[00:27:00] --- Simon Wilkinson has left: Lost connection [00:35:52] --- Claudio Bisegni has become available [00:57:51] --- dev-zero@jabber.org has left [01:35:54] --- dev-zero@jabber.org has become available [02:28:36] --- Simon Wilkinson has become available [02:58:48] --- Simon Wilkinson has left: Lost connection [03:49:26] --- Simon Wilkinson has become available [03:51:16] I wonder if, for 1.6, we should make the -L fileserver option the default. [04:50:50] --- Jeffrey Altman has become available [05:27:04] --- cclausen has left [05:33:29] --- Jeffrey Altman has left: Replaced by new connection [06:11:37] either -L, or, better, yet, the parameters discussed at the 08 workshop [06:13:06] --- Jeffrey Altman has become available [06:16:02] --- cclausen has become available [06:24:54] --- Russ has become available [06:29:53] hi folk, i have started today to make branch for OSX 10.5 preference pane. [06:36:18] Yeah, we should fix the standard parameter sets, rather than just making -L the default. [07:07:37] --- deason has become available [07:08:31] > hi folk, i have started today to make branch for OSX 10.5 preference pane yay! [07:08:53] submit pieces to gerrit as you get to that poin [07:09:05] point [07:09:16] ok [07:09:33] ideally, more small pushes sooner equals better. [07:09:36] i'm tryin to modify the Makefile to let xcode build the prefpane [07:09:46] then i must update the packager [07:10:34] xcodebuild called from Makefile was done elsewhere, istr. [07:10:35] yea i have read the Simon instruction, but my first push will be the total project:-) [07:10:44] yess i have done it [07:11:02] --- reuteras has left [07:11:03] --- stevenjenkins has left [07:11:11] --- haba has left [07:11:16] now i'm try fidn where is the code that copy all file into ${DEST} [07:11:26] i have never used make:-) [07:11:51] well, if you can get it building, i can write the dest and install and packaging targets [07:11:55] i have try also to compile afs on snow leopard but it didn't work [07:11:58] --- reuteras has become available [07:12:05] --- haba has become available [07:12:07] duplicate symbol errors? [07:12:20] ok it built [07:12:28] mmm i don't remeber [07:12:38] i can see if you whant [07:12:49] i ahve a virtual machine on my mac with snow server [07:12:53] me too [07:12:57] ok [07:13:01] if that's it, there's an open bug. [07:13:07] because i need to modify preference pane for snow [07:13:23] and see if work NSMenuExtra [07:13:30] ok [07:13:38] --- deason has left [07:13:50] i can push my base branch with only xcode compilation? [07:14:14] --- abo has left [07:14:15] --- Russ has left: Disconnected [07:14:32] --- deason has become available [07:14:42] --- stevenjenkins has become available [07:14:47] you mean without the glue to call xcodebuild, or with it, but no rules to do anythin after it's built? [07:15:09] --- abo has become available [07:15:13] only with the rule for compilation [07:15:22] I should push Toby's installer pane changes, because they've got make->xcodebuild hooks in. [07:15:57] sure, push, and simon, you too [07:16:02] ok, so i push only the modification for compile OpenAFS.prefPane [07:16:18] i will have to figure out dest/install. it will probably involve tar. or hey, macos ships rsync. i could use that [07:16:37] in fact, rsync will copy only files which changed, which is very make-compatible [07:16:43] yes, you can [07:17:24] Does our current style guide have anything to say about tabs? [07:17:26] remember, preference pane shuold be installed for all user [07:18:02] jhutz: It gives a crypto set of indent rules. [07:18:13] s/crypto/cryptic/. [07:18:36] prefs pane can't be installed for all users. many of them don't have macs. [07:18:50] However, the rule is that each level of indentation is 4 spaces. 8 spaces are collapsed into a tab. [07:19:26] Well, I assume it gives rules for emacs and/or indent. But, do we say anything about 0x09 in source files? Because I don't want to mention it in a comment I'm writing unless the style guide actually says don't do that. [07:19:27] i mean for all system user:-) [07:19:53] claudio, rdar://6885199 is the bug you probably want to file and request a dup on for the linker issue [07:20:09] jhutz: README.DEVEL is all we've got. [07:20:14] k [07:21:10] --- stevenjenkins has left [07:21:10] --- abo has left [07:21:14] But, if there's a patch with spaces inside a tab stop, or which uses diffferent indentation than 4 spaces / tab / tab + 4 spaces / 2 tabs, and so on, then I think that's worth a -1. [07:22:01] --- abo has become available [07:24:18] Yeah, I didn't look at it that closely. This was getting a -1 from me anyway for other reasons, on this pass. Hopefully the issues will get resolved, since I really do want to see this go in. [07:24:43] --- stevenjenkins has become available [07:25:13] --- manfred furuholmen has become available [07:26:35] So, the star next to a change means I want updates on it? [07:27:07] If you've reviewed, you'll be notified when that changes whatever happens. [07:27:32] The star can be used to mark a patch (that you've not reviewed) as something you're interested in. [07:29:02] --- deason has left [07:29:03] --- abo has left [07:29:30] --- deason has become available [07:29:43] --- abo has become available [07:30:03] --- deason has left [07:30:07] --- deason has become available [07:44:48] --- manfred furuholmen has left [07:46:18] --- manfred furuholmen has become available [08:00:51] --- reuteras has left [08:04:27] On the fileserver flags front, having just worked through with someone on IRC, I think -L is a huge improvement on what we ship as default at the moment. [08:04:51] I'm not sure if switching the -L numbers to be 'medium' makes sense, unless people can then suggest plausible new options for '-L', [08:06:44] --- abo has left [08:07:13] --- abo has become available [08:07:59] an info for the push operation, in the example there is write "HEAD:refs/for/" but the is the branch where i develop or the upstream branch? in my case i have branched on master and created the osx_preference_pane branch, where all work has been done [08:08:42] branch is the upstream branch. [08:08:47] So in your case would be 'master'. [08:08:58] ha ok so ipush onto master [08:09:01] --- deason has left [08:09:01] thank you simon [08:09:04] Yup. np. [08:09:55] --- deason has become available [08:11:44] --- stevenjenkins has left [08:13:16] i have seen that make clean from openasf root dir doesn't remove platform/DARWIN compiled content, if i call "make clean" from platform/DARWIN it work [08:14:57] If that's the case, it's broken. Either fix it :) , or open a bug in RT so it doesn't get forgotten about. [08:15:11] --- stevenjenkins has become available [08:15:14] ok thk [08:15:46] Preferably the former :-) [08:18:11] bug opened [08:31:22] --- Claudio Bisegni has left [08:40:56] --- deason has left [08:41:16] --- deason has become available [08:47:52] claudio, the make clean issue is now 198 in gerrit [09:01:00] You know, Tom does a much better job talking about the coherency model than I could. [09:03:09] --- manfred furuholmen has left [09:04:24] --- manfred furuholmen has become available [09:21:54] --- manfred furuholmen has left [09:24:00] --- manfred furuholmen has become available [09:34:08] Tom did an excellent job [09:36:41] --- manfred furuholmen has left [09:38:05] --- manfred furuholmen has become available [09:38:22] Anyone have any objection if I rewrite the src/packing/Darwin/Makefile to be generated by autoconf, and play better with the other children in the tree. In doing so, I can remove the APSL stuff, because I'll just write it from scratch. [09:38:56] The latter sounds like a win to me. [09:39:16] afssettings.m will still be APSL, sadly. [09:40:51] no objection from me [09:48:52] --- manfred furuholmen has left [09:53:28] --- manfred furuholmen has become available [09:59:15] yes, if you want to rewrite the darwin packaging makefile, please do,. i had a rewrite which went away when we did src/platform, and i never fixed it [10:04:58] --- phalenor has left [10:11:20] --- shadow@gmail.com/owl34D4DBE0 has left [10:16:17] --- shadow@gmail.com/owl34D4DBE0 has become available [10:36:37] Oh, lovely. We've got src/platform/DARWIN and src/packaging/MacOS [10:36:45] I wonder where the installer pane should live ... [10:37:54] Oh, and two copies of afssettings.m in the tree. Joyous. [10:39:36] src/packaging couldn't simply be e.g. DARWIN, LINUX. rehat, debian both linux [10:39:55] afssettings.m was supposed to move to (only) platform [10:40:16] I think I'll put the installer pane code in platform, so that it gets built on a make. [10:40:26] And reference it from buildpkg.sh [10:40:29] packaging was "only what you need to build packages, and not otherwise" platform was "tools which don't warrant a top directory but are platform-specific" [10:40:41] good idea [10:40:46] And I'll stick a patch in gerrit to remove the duplicates. [10:45:29] has anyone played with using branches to build for multiple platfroms from one git clone? [10:46:52] I don't see how that helps you. [10:47:34] Unless you're going to do a build, commit every file resulting from that build, and then switch to another branch. [10:47:58] I meant for personal building [10:48:20] as compared to needing to make clean every single time you wanted to use the same tree for different platforms [10:48:43] just do objdir builds. felix fixed them [10:48:59] oh, cool, I didn't notice that [10:49:01] or, failing that, use a shadow symlink tree [10:49:14] Did he really fix them? Do libafs_tree objdir builds also work? [10:49:48] 'git clone --mirror git://git.openafs.org/openafs.git' followed by 'git clone -s -l openafs.git git-foo1' 'git clone -s -l openafs.git git-foo2' etc will give you different trees [10:49:58] his fix broke nothing. i did not try libafs_tree [10:50:12] but objdir is probably preferable [10:50:47] well yeah, git-foo1 and git-foo2 would be entirely separate trees [10:51:03] which is what I'm doing currently, within VMs [10:51:34] i hear there's this distributed filesystem you can use [10:52:34] until you can ensure that I'll be able to not lose data when I have no net connection, that's great :) [10:53:12] well, if you don't use disconnected, you'll lose no data, but you won't be able to generate new data while offline [10:53:18] also "run your own local cell" [10:53:36] that's a good point [10:53:37] host runs cell. guests consume [11:00:09] > host runs cell. or you could have a guest whose job that is [11:04:00] --- tkeiser has become available [11:08:15] --- tkeiser has left [11:15:18] --- Claudio Bisegni has become available [11:16:18] derrick, ok for the clean issue thk [11:16:32] it's upstream now. git fetch will get it [11:16:40] k [11:21:54] derrick, remember that my Makefile.in lake for install part:-) [11:22:13] er, what? [11:22:49] the version of Makefile.in modify in mu commit doesn't implemente the install rule for preference pane [11:23:21] i have added only the compile rule in makefile [11:23:28] Derrick: change 199, what needs fixed in buildpkg? It seems to pick up afssettings from where src/platform puts it. [11:24:01] oh, i know. [11:24:05] one thing at a time tho [11:24:17] ok, np [11:24:25] you're right, i see from this build that the right thing seems to have happened [11:35:44] Do you want me to rebase 168? [11:35:58] without looking i'll assume yes [11:36:02] brb [11:39:48] --- Claudio Bisegni has left [12:27:41] and the regex patch asplode on "newer than 10.5" [12:29:08] That's dull. What have they broken on "newer" ? [12:30:09] nothing, apparently? configure did trt. hm. [12:31:14] configure:21514: checking for POSIX regex library configure:21525: result: yes but it's not defined. wtf? [12:35:14] let me check something. i suspect i see the answer [12:40:09] Found it? [12:40:50] stale checkout [12:40:57] regex.h in src/bucoord still there [12:41:17] Hmm. I would have thought it would get nuked. [12:41:39] it did, once i did the right git reset [12:41:48] Ah. Okay. [12:41:58] BTW: Have an AFS cellname prompter. Only a year or so delayed. [12:43:06] thanks [12:48:34] --- manfred furuholmen has left [13:00:38] One reason why we need more releases: The amount of time git cherry origin/openafs-stable-1_4_x origin/master takes to run! [13:06:10] --- deason has left [13:09:51] --- deason has become available [13:41:17] Why is the Return-Path on mail gerrit sends set to the person who submitted a patch, instead of to the gerrit or mailing list admin? There's nothing joe random can do with a bounce. [13:42:13] On which mail? [13:42:42] Mail that comes from the code review system, or mail that goes to openafs-cvs? [13:43:37] Mail that came from the code review system. [13:43:48] openafs-cvs is less important, because Mailman takes care of being sane [13:45:20] I suspect it's because gerrit just makes the message look like it's come from the user that submitted the patch. Gerrit actually does very little beyond set the From address, Java mail does most of the rest. [13:45:59] Okay, so why does 386fea8d have a >80 character subject? [13:47:00] Ah, no gap between the subject and the content. Tsk tsk. [14:00:32] With a little bit of scripting I now have a command which will tell you all the changes in a branch that haven't been cherry-picked to another, using both the patch ID, and the 'cherry-picked from' notation in the commit log. [14:05:56] ... and there are 1683 commits on head that aren't present in openafs-stable-1_4_x ... Ouch [14:05:58] So, you can tell us which patches on master haven't been pulled to 1.4.x? [14:06:10] Yes, all 1683 of them :) [14:06:53] Hm. Not all of those ever will be pulled over. It'd be nice to somehow be able to distinguish. [14:06:58] --- dev-zero@jabber.org has left [14:07:08] And yes, every time you add a feature I will ask for another. :-) [14:08:09] I really need to filter gerrit mail into its own mailbox. Or else change my preferences; every new change is too much. [14:09:54] It looks like the stable_1_4_x branch diverged from master in 2004, so that's 5 years worth of changes there. [14:10:52] --- deason has left [14:11:00] that is where the deltas come in. [14:11:20] --- deason has become available [14:11:25] --- mmeffie has become available [14:11:28] you should be able to take a delta name from master and determine if it was pulled (or an equivalent patch was applied) to 1.4 [14:14:50] --- Simon Wilkinson has left [14:14:50] --- Simon Wilkinson has become available [14:15:29] Ah, but I'm smarter than that :) [14:16:09] --- manfred furuholmen has become available [14:16:39] The git conversion tagged every pullup that was identifiable through having a corresponding delta name, with the magic 'cherry-picked from change $sha1', where $sha1 was the hash of the parent change on master. [14:17:56] So, you don't need deltas for this, either... [14:53:02] --- matt has become available [14:56:16] --- dev-zero@jabber.org has become available [15:13:15] --- mmeffie has left [15:27:20] --- deason has left [15:35:44] --- Russ has become available [16:12:21] --- deason has become available [16:12:26] --- manfred furuholmen has left [16:14:30] --- deason has left [16:42:12] --- cclausen has left [16:51:50] We should try and make sure that the information in http://www-01.ibm.com/support/docview.wss?rs=961&context=SSXMUG&dc=DA400&uid=swg27004254&loc=en_US&cs=UTF-8&lang=en&rss=ct961other doesn't go away. [16:57:16] that url doesn't work for me [16:58:12] How about http://tinyurl.com/lw76j4 [17:06:12] --- deason has become available [17:11:13] I can read that. I think we should strive to replicate it. [17:11:37] Put up a template in the wiki and encourage admins to help fill it in [17:57:29] --- matt has left [22:14:07] --- Russ has left: Disconnected [22:31:57] --- deason has left [22:37:56] --- reuteras has become available [23:02:04] --- dev-zero@jabber.org has left: Replaced by new connection [23:02:05] --- dev-zero@jabber.org has become available [23:35:49] --- dev-zero@jabber.org has left [23:47:54] --- Simon Wilkinson has left