| 03:42.41 | *** join/#asterisk-dev electronic_eel (~quassel@213.240.182.236) |
| 05:15.35 | *** join/#asterisk-dev tsearle (tsearlenim@gateway/shell/matrix.org/x-mbcvuglnhhpsqyzd) |
| 05:15.35 | *** mode/#asterisk-dev [+o tsearle] by ChanServ |
| 05:15.54 | *** join/#asterisk-dev DanJenkins[m] (dannimblea@gateway/shell/matrix.org/x-hejpnguilpgubjae) |
| 05:15.54 | *** mode/#asterisk-dev [+o DanJenkins[m]] by ChanServ |
| 13:29.27 | jkroon | when compiling a module external to asterisk, is there a clean way to determine the version of asterisk being compiled against during the pre-processor? |
| 15:03.42 | *** join/#asterisk-dev kharwell (uid358942@gateway/web/irccloud.com/x-kuhedtwvbsakvqfy) |
| 15:03.42 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 15:35.09 | *** join/#asterisk-dev bford (uid283514@gateway/web/irccloud.com/x-sujdrtgiuachspcr) |
| 15:35.09 | *** mode/#asterisk-dev [+o bford] by ChanServ |
| 17:23.56 | jkroon | kharwell, looking at 3d1bf3c537bba0416f691f48165fdd0a32554e8a, the change to app.h, what exactly was that aimed at? The old code works for me with gcc 10 (both c and c++) but the new code breaks when using c++ ...? Was there a specific issue being addressed there? |
| 17:26.54 | jkroon | I've done a small test file now (https://dpaste.org/NqRH) and the error I get (both g++-9 and g++-10): struct_hacks.C:6:28: error: types may not be defined in âsizeofâ expressions |
| 17:27.00 | jkroon | -x c works fine. |
| 17:27.33 | kharwell | I thought gcc 10 complained about the zero sized array not being at the end of the struct. |
| 17:27.55 | kharwell | It's been a while so I don't exactly recall if you are finding it does not |
| 17:29.19 | jkroon | am aiming to set something up for review, just want to make sure I don't make matters worse. |
| 17:30.57 | jkroon | https://dpaste.org/ydJ3 works with "gcc-10.2.0 -x c++ -Wall -c struct_hacks.C" |
| 17:31.08 | jkroon | just warns about var not actually being used. |
| 17:32.12 | jkroon | Will you be receptive if I push a review reverting just that one change from app.h from your larger commit? |
| 17:32.55 | kharwell | lemme see. I'm recompiling now to see if I get an error |
| 17:33.31 | jkroon | entire asterisk uses it as "C", I'm building an external C++ module against it. |
| 17:33.45 | jkroon | new code is find in "C", but not "C++" |
| 17:34.21 | kharwell | and current patch up for review doesn't fix it? https://gerrit.asterisk.org/c/asterisk/+/15399 |
| 17:35.18 | jkroon | didn't see that ... it's much overkill IMHO. |
| 17:35.58 | jkroon | simply reverting your original code change to app.h does the trick. |
| 17:37.15 | jkroon | plus, in that patch the usage of the macro is different for C and C++ users, making things a little confusion in my opinion. |
| 17:37.50 | kharwell | huh yeah I removed the change, and re-compiled and no errors. So don't remember what the motivation was heh. Albeit it's not good putting a zero sized array in the middle of a struct |
| 17:38.19 | jkroon | It's odd, but there's nothing technically *wrong* with it. |
| 17:38.30 | kharwell | since it seemingly wasn't a problem before, and is causing a problem now then I'd be fine just reverting that portion back. |
| 17:38.36 | jkroon | it might be an issue if there is static analysis done and someone indexes into the array ... |
| 17:39.35 | jkroon | Is it OK for me to -1 on that just to notify that I've seen it and think there is a simpler way and then submit a "competing" review? |
| 17:40.57 | kharwell | yeah that's fine, or if you simply want to comment on the current review and state reverting it fixes it is fine too. |
| 17:41.39 | kharwell | but if easier for you to just post a new review that'll work as well |
| 17:42.14 | jkroon | I'm just compile testing with both gcc 9 and 10 "quickly" |
| 17:44.21 | kharwell | good idea. I'll double check with gcc 9 too |
| 17:47.27 | jkroon | and of course now it insists on building with clang ... as that was the last patches I did. |
| 17:49.42 | kharwell | no problem with gcc 9 for me either |
| 17:50.11 | jkroon | compile issues in chan_oss and chan_audiosocket ... I'm assuming I can safely ignore since it's probably dependency issues on my side. |
| 17:52.34 | jkroon | no, that was caused by me being unable to {} match. |
| 18:01.59 | jkroon | https://gerrit.asterisk.org/c/asterisk/+/15441, thanks for the assist. |
| 18:03.57 | kharwell | no problem |
| 19:24.04 | jkroon | kharwell, another one, looks like C++ can now do the whole "struct foo bar = { .field = value, };" thing. In particular it looks like AST_MODULE_INFO could benefit from that. This works on g++ 9 and 10 at least. |
| 19:24.14 | jkroon | so how far back do we support? |
| 19:24.46 | jkroon | ditto with clang. |
| 19:26.23 | kharwell | whatcha mean how far back? |
| 19:26.38 | jkroon | what's the oldest gcc version we care about being able to compile on? |
| 19:27.29 | file | supporting things doesn't work like that, it's based on distributions so you'd have to map that to an old version |
| 19:27.30 | kharwell | oh hrm. not sure actually |
| 19:27.49 | file | The Asterisk project supports 32-bit and 64-bit x86 platforms using non-end of life CentOS, RHEL, Fedora, Ubuntu, and Debian Linux distributions. |
| 19:28.04 | jkroon | file, ok. |
| 19:31.25 | kharwell | is g++ an officially supported compiler though? |
| 19:31.52 | file | I would say C++ support is best effort, community supported |
| 19:32.24 | jkroon | kharwell, guessing no, since asterisk itself is purely c, however, i think enabling c++ to use asterisk headers and thus link against + with asterisk is a nice to have. |
| 19:34.33 | kharwell | ya I was more wondering with regards to backwards compatibility. If not officially supported then maybe not as big of deal if something doesn't compile for it on say CentOS 7 that has an older version |
| 19:35.10 | file | I think the standard supported platforms should be applied to it |
| 19:35.22 | jkroon | cool. if I bump into something that gives actual warnings/errors I'll see it's something that can be fixed/improved, otherwise, leave alone. |
| 19:36.31 | kharwell | file: yeah prob the most reasonable thing |
| 22:17.44 | *** join/#asterisk-dev jkroon (~jkroon@165.16.203.102) |
| 22:17.44 | *** mode/#asterisk-dev [+o jkroon] by ChanServ |
| 22:51.40 | *** join/#asterisk-dev jkroon (~jkroon@165.16.203.102) |
| 22:51.40 | *** mode/#asterisk-dev [+o jkroon] by ChanServ |