00:13.36 | *** join/#crosstool-ng al` (~al@208.85.112.101) |
00:13.36 | *** mode/#crosstool-ng [+o al`] by verne.freenode.net |
05:58.05 | *** join/#crosstool-ng bhundven (~bryan@unaffiliated/bhundven) |
06:36.13 | *** join/#crosstool-ng kos_tom (~thomas@col31-4-88-188-83-94.fbx.proxad.net) |
06:47.13 | *** join/#crosstool-ng y_morin (~ymorin@ARennes-256-1-140-186.w90-32.abo.wanadoo.fr) |
06:47.13 | *** mode/#crosstool-ng [+o y_morin] by ChanServ |
06:49.43 | *** join/#crosstool-ng y_morin (~ymorin@ARennes-256-1-140-186.w90-32.abo.wanadoo.fr) |
06:49.43 | *** mode/#crosstool-ng [+o y_morin] by ChanServ |
07:05.34 | *** join/#crosstool-ng y_morin (~ymorin@ARennes-256-1-140-186.w90-32.abo.wanadoo.fr) |
07:05.34 | *** mode/#crosstool-ng [+o y_morin] by ChanServ |
09:00.40 | bhundven | wtf... cell-card unlocking services are sketchy |
11:48.25 | *** join/#crosstool-ng smartin_ (52e7e007@gateway/web/freenode/ip.82.231.224.7) |
14:39.39 | *** join/#crosstool-ng al`_ (~al@208.85.112.101) |
15:43.31 | *** join/#crosstool-ng linuxjacques (6217fec1@nslu2-linux/jacques) |
17:23.24 | *** join/#crosstool-ng Buglouse (~bug@cpe-107-8-145-21.wi.res.rr.com) |
18:26.02 | bhundven|afk | y_morin: maybe I'm wrong? but why do we need --enable-symvers=gnu? |
18:26.15 | bhundven | lol |
18:34.04 | y_morin | bhundven: just a few minutes here, then dinner... |
18:34.10 | y_morin | bhundven: where's that ? |
18:34.31 | y_morin | OK, got it. |
18:36.29 | y_morin | bhundven: is that the same issue you are seeing? http://www.mail-archive.com/gcc-bugs@gcc.gnu.org/msg332020.html |
18:36.41 | y_morin | Oh, yes, it is... |
18:36.45 | y_morin | :-) |
18:37.11 | y_morin | GTG. BBL... |
18:44.01 | bhundven | yup |
20:29.19 | y_morin | bhundven: back! And bak to this symver issue... Lemme see... |
20:31.09 | bhundven | ok |
20:31.23 | y_morin | Wher do you see this issue? In the java configure? |
20:31.29 | bhundven | yes |
20:31.41 | bhundven | so, I don't care if symvers is on or off, or gnu, or sun. |
20:31.54 | bhundven | but my point to gcc is if it is correct or not. |
20:31.57 | y_morin | Strange. Here (gcc-4.6.0), the java configure does not check for symvers. |
20:32.08 | y_morin | Yes, I understand. |
20:32.21 | y_morin | But I'm trying to pinpoint the issue as well. |
20:34.25 | bhundven | if |
20:34.30 | bhundven | symvers=gnu |
20:34.48 | bhundven | libjava only checks for symvers=yes, symvars=no, or symvars=<error> |
20:34.58 | bhundven | or not set,: --disable-symvars |
20:35.12 | bhundven | which is equivilent to no |
20:35.25 | y_morin | But that's new in 4.6.1. |
20:35.29 | bhundven | 4.6.0 |
20:35.33 | y_morin | 4.6.0 did not have that check. |
20:35.45 | bhundven | ok.... lemme double check |
20:36.19 | y_morin | In 4.6.0, the test is only in libstdc++ and libgomp |
20:37.09 | bhundven | ok. so yes |
20:37.11 | bhundven | 4.6.1 |
20:37.15 | y_morin | In 4.6.1, it's also in libssp, libjava, libquadmath, libfortran. |
20:37.20 | bhundven | but, |
20:37.27 | bhundven | for libjava, it only checks for yes or no |
20:37.34 | y_morin | And now the real question: why do we set it? |
20:37.39 | bhundven | exactly |
20:37.50 | bhundven | for you, the question is why |
20:37.56 | bhundven | for gcc, the question is consistency |
20:38.14 | bhundven | if 'sh' is the only platform that really needs it |
20:38.23 | y_morin | I think it is a 'leftover' from the initial import from the original crosstool. Lemme check... |
20:38.30 | bhundven | then, I should be able to build with --enable-symvers=gnu and build libjava |
20:38.50 | bhundven | but, if that is no longer true... |
20:39.50 | bhundven | I think I'm looking at this like a logic problem. You can set it but it doesn't work. You might need it, but not sure. We set it, but not just to yes, but to gnu. |
20:40.09 | y_morin | bhundven: Yes, it's a leftover from the original code. |
20:40.18 | y_morin | And it seems it was needed for sh4 |
20:40.30 | bhundven | might not be anymore? |
20:40.56 | y_morin | From the original code: |
20:40.58 | y_morin | # --enable-symvers=gnu really only needed for sh4 to work around a detection problem |
20:40.59 | y_morin | <PROTECTED> |
20:41.03 | bhundven | yup, I read that |
20:41.09 | bhundven | I also noticed it on kegel's site |
20:41.30 | bhundven | I also noticed it on the blackfin build |
20:41.52 | y_morin | I'm no super-H expert by a large margin |
20:42.16 | bhundven | me neither |
20:42.39 | y_morin | OK, so what now? Make it a choice option: auto, yes, no, gnu, others... |
20:43.01 | bhundven | well |
20:43.09 | bhundven | from the bug, he asked why I am setting it manually |
20:43.24 | bhundven | so, maybe it should not be an option |
20:43.25 | bhundven | but I am waiting for a response |
20:43.30 | y_morin | From me? |
20:43.34 | bhundven | from gcc bug |
20:43.39 | y_morin | OK. |
20:45.03 | y_morin | And yes, the question is two-fold: consistency in gcc on one side, and reason for using it in crosstool-NG. |
20:45.13 | bhundven | exactly |
20:45.56 | y_morin | I think we have the answer to the second part: legacy code for a time where sh4 needed it, and it did not break other archs and other components. |
20:46.42 | bhundven | then we wait for ro's response |
20:46.55 | y_morin | Yep. |
20:46.59 | bhundven | I'm writing a re-write to his questions. |
20:47.06 | y_morin | goes get a coffee! ;-) |
20:47.09 | bhundven | I don't think my answers were good enough |
20:47.19 | y_morin | I'll follow the thread. Can you CC me, please? |
20:47.23 | bhundven | sure |
20:47.27 | y_morin | Thx. |
20:48.43 | y_morin | bhundven: use my alternate email: yann.morin.1998@free.fr for this thread, please. |
20:49.00 | bhundven | ok |
20:49.29 | y_morin | Do you have a pointer to the blackfin build scripts? |
20:49.48 | bhundven | I'll see if I can find it again |
20:50.52 | *** join/#crosstool-ng kos_tom (~thomas@2a01:e35:2f7b:7480:6aa3:c4ff:fe95:c756) |
20:51.07 | bhundven | oops |
20:51.09 | bhundven | I read it wrong |
20:51.14 | bhundven | they are disabling it |
20:51.21 | bhundven | so we are the only ones |
20:51.28 | bhundven | and maybe sh4 |
20:51.37 | bhundven | I don't see it in buildroot |
20:51.42 | y_morin | Let's look at buildroot |
20:51.43 | y_morin | Ok. |
20:53.17 | y_morin | Nope, indeed, they are silent about this. |
20:53.48 | y_morin | bhundven: Then I'll remove it from the build, and see how the current sh4 sample behaves. |
20:53.56 | bhundven | ok |
20:55.41 | y_morin | likes it: ct-ng build.18 :-) |
20:55.51 | y_morin | Unfortunately, that does not speed up the network... |
20:56.41 | bhundven | GCC Bugzilla was unable to make any match at all for one or more of the names and/or email addresses you entered on the previous page. |
20:56.44 | bhundven | Please go back and try other names or email addresses. |
20:56.47 | bhundven | CC: |
20:56.49 | bhundven | yann.morin.1998@free.fr did not match anything |
20:56.56 | bhundven | I hate bugzilla |
20:57.16 | bhundven | is that the email address associated with your gcc bugzilla login? |
20:57.23 | y_morin | So one has to be already registered to be CCed? Sigh... |
20:57.28 | bhundven | yea |
20:57.35 | y_morin | Lemme check... |
20:58.36 | y_morin | No, I am not registered to that BZ. |
20:58.46 | y_morin | I'll do, give me a few secs... |
20:58.50 | bhundven | k |
20:59.58 | y_morin | OK, registered. |
21:03.01 | bhundven | sent my re-answer |
21:03.05 | bhundven | with you cc'd |
21:03.06 | y_morin | Yep. |
21:03.08 | y_morin | Reading... |
21:03.55 | bhundven | I'm sure I have other problems with that build, besides symvers |
21:03.57 | bhundven | :-) |
21:04.33 | y_morin | bhundven: and probably also in other places than libjava: the new compojnents that check symver do it the way libjava does, not the way the 'old' libs did. |
21:04.45 | bhundven | hmm |
21:04.51 | y_morin | So I guess even if libjava got fixed, then the same issue would arise on those other libs. |
21:05.08 | y_morin | To be confirmed, of course, just reading the code here... |
21:05.19 | bhundven | probably true. |
21:07.51 | y_morin | Quoting Rainer: "anyone who undertakes to build GCC himself needn't be afraid from Perl" |
21:07.54 | y_morin | Lol! :-) |
21:08.19 | bhundven | I'm still more afraid of perl |
21:08.36 | y_morin | Yeah. I guess he was being humourous! :-) |
21:08.41 | bhundven | :) |
21:08.45 | bhundven | I get it |
21:09.12 | bhundven | i lol'd |
21:09.40 | bhundven | needs speach-to-text... |
21:09.55 | y_morin | OK, sh4 sample confirmed working with no change. Removing enable-symver and re-testing... |
21:10.58 | bhundven | and none of the samples we have, have java (gcj) enabled |
21:11.37 | y_morin | Yes: i686-nptl-linux-gnu/crosstool.config:CT_CC_LANG_JAVA=y |
21:11.44 | bhundven | oh |
21:11.48 | bhundven | :) heh |
21:12.00 | y_morin | 11 of them actually do. |
21:12.05 | bhundven | oh |
21:12.29 | y_morin | grep -r LANG_JAVA=y samples.mk |
21:12.35 | bhundven | so if they all were updated to 4.6.1, they would all break... right now. |
21:12.41 | y_morin | Yes. |
21:13.21 | y_morin | And still this annoying non-error on the sh4 build: |
21:13.23 | y_morin | [ERROR] st-dummy.c:1:23: error: sys/types.h: No such file or directory |
21:13.25 | y_morin | [ERROR] st-dummy.c:4:19: error: stdio.h: No such file or directory |
21:13.26 | y_morin | [ERROR] st-dummy.c:5:18: error: time.h: No such file or directory |
21:13.28 | y_morin | [ERROR] st-dummy.c:6:20: error: signal.h: No such file or directory |
21:13.35 | y_morin | And the build continues... |
21:13.38 | y_morin | As usual... |
21:14.05 | bhundven | interesting |
21:14.23 | bhundven | does that only happen on static core? |
21:14.34 | y_morin | This non-error has been here for ages. And yes, only static core. |
21:14.57 | y_morin | With or without // build. |
21:15.15 | y_morin | On x86 and x86_64. And only for sh4. |
21:16.36 | y_morin | Oh, I'm using gcc-4.3.2, as the sample is configured for. I'm not upgrading now, I'm trying not to pile up issues! ;-) |
21:16.53 | bhundven | hmm |
21:17.25 | bhundven | probably still a valid test |
21:17.29 | y_morin | glibc build OK so far. Starting build of gcc... |
21:17.51 | y_morin | Yes, probably stil a valid test, as it will prove non-gression. |
21:18.01 | bhundven | yup |
21:19.06 | bhundven | ... finding what fixed it would be trivial ... lots of bisecting. |
21:19.20 | bhundven | probably not worth it |
21:19.24 | y_morin | OK. Build was entirely successful. |
21:19.37 | y_morin | I suggest that we trash that option now. |
21:19.43 | bhundven | agreed |
21:19.49 | bhundven | my gcc question still persists. |
21:19.57 | y_morin | It leaves the gcc inconsistency, but that's another issue. |
21:20.01 | y_morin | Right. |
21:20.02 | bhundven | yup |
21:20.34 | y_morin | One day, I'll be typing faster than you are, and our messages won't be inverted anymore! ;-) |
21:20.49 | bhundven | haha, and I'm eating a sandwich at the same time |
21:20.57 | bhundven | 8-) |
21:21.01 | y_morin | cries... |
21:24.26 | bhundven | sudo apt-get install typespeed |
21:25.13 | y_morin | Lol. My brother (coding next to me) wondered what happened! :-) |
21:26.02 | y_morin | Well, well, well. typespeed *is* an actual package! |
21:27.05 | bhundven | damn, and I still suck at typespeed |
21:27.46 | bhundven | Rank: Average |
21:27.46 | bhundven | Score: 326 |
21:27.46 | bhundven | WPM: 43 |
21:27.47 | bhundven | CPS: 3.594 |
21:27.47 | bhundven | Typo ratio: 10.0% |
21:27.49 | bhundven | Typorank: Monkey |
21:27.51 | bhundven | :( |
21:28.16 | bhundven | that was with unix commands |
21:29.03 | y_morin | typespeed is cheating! It uses the american spelling! But I asked for 'english' words! |
21:29.33 | y_morin | generalise <- correct spelling in english. Only americaners use a 'z'! :-p |
21:29.46 | bhundven | haha |
21:29.55 | y_morin | won't probably code much tonight! :-) |
21:30.17 | bhundven | you should have used the french words |
21:30.56 | bhundven | find a baseline for your typing, then try something harder to help you figure out what you need to do better. |
21:31.21 | bhundven | also takes his own advise ;) |
21:32.08 | y_morin | ipchains? Really, anyone stil using ipchains? |
21:33.23 | y_morin | You achieved: |
21:33.25 | y_morin | <PROTECTED> |
21:33.26 | y_morin | <PROTECTED> |
21:33.28 | y_morin | <PROTECTED> |
21:33.29 | y_morin | <PROTECTED> |
21:33.31 | y_morin | <PROTECTED> |
21:33.32 | y_morin | <PROTECTED> |
21:33.34 | y_morin | <PROTECTED> |
21:33.35 | y_morin | Hmm... :-( |
21:34.05 | y_morin | WTF with the formating... |
21:34.46 | y_morin | The issue is not really with typing speed, but with thought speed. And I do need training! :-) |
21:34.58 | bhundven | me too |
21:35.01 | y_morin | I almost know what to say, not how to say it... |
21:36.42 | y_morin | And typing french words on a qwerty kbd... :-/ |
21:36.57 | bhundven | oh, yea, you have to use a modifier key, huh? |
21:37.17 | y_morin | Yep, except that under that new fancy KDE, meta no longer works... :-( |
21:37.40 | bhundven | or isn't there a special fr keyboard where you don't need the modifier key? |
21:37.49 | bhundven | s/modifier/meta/ |
21:37.49 | ctngbot | bhundven meant: or isn't there a special fr keyboard where you don't need the meta key? |
21:37.56 | bhundven | whoa |
21:38.06 | bhundven | 1 bot at a time, please... |
21:38.08 | y_morin | There are french keyboards, but I prefer qwerty. |
21:38.33 | y_morin | bhundven: yes, indeed. I forgot to ask infobot to leave. |
21:38.56 | y_morin | (I don't manage it, it was there in the interim until ctngbot was operational). Sorry... |
21:39.05 | bhundven | np |
21:39.33 | bhundven | I also didn't know that they understood regex |
21:39.34 | y_morin | you get mail. |
21:40.10 | bhundven | looks good to me! |
21:40.36 | bhundven | (as we say at my work:) ship it! |
21:40.51 | bhundven | lol |
21:41.03 | y_morin | Thanks! Pushing... |
21:41.33 | y_morin | You may want to add acked/reviewed? |
21:42.48 | bhundven | emailed reviewed-by: |
21:46.34 | y_morin | Pushed! :-) |