Re: [Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support
Hi, It's lots of text that is totally uninteresting if it succeeds, but that > you need to scroll over when looking for the reason why a build fails - > which is the reason why it is the way it is today. > I have sent v2 which doesn't change the behavior for non-Windows builds. If you volunteer to update MSVC every time we add or change a source file, > I'm happy to send you a HEADS UP notice. But I am not doing it, and I will > not request it from any contributor who is not using windows. > All right, that would work for me. We don't add C files that often, I could take care of fixing Visual Studio build after getting a mail from travis / buildbot (when we'll have MSVC there). v2: https://patchwork.openvpn.net/patch/886/ happy travis: https://travis-ci.org/lstipakov/openvpn/builds/608949725 -- -Lev ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support
Hi On Thu, Nov 7, 2019 at 7:43 AM Lev Stipakov wrote: > Hi, > > >> I'm a bit unhappy with that one, as it changes behaviour for all >> non-windows >> builds (including all the openssl build output even if it succeeds). >> > > The only place it changes behavior is this > > install: >- if [ ! -z "${CHOST}" ]; then unset CC; fi > - - .travis/build-deps.sh > build-deps.log 2>&1 || (cat build-deps.log && > exit 1) > + - .travis/build-deps.sh > > I don't see it as an issue to print output when building dependencies. The > reason why > it is done is that travis aborts build if there is no output for more than > 10 minutes. > > >> Besides this, we need to fix this whole MSVC mess - all other platforms >> are just done with "add new source file to the Makefile.ac" and done >> (including mingw builds), and then MSVC is broken again, and this will >> happen again and again. > > >> Is there no reasonable way to build these project files from Makefile.ac? >> > > I see no reasonable way. Selva, Simon - opinions? > Personally, I come from the Unix world, work on Windows only out of necessity, and either avoid MSVC or leave it to others to figure out as far as possible. So my opinion may not count for much. That said, short of moving to a more Windows-friendly build system such as CMake, I see no good options. But the status quo looks good enough to me -- i.e., just do a patch to fix the project files when you notice a missing entry. Selva ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support
Hi, On 07/11/2019 14:36, Simon Rozman wrote: > I revoke this “master.c” idea. It makes incremental compiling ridiculously > slow. It might work for production builds, but definitely not for development. It also breaks scoping of variables/function. So, I am glad you found your own reason to reject that :-) -- Antonio Quartulli ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support
I revoke this “master.c” idea. It makes incremental compiling ridiculously slow. It might work for production builds, but definitely not for development. Best regards, Simon From: Simon Rozman [mailto:si...@rozman.si] Sent: Thursday, November 7, 2019 2:06 PM To: 'Lev Stipakov' ; 'Gert Doering' Cc: 'Antonio Quartulli' ; 'openvpn-devel' Subject: Re: [Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support We could introduce a master.c file which would include: #include "argv.c" #include "auth_token.c" #include "base64.c" #include "block_dns.c" #include "buffer.c" . . . #include "status.c" #include "tls_crypt.c" #include "tun.c" #include "win32.c" And then have Makefile.ac and MSVC project files compile the master.c only. Best regards, Simon Besides this, we need to fix this whole MSVC mess - all other platforms are just done with "add new source file to the Makefile.ac" and done (including mingw builds), and then MSVC is broken again, and this will happen again and again. Is there no reasonable way to build these project files from Makefile.ac? I see no reasonable way. Selva, Simon - opinions? We could probably have a templatized project file and script which inserts source files to there based on content of Makefile.ac, but I don't like it. I consider necessity of modifying VS project a lesser evil. (you open it and it just works) -Lev smime.p7s Description: S/MIME cryptographic signature ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support
Hi, On 07/11/2019 12:42, Lev Stipakov wrote: Hi, I'm a bit unhappy with that one, as it changes behaviour for all non-windows builds (including all the openssl build output even if it succeeds). The only place it changes behavior is this install: - if [ ! -z "${CHOST}" ]; then unset CC; fi - - .travis/build-deps.sh > build-deps.log 2>&1 || (cat build-deps.log && exit 1) + - .travis/build-deps.sh I don't see it as an issue to print output when building dependencies. The reason why it is done is that travis aborts build if there is no output for more than 10 minutes. Can travis be instructed to wait longer ? ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support
We could introduce a master.c file which would include: #include "argv.c" #include "auth_token.c" #include "base64.c" #include "block_dns.c" #include "buffer.c" . . . #include "status.c" #include "tls_crypt.c" #include "tun.c" #include "win32.c" And then have Makefile.ac and MSVC project files compile the master.c only. Best regards, Simon Besides this, we need to fix this whole MSVC mess - all other platforms are just done with "add new source file to the Makefile.ac" and done (including mingw builds), and then MSVC is broken again, and this will happen again and again. Is there no reasonable way to build these project files from Makefile.ac? I see no reasonable way. Selva, Simon - opinions? We could probably have a templatized project file and script which inserts source files to there based on content of Makefile.ac, but I don't like it. I consider necessity of modifying VS project a lesser evil. (you open it and it just works) -Lev smime.p7s Description: S/MIME cryptographic signature ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support
Hi, On Thu, Nov 07, 2019 at 02:42:17PM +0200, Lev Stipakov wrote: > > I'm a bit unhappy with that one, as it changes behaviour for all > > non-windows > > builds (including all the openssl build output even if it succeeds). > > The only place it changes behavior is this > > install: >- if [ ! -z "${CHOST}" ]; then unset CC; fi > - - .travis/build-deps.sh > build-deps.log 2>&1 || (cat build-deps.log && > exit 1) > + - .travis/build-deps.sh > > I don't see it as an issue to print output when building dependencies. The > reason why > it is done is that travis aborts build if there is no output for more than > 10 minutes. It's lots of text that is totally uninteresting if it succeeds, but that you need to scroll over when looking for the reason why a build fails - which is the reason why it is the way it is today. [..] > We could probably have a templatized project file and script which inserts > source files to there > based on content of Makefile.ac, but I don't like it. I consider necessity > of modifying VS project a lesser evil. > > (you open it and it just works) I'm not taking responsibility for not breaking MSVC builds. This is a world alien to me - if you want me to not break things, it needs to be "without fiddling XML files". If you volunteer to update MSVC every time we add or change a source file, I'm happy to send you a HEADS UP notice. But I am not doing it, and I will not request it from any contributor who is not using windows. gert -- "If was one thing all people took for granted, was conviction that if you feed honest figures into a computer, honest figures come out. Never doubted it myself till I met a computer with a sense of humor." Robert A. Heinlein, The Moon is a Harsh Mistress Gert Doering - Munich, Germany g...@greenie.muc.de signature.asc Description: PGP signature ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support
Hi, > I'm a bit unhappy with that one, as it changes behaviour for all > non-windows > builds (including all the openssl build output even if it succeeds). > The only place it changes behavior is this install: - if [ ! -z "${CHOST}" ]; then unset CC; fi - - .travis/build-deps.sh > build-deps.log 2>&1 || (cat build-deps.log && exit 1) + - .travis/build-deps.sh I don't see it as an issue to print output when building dependencies. The reason why it is done is that travis aborts build if there is no output for more than 10 minutes. > Besides this, we need to fix this whole MSVC mess - all other platforms > are just done with "add new source file to the Makefile.ac" and done > (including mingw builds), and then MSVC is broken again, and this will > happen again and again. > Is there no reasonable way to build these project files from Makefile.ac? > I see no reasonable way. Selva, Simon - opinions? We could probably have a templatized project file and script which inserts source files to there based on content of Makefile.ac, but I don't like it. I consider necessity of modifying VS project a lesser evil. (you open it and it just works) -Lev ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support
Hi, On Thu, Nov 07, 2019 at 02:18:25PM +0200, Lev Stipakov wrote: > This broke (again) Visual Studio build: > > > error LNK2019: unresolved external symbol vlan_process_outgoing_tun > referenced in function multi_process_outgoing_tun Can you please send a patch that adds vlan.c to the list of source code modules to be compiled and linked? > Can we please ack and merge this patch > https://patchwork.openvpn.net/patch/868/ to catch these things earlier? I'm a bit unhappy with that one, as it changes behaviour for all non-windows builds (including all the openssl build output even if it succeeds). Besides this, we need to fix this whole MSVC mess - all other platforms are just done with "add new source file to the Makefile.ac" and done (including mingw builds), and then MSVC is broken again, and this will happen again and again. Is there no reasonable way to build these project files from Makefile.ac? gert -- "If was one thing all people took for granted, was conviction that if you feed honest figures into a computer, honest figures come out. Never doubted it myself till I met a computer with a sense of humor." Robert A. Heinlein, The Moon is a Harsh Mistress Gert Doering - Munich, Germany g...@greenie.muc.de signature.asc Description: PGP signature ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support
This broke (again) Visual Studio build: > error LNK2019: unresolved external symbol vlan_process_outgoing_tun referenced in function multi_process_outgoing_tun Can we please ack and merge this patch https://patchwork.openvpn.net/patch/868/ to catch these things earlier? ke 6. marrask. 2019 klo 22.42 Gert Doering (g...@greenie.muc.de) kirjoitti: > Acked-by: Gert Doering > > Stared at the code (twice now), run t_client and t_server tests. > > This patch does not really *do* much yet, but it lays the groundwork > for future work - the "broadcast only to clients in the same vlan" > part is there, but it's always called with "0" (= all clients). As > far as I can see, the only notable behavioural change we have so far > is "if a client is assigned a pvid (!= the global pvid), it will not > be able to communicate with the TAP interface" (check in vlan.c, > vlan_process_outgoing_tun()), but client-to-client is still allowed, > and there is no per-vlan MAC learning yet either. > > (Most notably, it only adds options and code relevant for TAP mode) > > Your patch has been applied to the master branch. > > commit 99f28081477ca325a14b13c38abec2c9b619eb01 > Author: Antonio Quartulli > Date: Wed Oct 9 16:34:15 2019 +0200 > > VLAN: add basic VLAN tagging support > > Signed-off-by: Fabian Knittel > Signed-off-by: Antonio Quartulli > Acked-by: Gert Doering > Message-Id: <20191009143422.9419-...@unstable.cc> > URL: > https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg18924.html > Signed-off-by: Gert Doering > > > -- > kind regards, > > Gert Doering > > > > ___ > Openvpn-devel mailing list > Openvpn-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/openvpn-devel > -- -Lev ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
[Openvpn-devel] [PATCH applied] Re: VLAN: add basic VLAN tagging support
Acked-by: Gert Doering Stared at the code (twice now), run t_client and t_server tests. This patch does not really *do* much yet, but it lays the groundwork for future work - the "broadcast only to clients in the same vlan" part is there, but it's always called with "0" (= all clients). As far as I can see, the only notable behavioural change we have so far is "if a client is assigned a pvid (!= the global pvid), it will not be able to communicate with the TAP interface" (check in vlan.c, vlan_process_outgoing_tun()), but client-to-client is still allowed, and there is no per-vlan MAC learning yet either. (Most notably, it only adds options and code relevant for TAP mode) Your patch has been applied to the master branch. commit 99f28081477ca325a14b13c38abec2c9b619eb01 Author: Antonio Quartulli Date: Wed Oct 9 16:34:15 2019 +0200 VLAN: add basic VLAN tagging support Signed-off-by: Fabian Knittel Signed-off-by: Antonio Quartulli Acked-by: Gert Doering Message-Id: <20191009143422.9419-...@unstable.cc> URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg18924.html Signed-off-by: Gert Doering -- kind regards, Gert Doering ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel