looking into doing a sync of the trunk and the 2.4 branch with regards to mod_lua, as well as look into perhaps extending the number of places where hooks can be positioned early/late, as per Bill's suggestion at the mod_lua talk. On a side note, I was humored as well as pleased to be asked about whether mod_lua is a viable alternative to node.js - of course this depends immensely on what exactly you're trying to accomplish, but hey...anything is possible with mod_lua, or at least that's my overly biased opinion ;) So, what does the future hold for httpd? Well, one of the interesting times during the conference was Roy's tumble-weed moment where he inquired on httpd 3.0 and the entire ballroom went dead silent. I think this illustrates that while we do have a vibrant and healthy community, we may not know exactly where we're going in the future, and this may be a concern to some of our users. Will we be rewriting httpd completely, will we just do a 2.6 that follows in the footsteps of 2.4, or...what will we be doing? At some point, it would be nice if we could either get a working group going or start a discussion of what we'd want the next httpd to be and how we'll get there. I know much of this hinges on the HTTP/2.0 work being done by Roy et al. and so I'm not trying to force people to think about it right now, but at some point, we need to figure out where we want to be once HTTP/2.0 becomes a reality. Maybe we'll be the first major web server supporting it? Who knows. Anywho, thanks for a truly awesome ApacheCon, and I really hope to see the same cool httpd gang next time, and hopefully more of the committers that have contributed so much brilliant code to the httpd project over the last year. One thing that I would have liked to have happen was an actual httpd team get-together where we could say hello and talk a bit more about httpd and what we're each working/tinkering on at the moment. This will naturally have to be planned in advance, which seems to be where it all failed. I'll do my best to hash out some plans for the next meetup, so we can get down to business and discuss what awesome plans we have for httpd. I promise, I will work on improving my presentations (my mod_lua talk went okay, but the HRM talk tanked due to me losing my slides - talk about bad luck) and I hope/know that the rest of you will just be awesome as usual next time around. With regards, Daniel.
Re: svn commit: r1451633 - in /httpd/httpd/trunk: include/ap_mmn.h modules/proxy/mod_proxy.h modules/proxy/proxy_util.c
Am 02.03.2013 15:12, schrieb Jim Jagielski: Is there any way you could add the #ifdef stuff? Since I lack a Windows and/or NetWare system, it would be better, I think, if someone who did actually fixed this instead of us simply removing it, and the person who fixed it was able to test the fix :) sure ... 1st it breaks here: CC proxy_util.c ### mwccnlm Compiler: #File: proxy_util.c # - #2405: const socklen_t addrlen = APR_OFFSETOF(struct sockaddr_un, sun_path) # Error: ^^^ # ';' expected # Too many errors printed, aborting program this is because it should be apr_socklen_t; but when I fix this then next is with the struct sockaddr_un ... I've just ifdef'd this problemaric function which makes proxy_util.c compile again for NetWare; lets see what Gregg says for Windows ... http://svn.apache.org/r1451921 Gün.
Re: svn commit: r1451633 - in /httpd/httpd/trunk: include/ap_mmn.h modules/proxy/mod_proxy.h modules/proxy/proxy_util.c
On 3/2/2013 11:21 AM, Guenter Knauf wrote: Am 02.03.2013 15:12, schrieb Jim Jagielski: Is there any way you could add the #ifdef stuff? Since I lack a Windows and/or NetWare system, it would be better, I think, if someone who did actually fixed this instead of us simply removing it, and the person who fixed it was able to test the fix :) sure ... 1st it breaks here: CC proxy_util.c ### mwccnlm Compiler: #File: proxy_util.c # - #2405: const socklen_t addrlen = APR_OFFSETOF(struct sockaddr_un, sun_path) # Error: ^^^ # ';' expected # Too many errors printed, aborting program this is because it should be apr_socklen_t; but when I fix this then next is with the struct sockaddr_un ... I've just ifdef'd this problemaric function which makes proxy_util.c compile again for NetWare; lets see what Gregg says for Windows ... http://svn.apache.org/r1451921 Compiles fine on Windows