choose-tree automatic update
Hi, If I leave choose-tree open and create other sessions, how do I get the already open choose-tree to update? Did this used to work before Thomas changed the code for choose-tree? Thanks, Jason -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris
In the first diff it should be enough just to remove the #include line because compat.h already includes it, so I have done that. The lockf change has gone into the OpenBSD repo rather than portable so will be pulled through to git in the next sync. Applied the cfmakeraw change as-is. Many thanks for your work! On Sun, Apr 21, 2013 at 10:30:41AM +0200, Dagobert Michelsen wrote: Hi John, Am 20.04.2013 um 21:40 schrieb John Long codeb...@inbox.lv: On Fri, Apr 19, 2013 at 04:14:16PM +0200, Dagobert Michelsen wrote: I rewrote the patch to use contrib/ substitutions for cfmakeraw and use lockf: https://sourceforge.net/p/tmux/tmux-code/merge-requests/3/ Thank you. Is there some way to download all the changes so I can apply them locally or do I just need to cut and paste from the web pages? From my understanding it should be possible to just accept the pull request. However, I have now formatted patches for the three functional units: http://sourceforge.net/apps/trac/gar/browser/csw/mgar/pkg/tmux/trunk/files/0001-Include-of-paths.h-if-its-presence-has-been-detected.patch http://sourceforge.net/apps/trac/gar/browser/csw/mgar/pkg/tmux/trunk/files/0002-Use-lockf-instead-of-flock-as-it-is-POSIX-compliant.patch http://sourceforge.net/apps/trac/gar/browser/csw/mgar/pkg/tmux/trunk/files/0003-Provide-replacement-for-cfmakeraw.patch Just let me know if you have any issues with the patches. Best regards -- Dago -- You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process. - xkcd #896 -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris
On Sun, Apr 21, 2013 at 10:55:29PM +0200, Dagobert Michelsen wrote: Hi John, Am 21.04.2013 um 17:26 schrieb John Long codeb...@inbox.lv: On Sun, Apr 21, 2013 at 05:16:09PM +0200, Dagobert Michelsen wrote: Am 21.04.2013 um 12:48 schrieb John Long: All patches are against HEAD: https://sourceforge.net/p/tmux/tmux-code/ci/master/tree/ Ok, downloaded the tarball? I suggest using git when doing stuff with git :-) I don't know how to use git. Isn't using the link you posted and downloading the tarball using git? I thought that got me the version your patches apply to. Make sure to call autogen.sh before building after applying the patches. The 0003-Provide-replacement-for-cfmakeraw.patch patch still doesn't apply: error: patch failed: Makefile.am:231 error: Makefile.am: patch does not apply What should I be looking for to fix this? Works for me: dam@login [login]:/home/dam/mgar/pkg/tmux/trunk cd tmux-tmux-code-c24b58e2ee8691870736959deb252c225b205b4d dam@login [login]:/home/dam/mgar/pkg/tmux/trunk/tmux-tmux-code-c24b58e2ee8691870736959deb252c225b205b4d gpatch -p1 ../files/0003-Provide-replacement-for-cfmakeraw.patch patching file Makefile.am Hunk #1 succeeded at 233 with fuzz 2 (offset 2 lines). patching file compat.h patching file compat/cfmakeraw.c patching file configure.ac I didn't use gpatch. I used git apply. Your way works better ;-) The build gets a lot further then breaks with a nawk syntax error. if test xman = xmdoc; then \ cp tmux.1.in.tmux.1; \ else \ nawk -fmdoc2man.awk tmux.1.in tmux.1; \ fi nawk: syntax error at source line 1 context is . \ $Id$ 4 missing ]'s nawk: bailing out at source line 1 Thank you. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris
On Mon, Apr 22, 2013 at 09:45:29AM +0100, Nicholas Marriott wrote: In the first diff it should be enough just to remove the #include line because compat.h already includes it, so I have done that. The lockf change has gone into the OpenBSD repo rather than portable so will be pulled through to git in the next sync. How can I tell when this happens? I just downloaded a tarball from http://sourceforge.net/p/tmux/tmux-code/ci/master/tree/ and the build breaks on client.c because it can't find flock. Thanks. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris
Hi John, Am 22.04.2013 um 12:46 schrieb John Long codeb...@inbox.lv: On Sun, Apr 21, 2013 at 10:55:29PM +0200, Dagobert Michelsen wrote: Hi John, Am 21.04.2013 um 17:26 schrieb John Long codeb...@inbox.lv: On Sun, Apr 21, 2013 at 05:16:09PM +0200, Dagobert Michelsen wrote: Am 21.04.2013 um 12:48 schrieb John Long: All patches are against HEAD: https://sourceforge.net/p/tmux/tmux-code/ci/master/tree/ Ok, downloaded the tarball? I suggest using git when doing stuff with git :-) I don't know how to use git. Isn't using the link you posted and downloading the tarball using git? No, that is just the files from git. Using git is facilitating the command listed on the webpage git clone git://git.code.sf.net/p/tmux/tmux-code tmux-tmux-code or something. Learning git pays off, but has a steep learning curve. I suggest the books and documents from Scott Chacon. I thought that got me the version your patches apply to. Make sure to call autogen.sh before building after applying the patches. The 0003-Provide-replacement-for-cfmakeraw.patch patch still doesn't apply: error: patch failed: Makefile.am:231 error: Makefile.am: patch does not apply What should I be looking for to fix this? Works for me: dam@login [login]:/home/dam/mgar/pkg/tmux/trunk cd tmux-tmux-code-c24b58e2ee8691870736959deb252c225b205b4d dam@login [login]:/home/dam/mgar/pkg/tmux/trunk/tmux-tmux-code-c24b58e2ee8691870736959deb252c225b205b4d gpatch -p1 ../files/0003-Provide-replacement-for-cfmakeraw.patch patching file Makefile.am Hunk #1 succeeded at 233 with fuzz 2 (offset 2 lines). patching file compat.h patching file compat/cfmakeraw.c patching file configure.ac I didn't use gpatch. I used git apply. Your way works better ;-) You can't use git apply if you don't have a git repository, obviously ;-) The build gets a lot further then breaks with a nawk syntax error. if test xman = xmdoc; then \ cp tmux.1.in.tmux.1; \ else \ nawk -fmdoc2man.awk tmux.1.in tmux.1; \ fi nawk: syntax error at source line 1 context is . \ $Id$ 4 missing ]'s nawk: bailing out at source line 1 The script is for gawk, not nawk. Feel free to use e.g. the one from OpenCSW. Best regards -- Dago Thank you. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process. - xkcd #896 -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris
Hi Dagobert, On Mon, Apr 22, 2013 at 01:00:19PM +0200, Dagobert Michelsen wrote: No, that is just the files from git. Using git is facilitating the command listed on the webpage git clone git://git.code.sf.net/p/tmux/tmux-code tmux-tmux-code or something. Learning git pays off, but has a steep learning curve. I suggest the books and documents from Scott Chacon. Thanks but this is a lot more than anybody should have to do to get a working copy of tmux. I'm not ever going to be a Linux developer and learning their tools (or junking up my Solaris boxes with gnu userland just to be able to build non-portable apps) is not on my list. I built git at the request of one of the developers on this list and it was the biggest pain in the ass of any piece of software I've built on Solaris. The git developers obviously have no interest and possibly no awareness in anything but Linux. And they're hardly alone. Anyway, nothing happened. I set up a shell account and gave access but nothing was done. After 6 months I was pretty hyped to see your patches. nawk -fmdoc2man.awk tmux.1.in tmux.1; \ fi nawk: syntax error at source line 1 context is . \ $Id$ 4 missing ]'s nawk: bailing out at source line 1 The script is for gawk, not nawk. Feel free to use e.g. the one from OpenCSW. I'm sure you're no stranger to this since you put so much effort into packaging software for Solaris, but I'm always amazed how much non-portable code and how much All the World R Belong 2 Linux stuff is written. I've been trying to get tmux on Solaris for a long time and nobody has been able or interested enough to fix it. Thanks for all your work on Solaris packaging, fixes, etc. I don't use any of it because I don't want the infrastructure and I don't want to make Solaris' preexisting package hell any worse, but I still greatly appreciate your efforts because a lot of the good work you do trickles down (up?) to other projects and possibly wakes people up to the fact Linux is not the only POSIX-like OS and some of us still prefer Solaris, right to the bitter end. I'll try building gawk myself and if that's not enough I'll give up until anybody has interest in making tmux work with standard UNIX stuff without requiring a gnu userland. If I wanted gnu crap I'd be running Linux. Yeah I know tmux runs on OpenBSD, that's where I found out about it. But it's less of a pain in the ass there since tmux cares about OpenBSD as a primary platform. By the way have you tested your tmux builds on Solaris Intel? Older versions of tmux build and run fine on my SPARC boxes but on Intel there is a problem of various keys (especially backspace) going in the wrong direction and not behaving nicely. I don't know where this is coming from but after this discussion I guess I'm missing other pieces of gnu crap that are needed for tmux to work. /jl -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
The script is for gawk, not nawk. Feel free to use e.g. the one from OpenCSW. I built the latest copy of gawk and now tmux builds ok, but I am back to the same problem I reported last year, which is the backspace key goes forward. Do you have any idea what could be causing this? It's been like this for me on Solaris Intel. On SPARC this has not been a problem. Thank you. /jl -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris
The script isn't for awk, I think we just invoke it in a way nawk isn't pleased with. Try this please: You will need to run autogen.sh again. diff --git a/Makefile.am b/Makefile.am index c5369fd..c131940 100644 --- a/Makefile.am +++ b/Makefile.am @@ -242,7 +242,7 @@ tmux.1: tmux.1.in if test x@MANFORMAT@ = xmdoc; then \ cp tmux.1.in tmux.1; \ else \ - $(AWK) -fmdoc2man.awk tmux.1.in tmux.1; \ + $(AWK) -fmdoc2man.awk tmux.1.in tmux.1; \ fi # Update SF web site. On Mon, Apr 22, 2013 at 10:46:22AM +, John Long wrote: On Sun, Apr 21, 2013 at 10:55:29PM +0200, Dagobert Michelsen wrote: Hi John, Am 21.04.2013 um 17:26 schrieb John Long codeb...@inbox.lv: On Sun, Apr 21, 2013 at 05:16:09PM +0200, Dagobert Michelsen wrote: Am 21.04.2013 um 12:48 schrieb John Long: All patches are against HEAD: https://sourceforge.net/p/tmux/tmux-code/ci/master/tree/ Ok, downloaded the tarball? I suggest using git when doing stuff with git :-) I don't know how to use git. Isn't using the link you posted and downloading the tarball using git? I thought that got me the version your patches apply to. Make sure to call autogen.sh before building after applying the patches. The 0003-Provide-replacement-for-cfmakeraw.patch patch still doesn't apply: error: patch failed: Makefile.am:231 error: Makefile.am: patch does not apply What should I be looking for to fix this? Works for me: dam@login [login]:/home/dam/mgar/pkg/tmux/trunk cd tmux-tmux-code-c24b58e2ee8691870736959deb252c225b205b4d dam@login [login]:/home/dam/mgar/pkg/tmux/trunk/tmux-tmux-code-c24b58e2ee8691870736959deb252c225b205b4d gpatch -p1 ../files/0003-Provide-replacement-for-cfmakeraw.patch patching file Makefile.am Hunk #1 succeeded at 233 with fuzz 2 (offset 2 lines). patching file compat.h patching file compat/cfmakeraw.c patching file configure.ac I didn't use gpatch. I used git apply. Your way works better ;-) The build gets a lot further then breaks with a nawk syntax error. if test xman = xmdoc; then \ cp tmux.1.in.tmux.1; \ else \ nawk -fmdoc2man.awk tmux.1.in tmux.1; \ fi nawk: syntax error at source line 1 context is . \ $Id$ 4 missing ]'s nawk: bailing out at source line 1 Thank you. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
What was the mail subject of the previous thread? Can you send me stty -a output inside and outside tmux? On Mon, Apr 22, 2013 at 11:47:32AM +, John Long wrote: The script is for gawk, not nawk. Feel free to use e.g. the one from OpenCSW. I built the latest copy of gawk and now tmux builds ok, but I am back to the same problem I reported last year, which is the backspace key goes forward. Do you have any idea what could be causing this? It's been like this for me on Solaris Intel. On SPARC this has not been a problem. Thank you. /jl -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: choose-tree automatic update
Hi, I think this is because when in choose-tree nothing can be updated, yes? I thought I remember this working once before in the past? Can someone look in to this? Jason - Original Message - From: Jason Timrod jtim...@yahoo.com To: tmux-users@lists.sourceforge.net tmux-users@lists.sourceforge.net Cc: Sent: Monday, April 22, 2013 8:27 AM Subject: choose-tree automatic update Hi, If I leave choose-tree open and create other sessions, how do I get the already open choose-tree to update? Did this used to work before Thomas changed the code for choose-tree? Thanks, Jason -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: FAO Thomas - code more carefully!
Hi, I've not herd anything from NicM yet. Does this mean my points here are being addressed? Jason - Original Message - From: Jason Timrod jtim...@yahoo.com To: Thomas Adam tho...@xteddy.org Cc: tmux-users@lists.sourceforge.net tmux-users@lists.sourceforge.net Sent: Sunday, April 21, 2013 6:29 PM Subject: Re: FAO Thomas - code more carefully! Hi, Well, I think I would like to hear from NicM (or the person in charge). The choose-tree -u fix I think should not go in to tmux. The more code ive seen from you, the worse things get because theres been lots of fixes on top of your own code from yourself and others. It shows poor programming. Jason From: Thomas Adam tho...@xteddy.org To: Jason Timrod jtim...@yahoo.com Cc: tmux-users@lists.sourceforge.net tmux-users@lists.sourceforge.net Sent: Sunday, April 21, 2013 5:16 PM Subject: Re: FAO Thomas - code more carefully! Hello, On Sun, Apr 21, 2013 at 05:02:50AM -0700, Jason Timrod wrote: Hi, Ive enjoyed using the devel version of tmux for ages - but lately ive had problems with it: I have to say that I am very surprised by this email, and am even more puzzled as to why you feel that I am solely responsible for you perceived problems. Yes, I may well have contributed to the areas you're referring to but you have to be a little more specific of my own personal involvement before you start to single me out as being harmful to this project. Why is there changes to choose-tree to fix -u? Do we need this? I do not think we do - I like the way it works now. Well I think you're in the minority. This is a bug-fix because the placement of the selected item is consistent with how the old choose-{session,window} commands worked in the past; highlighting those entries which were the current item. That this has now been folded in to a tree applies just as well when expanding/collapsing entries -- and hence by extension, using '-u'. Why do we need to support hooks? These have lots of huge changes - and most of the fixes in the devel of tmux are from Thomas's previous efforts to fix his mistakes. I can't say I follow you here. Hooks would be useful in so many different scenarios when different commands are run in tmux. It's something people have been wanting to do for a while now, and I've finally devised a solution to do it. It might also mean reducing a number of internal commands in tmux over time. As for these previous efforts you allude to, can you please show me where in the past I have done something very bad for this project? The choose-mode had a rewrite earlier this year - why? Thomas - did you break it when choose-tree came about and NicM fixed it? No -- internal improvements to different API parts are always on-going by myself, NicM, and other people who send in patches. What NicM and I were doing there was to simplify the fact that certain callbacks could be generalised, and hence reduce the code churn a little between the different choose-* commands. The git history also got messed up - and there was a thread about this on the mailing list you did not reply to? Why? I don't need to reply to everything. And now we have -u for choose-tree to fix it from today - I am worried Thomas that you are causing harm to tmux and not good. Can you please qualify this with something tangible? I notice that there was a fix added for missing -V as well. how did this happen? I mis-read the merge conflict, and chose the wrong one. Someone on IRC noticed this, and I fixed it. Why is this work not being reviewed? It is not good enough that you are creating all this badness for tmux. Does no one check your work? It would make tmux all the better for it!! My apologies to you and to anyone else who might feel my contributions to tmux are not as good as they could be, but I don't feel as though I need to justify this in the slightest. I am equally as puzzled as to why you think there's no review process. There is, and that's one of the most pleasurable parts about this project -- the review! Contrary to your own observations, my code always undergoes review because I do not have a commit-bit to the OpenBSD repository to make such changes -- and perhaps more importantly tmux is not my project, I only help out where/when I can, and see fit to do so. Note as well that off-list. NicM and I are quite often banding about patches between one another for peer-review/opinions, etc. So I am often involved indirectly in features I myself did not write. That's actually quite a high privilege in my eyes, that I'm able to help out like that. It should show some level of maturity and trust in the fact that I can make a positive impact to this project; and that justifying this to you for your own misunderstandings about my work on this project, is rather insulting. My involvement in this project is quite minimal; all things considered, and to think otherwise that I might have caused this
Re: Integration of patches for Solaris
Hi Nicholas, Thank you very much. In the meantime after I got Dagobert's answer and had not yet received your post, I compiled and installed the latest gawk and the tmux build goes ok now. It still looks we're in somewhat of a flux regarding which of Dagobert's patches made it to head because without me patching the build breaks and with patching some of them apply, some don't but in the end it builds with no errors after putting gawk ahead of Solaris awk in the path. Unfortunately, I am having the same problem I have always had on Solaris Intel: the backspace key goes in the wrong direction. This is real odd because on Solaris SPARC it works ok. Do you or anyone else have any idea what could be causing this? I remember at one point you suggested me building against gnu curses rather than Solaris curses and I thought I tried that, but maybe I didn't do all that was needed. Can you please explain how I could compile and link against gnu curses rather than Solaris curses so I can test your earlier theory? Thanks alot. /jl On Mon, Apr 22, 2013 at 12:53:28PM +0100, Nicholas Marriott wrote: The script isn't for awk, I think we just invoke it in a way nawk isn't pleased with. Try this please: You will need to run autogen.sh again. diff --git a/Makefile.am b/Makefile.am index c5369fd..c131940 100644 --- a/Makefile.am +++ b/Makefile.am @@ -242,7 +242,7 @@ tmux.1: tmux.1.in if test x@MANFORMAT@ = xmdoc; then \ cp tmux.1.in tmux.1; \ else \ - $(AWK) -fmdoc2man.awk tmux.1.in tmux.1; \ + $(AWK) -fmdoc2man.awk tmux.1.in tmux.1; \ fi # Update SF web site. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel]
On Mon, Apr 22, 2013 at 12:54:59PM +0100, Nicholas Marriott wrote: What was the mail subject of the previous thread? Same as above, minus the [backspace key.. etc. Can you send me stty -a output inside and outside tmux? Yes, thank you here it is: This is the stty -a output from an xterm when initially connecting to the Solaris box: speed 38400 baud; rows = 24; columns = 80; ypixels = 316; xpixels = 484; csdata ? eucw 1:0:0:0, scrw 1:0:0:0 intr = ^c; quit = ^\; erase = ^?; kill = ^u; eof = ^d; eol = -^?; eol2 = undef; swtch = undef; start = ^q; stop = ^s; susp = ^z; dsusp = ^y; rprnt = ^r; flush = ^o; werase = ^w; lnext = ^v; parenb -parodd cs8 -cstopb -hupcl cread -clocal -loblk -crtscts -crtsxoff -parext -ignbrk brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr icrnl -iuclc ixon -ixany -ixoff -imaxbel isig icanon -xcase echo echoe echok -echonl -noflsh -tostop echoctl -echoprt echoke -defecho -flusho -pendin iexten opost -olcuc onlcr -ocrnl -onocr -onlret -ofill -ofdel tab3 This is the stty -a output inside the compiled (but not installed) copy of tmux: speed 9600 baud; rows = 23; columns = 80; ypixels = 0; xpixels = 0; csdata ? eucw 1:0:0:0, scrw 1:0:0:0 intr = ^c; quit = ^\; erase = ^?; kill = ^u; eof = ^d; eol = -^?; eol2 = undef; swtch = undef; start = ^q; stop = ^s; susp = ^z; dsusp = ^y; rprnt = ^r; flush = ^o; werase = ^w; lnext = ^v; -parenb -parodd cs8 -cstopb -hupcl cread -clocal -loblk -crtscts -crtsxoff -parext -ignbrk brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr icrnl -iuclc ixon -ixany -ixoff imaxbel isig icanon -xcase echo echoe echok -echonl -noflsh -tostop echoctl -echoprt echoke -defecho -flusho -pendin iexten opost -olcuc onlcr -ocrnl -onocr -onlret -ofill -ofdel tab3 -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
On Mon, Apr 22, 2013 at 12:54:59PM +0100, Nicholas Marriott wrote: What was the mail subject of the previous thread? Oh, sorry, I suppose you mean the thread from last year? Let me check and I'll post here again as soon as I find it. Thanks. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
On Mon, Apr 22, 2013 at 12:54:59PM +0100, Nicholas Marriott wrote: What was the mail subject of the previous thread? This was my first email to the list, with a bad subject line since I was posting past my bedtime: http://sourceforge.net/mailarchive/message.php?msg_id=29643966 Sometime after that I got an older version working on Solaris SPARC. My next plea for help was here since having tmux on Solaris Intel would be so wonderful: http://sourceforge.net/mailarchive/message.php?msg_id=30150869 Thanks again. /jl -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris
On Mon, Apr 22, 2013 at 12:00:24PM +, John Long wrote: Hi Nicholas, Thank you very much. In the meantime after I got Dagobert's answer and had not yet received your post, I compiled and installed the latest gawk and the tmux build goes ok now. It still looks we're in somewhat of a flux regarding which of Dagobert's patches made it to head because without me patching the build breaks and with patching some of them apply, some don't but in the end it builds with no errors after putting gawk ahead of Solaris awk in the path. Ok well I think this is right so I will apply it and if it is still busted I guess next time someone tries to build on Solaris they will complain. Unfortunately, I am having the same problem I have always had on Solaris Intel: the backspace key goes in the wrong direction. This is real odd because on Solaris SPARC it works ok. Do you or anyone else have any idea what could be causing this? I remember at one point you suggested me building against gnu curses rather than Solaris curses and I thought I tried that, but maybe I didn't do all that was needed. Can you please explain how I could compile and link against gnu curses rather than Solaris curses so I can test your earlier theory? Well, do you already compile and link against libevent manually? You just need to do the same but with ncurses. Build it into somewhere eg /opt/ncurses: # ./configure --prefix=/opt/ncurses # make # make install I don't think ncurses has any dependencies but if it does you will need to either install them or figure out how to tell ncurses to disable them. Then build tmux. I would try a static build so you don't need to faff about with LD_LIBRARY_PATH: $ CFLAGS=-I/opt/ncurses/include LDFLAGS=-L/opt/ncurses/lib ./configure --prefix=/opt/tmux --enable-static $ make $ make install If you already set CFLAGS and LDFLAGS for libevent you'll need to include those bits too (eg CFLAGS=-I/path/to/ncurses/include -I/path/to/libevent/include). If that doesn't work try removing --enable-static but you'll need to run tmux with LD_LIBRARY_PATH pointing to /opt/ncurses/lib. I haven't used Solaris for a while but tmux definitely worked on Solaris 8, 9 and 10 a couple of years ago and I have heard few other complaints until now. Thanks alot. /jl On Mon, Apr 22, 2013 at 12:53:28PM +0100, Nicholas Marriott wrote: The script isn't for awk, I think we just invoke it in a way nawk isn't pleased with. Try this please: You will need to run autogen.sh again. diff --git a/Makefile.am b/Makefile.am index c5369fd..c131940 100644 --- a/Makefile.am +++ b/Makefile.am @@ -242,7 +242,7 @@ tmux.1: tmux.1.in if test x@MANFORMAT@ = xmdoc; then \ cp tmux.1.in tmux.1; \ else \ - $(AWK) -fmdoc2man.awk tmux.1.in tmux.1; \ + $(AWK) -fmdoc2man.awk tmux.1.in tmux.1; \ fi # Update SF web site. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
The stty looks fine I think, but I wonder if you backspace key is sending ^H rather than ^?. I wonder if that is the case and your TERM has left or right set to ^H. Can you start tmux and then do tmux info and send me the output? On Mon, Apr 22, 2013 at 12:16:55PM +, John Long wrote: On Mon, Apr 22, 2013 at 12:54:59PM +0100, Nicholas Marriott wrote: What was the mail subject of the previous thread? This was my first email to the list, with a bad subject line since I was posting past my bedtime: http://sourceforge.net/mailarchive/message.php?msg_id=29643966 Sometime after that I got an older version working on Solaris SPARC. My next plea for help was here since having tmux on Solaris Intel would be so wonderful: http://sourceforge.net/mailarchive/message.php?msg_id=30150869 Thanks again. /jl -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris
Hi John, Am 22.04.2013 um 13:28 schrieb John Long codeb...@inbox.lv: On Mon, Apr 22, 2013 at 01:00:19PM +0200, Dagobert Michelsen wrote: No, that is just the files from git. Using git is facilitating the command listed on the webpage git clone git://git.code.sf.net/p/tmux/tmux-code tmux-tmux-code or something. Learning git pays off, but has a steep learning curve. I suggest the books and documents from Scott Chacon. Thanks but this is a lot more than anybody should have to do to get a working copy of tmux. I'm not ever going to be a Linux developer and learning their tools (or junking up my Solaris boxes with gnu userland just to be able to build non-portable apps) is not on my list. I built git at the request of one of the developers on this list and it was the biggest pain in the ass of any piece of software I've built on Solaris. The git developers obviously have no interest and possibly no awareness in anything but Linux. And they're hardly alone. Anyway, nothing happened. I set up a shell account and gave access but nothing was done. After 6 months I was pretty hyped to see your patches. Short answer: Just wait until all the patches have been integrated, released possibly in tmux 1.9 and I'll release tmux at OpenCSW so you can do pkgutil -i tmux in the near future without further compilation. Just as a reminder: we already did all the tedious porting of git, so pkgutil -i git is sufficient if you need it in the future. nawk -fmdoc2man.awk tmux.1.in tmux.1; \ fi nawk: syntax error at source line 1 context is . \ $Id$ 4 missing ]'s nawk: bailing out at source line 1 The script is for gawk, not nawk. Feel free to use e.g. the one from OpenCSW. I'm sure you're no stranger to this since you put so much effort into packaging software for Solaris, but I'm always amazed how much non-portable code and how much All the World R Belong 2 Linux stuff is written. I've been trying to get tmux on Solaris for a long time and nobody has been able or interested enough to fix it. Yes. Thanks for all your work on Solaris packaging, fixes, etc. I don't use any of it because I don't want the infrastructure and I don't want to make Solaris' preexisting package hell any worse, I really don't understand what you mean here... but I still greatly appreciate your efforts because a lot of the good work you do trickles down (up?) to other projects and possibly wakes people up to the fact Linux is not the only POSIX-like OS and some of us still prefer Solaris, right to the bitter end. I'll try building gawk myself and if that's not enough I'll give up until anybody has interest in making tmux work with standard UNIX stuff without requiring a gnu userland. GNU AWK is for generating the manpage only. It is not used to run tmux. By the way have you tested your tmux builds on Solaris Intel? Older versions of tmux build and run fine on my SPARC boxes but on Intel there is a problem of various keys (especially backspace) going in the wrong direction and not behaving nicely. I just tried, works like charm on both spark and i386. I don't know where this is coming from but after this discussion I guess I'm missing other pieces of gnu crap that are needed for tmux to work. You don't need any GNU stuff to *run* tmux, just libevent. Best regards -- Dago -- You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process. - xkcd #896 -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
On Mon, Apr 22, 2013 at 01:38:37PM +0100, Nicholas Marriott wrote: Hmm this looks fine too. If you do cat outside tmux and press backspace, do you get ^H or nothing printed? It does what I would expect, it erases the t in cat and leaves me with ca, no funny characters are printed. When you say the backspace goes forward, do you mean that the cursor moves right but the character is still deleted, or the cursor moves but the character isn't deleted, The cursor moves to the right and no characters are deleted. Really quite unnerving! ;-) or what? Does it make any difference if you are at the shell prompt or if you are typing after running cat? Not sure what you meant. In tmux when I press backspace at the shell prompt I don't see anything happening. If I type at least one character then backspace, then the cursor moves to the right one character. For example if I type the letter c and backspace I get %c ^ | cursor points here, two chars after the c. It would seem, when I press backspace repeatedly the cursor goes forward as many characters as were typed. It just seems to be going in the wrong direction, and not deleting anything. Very very weird. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Make tmux.conf relative to sysconfdir
Hi, I noticed that the location of tmux.conf is not relative to sysconfdir, but hardcoded to /etc. This patch makes the location configurable: https://sourceforge.net/u/dmichelsen/tmux/ci/5547c378f07ecb7b64e2c0f8a0fec345c75873c4/ The patch can also be downloaded here: https://sourceforge.net/apps/trac/gar/browser/csw/mgar/pkg/tmux/trunk/files/0001-Make-location-of-tmux.conf-relative-to-sysconfdir.patch?rev=20839 It would be nice if it could be applied to HEAD. Best regards -- Dago -- You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process. - xkcd #896 -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris
On Mon, Apr 22, 2013 at 02:48:04PM +0200, Dagobert Michelsen wrote: Thanks for all your work on Solaris packaging, fixes, etc. I don't use any of it because I don't want the infrastructure and I don't want to make Solaris' preexisting package hell any worse, I really don't understand what you mean here... I mean for some (all?) of the csw stuff we first have to install your package tools and some prereq (base) packages and then add /opt/csw to the path, correct? I'm saying for me /usr/bin, /usr/local/bin, /usr/sfw/bin, /opt/firefox, /opt/solstudio, /opt/sfw etc. is enough path hell! Why doesn't everyone use /usr/local and be done with it? This has nothing to do with you, I just prefer not to make things any worse given Solaris path hell. Not that another path would really matter, it's just something I prefer not to participate in. And I guess this discussion should be offline, since the tmux people are probably mostly not interested. My main point is thanks for the work you do, I believe most Solaris users benefit from it in numerous ways whether they use your csw project directly or not. By the way have you tested your tmux builds on Solaris Intel? Older versions of tmux build and run fine on my SPARC boxes but on Intel there is a problem of various keys (especially backspace) going in the wrong direction and not behaving nicely. I just tried, works like charm on both spark and i386. That's really odd. I wonder what could be causing this. You don't need any GNU stuff to *run* tmux, just libevent. That's good in this specific case but generally it's sort of academic. Needing gnu stuff to build something is just as bad as needing gnu stuff to run something, unless you have a dedicated build machine. Fortunately, Nicholas has a fix that should work with nawk and I'm sure I'll have an opportunity to test it, although I now have gawk installed as well. Thanks again. /jl -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
On Mon, Apr 22, 2013 at 12:48:47PM +, John Long wrote: On Mon, Apr 22, 2013 at 01:38:37PM +0100, Nicholas Marriott wrote: Hmm this looks fine too. If you do cat outside tmux and press backspace, do you get ^H or nothing printed? It does what I would expect, it erases the t in cat and leaves me with ca, no funny characters are printed. When you say the backspace goes forward, do you mean that the cursor moves right but the character is still deleted, or the cursor moves but the character isn't deleted, The cursor moves to the right and no characters are deleted. Really quite unnerving! ;-) or what? Does it make any difference if you are at the shell prompt or if you are typing after running cat? Not sure what you meant. In tmux when I press backspace at the shell prompt I don't see anything happening. If I type at least one character then backspace, then the cursor moves to the right one character. For example if I type the letter c and backspace I get %c ^ | cursor points here, two chars after the c. It would seem, when I press backspace repeatedly the cursor goes forward as many characters as were typed. It just seems to be going in the wrong direction, and not deleting anything. Very very weird. Ok, but is that behaviour just at the shell prompt, or if you run cat then do the same, ie you are typing in cat rather than at the shell, does the cursor also move wrongly? Eg do: $ cat abcdefbackspace Instead of at the shell prompt. I'm trying to work out if this is because tmux is getting the wrong thing from zsh, or getting the right thing from zsh but displaying the wrong thing for xterm. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
On Mon, Apr 22, 2013 at 12:48:47PM +, John Long wrote: On Mon, Apr 22, 2013 at 01:38:37PM +0100, Nicholas Marriott wrote: Hmm this looks fine too. If you do cat outside tmux and press backspace, do you get ^H or nothing printed? It does what I would expect, it erases the t in cat and leaves me with ca, no funny characters are printed. I guess I misunderstood you again. If I do cat (as opposed to typing cat) backspace doesn't print anything. or what? Does it make any difference if you are at the shell prompt or if you are typing after running cat? No, there is no visible difference whether outside tmux or doing cat from tmux. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Make tmux.conf relative to sysconfdir
Hi I think this might be better than generating tmux.h, although I haven't tested it. You get the idea though :-). diff --git a/Makefile.am b/Makefile.am index c131940..d5bad20 100644 --- a/Makefile.am +++ b/Makefile.am @@ -15,7 +15,7 @@ dist-hook: find $(distdir) -name .svn -type d|xargs rm -Rf # Preprocessor flags. -CPPFLAGS += @XOPEN_DEFINES@ +CPPFLAGS += @XOPEN_DEFINES@ -DTMUX_CONF=\$(sysconfdir)/tmux.conf\ # glibc as usual does things ass-backwards and hides useful things by default, # so everyone has to add this. diff --git a/server.c b/server.c index 4bfa918..bd28d51 100644 --- a/server.c +++ b/server.c @@ -170,13 +170,13 @@ server_start(int lockfd, char *lockfile) cfg_references = 1; ARRAY_INIT(cfg_causes); - if (access(SYSTEM_CFG, R_OK) == 0) { - if (load_cfg(SYSTEM_CFG, cfg_cmd_q, cause) == -1) { - xasprintf(cause, %s: %s, SYSTEM_CFG, cause); + if (access(TMUX_CONF, R_OK) == 0) { + if (load_cfg(TMUX_CONF, cfg_cmd_q, cause) == -1) { + xasprintf(cause, %s: %s, TMUX_CONF, cause); ARRAY_ADD(cfg_causes, cause); } } else if (errno != ENOENT) { - xasprintf(cause, %s: %s, SYSTEM_CFG, strerror(errno)); + xasprintf(cause, %s: %s, TMUX_CONF, strerror(errno)); ARRAY_ADD(cfg_causes, cause); } if (cfg_file != NULL) { diff --git a/tmux.c b/tmux.c index 2916bbb..606c574 100644 --- a/tmux.c +++ b/tmux.c @@ -363,7 +363,7 @@ main(int argc, char **argv) if (pw != NULL) home = pw-pw_dir; } - xasprintf(cfg_file, %s/%s, home, DEFAULT_CFG); + xasprintf(cfg_file, %s/.tmux.conf, home); if (access(cfg_file, R_OK) != 0 errno == ENOENT) { free(cfg_file); cfg_file = NULL; diff --git a/tmux.h b/tmux.h index f0b9edf..fc5561b 100644 --- a/tmux.h +++ b/tmux.h @@ -39,10 +39,6 @@ extern char*__progname; extern char **environ; -/* Default configuration files. */ -#define DEFAULT_CFG .tmux.conf -#define SYSTEM_CFG /etc/tmux.conf - /* Default prompt history length. */ #define PROMPT_HISTORY 100 On Mon, Apr 22, 2013 at 02:56:17PM +0200, Dagobert Michelsen wrote: Hi, I noticed that the location of tmux.conf is not relative to sysconfdir, but hardcoded to /etc. This patch makes the location configurable: https://sourceforge.net/u/dmichelsen/tmux/ci/5547c378f07ecb7b64e2c0f8a0fec345c75873c4/ The patch can also be downloaded here: https://sourceforge.net/apps/trac/gar/browser/csw/mgar/pkg/tmux/trunk/files/0001-Make-location-of-tmux.conf-relative-to-sysconfdir.patch?rev=20839 It would be nice if it could be applied to HEAD. Best regards -- Dago -- You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process. - xkcd #896 -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
On Mon, Apr 22, 2013 at 01:58:33PM +0100, Nicholas Marriott wrote: Ok, but is that behaviour just at the shell prompt, or if you run cat then do the same, ie you are typing in cat rather than at the shell, does the cursor also move wrongly? Eg do: $ cat abcdefbackspace Instead of at the shell prompt. Yes, I'm sorry, I misunderstood you earlier (face-palm). I tried this now with and without tmux and backspace works fine in both cases, going backwards and erasing the characters. I'm trying to work out if this is because tmux is getting the wrong thing from zsh, or getting the right thing from zsh but displaying the wrong thing for xterm. Wondering about this earlier, I switched to bash before invoking tmux and I had the same issue with backspace going forward and not erasing. Thank you. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
On Mon, Apr 22, 2013 at 01:03:53PM +, John Long wrote: On Mon, Apr 22, 2013 at 12:48:47PM +, John Long wrote: On Mon, Apr 22, 2013 at 01:38:37PM +0100, Nicholas Marriott wrote: Hmm this looks fine too. If you do cat outside tmux and press backspace, do you get ^H or nothing printed? It does what I would expect, it erases the t in cat and leaves me with ca, no funny characters are printed. I guess I misunderstood you again. If I do cat (as opposed to typing cat) backspace doesn't print anything. or what? Does it make any difference if you are at the shell prompt or if you are typing after running cat? No, there is no visible difference whether outside tmux or doing cat from tmux. This isn't clear. Are you saying the problem only occurs at the zsh prompt and not when you do it with cat? If I do this at my shell prompt, where X is the cursor: $ abcdefXbackspace I get: $ abcdeX If I do this instead: $ cat abcdefXbackspace I get $ cat abcdeX So for you, the first case is broken and the second is normal. If you do both inside tmux. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
On Mon, Apr 22, 2013 at 01:08:54PM +, John Long wrote: I'm trying to work out if this is because tmux is getting the wrong thing from zsh, or getting the right thing from zsh but displaying the wrong thing for xterm. Wondering about this earlier, I switched to bash before invoking tmux and I had the same issue with backspace going forward and not erasing. However: If I invoke bash *inside* tmux then backspace starts working correctly again! What's going on??? -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
tmux on Raspberry Pi
I was able to install tmux on a Raspberry Pi running the recommended Wheezy Debian distribution. However, I wasn't able to start it. When I enter tmux new -s foo, it just outputs [exited]. If I run the same command with sudo, it works. Maybe I need to change permissions on some files. Any idea which and why I need to do that? -- R. Mark Volkmann Object Computing, Inc. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
On Mon, Apr 22, 2013 at 02:15:49PM +0100, Nicholas Marriott wrote: Ok so the problem only occurs with zsh. If you reproduce the problem inside tmux and then press C-b r (where C-b is your tmux prefix key), does the cursor move to the right place? No, it doesn't move at all. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
Ok so zsh is probably sending something weird. Can you run script then reproduce then exit script and send me the typescript file, eg something like: $ script Script started, output file is typescript $ abcdefbackspace $ exit On Mon, Apr 22, 2013 at 01:25:42PM +, John Long wrote: On Mon, Apr 22, 2013 at 02:15:49PM +0100, Nicholas Marriott wrote: Ok so the problem only occurs with zsh. If you reproduce the problem inside tmux and then press C-b r (where C-b is your tmux prefix key), does the cursor move to the right place? No, it doesn't move at all. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
On Mon, Apr 22, 2013 at 02:11:31PM +0100, Nicholas Marriott wrote: This isn't clear. Are you saying the problem only occurs at the zsh prompt and not when you do it with cat? Yes. If I do this at my shell prompt, where X is the cursor: $ abcdefXbackspace I get: $ abcdeX If I do this instead: $ cat abcdefXbackspace I get $ cat abcdeX So for you, the first case is broken and the second is normal. If you do both inside tmux. Yes, correct. And only with zsh. In bash backspace works as expected. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: [PATCH] FIX window resizing when a grouped session's child dies
Applied, thanks, although I left out the comment. On Fri, Apr 19, 2013 at 01:25:27AM -0700, Daniel Ralston wrote: Hi all, I found and fixed a bug where window sizes weren't being recalculated in grouped sessions when a session's child dies, killing the window. I've included the repro steps in the commit message. Let me know if you'd like any changes. Daniel From d7c541ad37dbe14dc50a2878448f5de0b753286c Mon Sep 17 00:00:00 2001 From: Daniel Ralston wubbul...@gmail.com Date: Fri, 19 Apr 2013 01:13:07 -0700 Subject: [PATCH] [PATCH] FIX window resizing bug when a grouped session's child dies When the child of a grouped session dies, and the session's client switches to the same window as the client of another session it is grouped with, sizes should be recalculated the same as if the user had switched to that other window manually. Steps to Reproduce Bug: 1. In an xterm: $ tmux new -s TEST_SESSION 2. Make sure the tmux session has aggressive-resizing enabled: prefix-: aggressive resizing on 3. In another xterm: $ tmux new -t TEST_SESSION 4. Create a new window running a shell (or whatever program you want) in the second xterm, and switch to it so the two tmux clients are viewing different windows. 5. Resize the second xterm so it is smaller than the first. 6. Kill the program in the second xterm so that its window closes. This will cause it to switch to the same window as the first tmux session. 7. In the first xterm, even though a smaller tmux session is now viewing that window, the size has not been updated. --- server-fn.c | 6 ++ 1 file changed, 6 insertions(+) diff --git a/server-fn.c b/server-fn.c index 566925f..4db4eb5 100644 --- a/server-fn.c +++ b/server-fn.c @@ -283,6 +283,12 @@ server_kill_window(struct window *w) if (options_get_number(s-options, renumber-windows)) session_renumber_windows(s); } + +/* If this session is grouped with another session, this + * session's client is switching to the same window that the + * other session's client is focused on, and two two clients + * are different sizes, we need to recalculate sizes. */ +recalculate_sizes(); } int -- 1.8.2.1 -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
On Mon, Apr 22, 2013 at 02:31:13PM +0100, Nicholas Marriott wrote: Ok so zsh is probably sending something weird. Can you run script then reproduce then exit script and send me the typescript file, eg something like: $ script Script started, output file is typescript $ abcdefbackspace $ exit This is from xterm Script started on Mon Apr 22 13:42:39 2013 [m[m[m[J$ [Kaabcdef zsh: command not found: abcde [m[m[m[J$ [Keexit script done on Mon Apr 22 13:43:20 2013 This is from tmux Script started on Mon Apr 22 13:43:12 2013 $ abcdef zsh: command not found: abcde $ exit script done on Mon Apr 22 13:42:50 2013 -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
I'm sending this again as attachments since the characters may not be displayable in email. First attachment is in xterm, second in tmux. Not sure if attachments will go through to the mailing list though. If it doesn't work let me know and I'll send it to you directly. Script started on Mon Apr 22 13:42:39 2013 [m[m[m[J$ [Kaabcdef zsh: command not found: abcde [m[m[m[J$ [Keexit script done on Mon Apr 22 13:42:50 2013 Script started on Mon Apr 22 13:43:12 2013 $ abcdef zsh: command not found: abcde $ exit script done on Mon Apr 22 13:43:20 2013 -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
Hmm this is weird. What is TERM set to inside tmux? Do you have the infocmp command? If so please run infocmp inside tmux and send me the output. On Mon, Apr 22, 2013 at 01:55:18PM +, John Long wrote: I'm sending this again as attachments since the characters may not be displayable in email. First attachment is in xterm, second in tmux. Not sure if attachments will go through to the mailing list though. If it doesn't work let me know and I'll send it to you directly. Script started on Mon Apr 22 13:42:39 2013 [m[m[m[J$ [Kaabcdef zsh: command not found: abcde [m[m[m[J$ [Keexit script done on Mon Apr 22 13:42:50 2013 Script started on Mon Apr 22 13:43:12 2013 $ abcdef zsh: command not found: abcde $ exit script done on Mon Apr 22 13:43:20 2013 -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: Integration of patches for Solaris [backspace key goes forward on Solaris Intel
Hello John, Note, that I didn't read the whole backlog. From what I gathered, your zsh is misbehaving inside of tmux. I can guess what might be going on. Maybe I'm onto something or maybe not. I guess we'll only know if you try. ;) John Long wrote: [...] Yes, correct. And only with zsh. In bash backspace works as expected. Okay. Bash uses the readline library for input handling. Zsh does not. It has its own line editor implementation (zle: man zshzle). So, generally, if bash works that doesn't mean anything for zsh. Now several things could be going on here. My first and foremost guess is, that you're actually using vi-mode without realising it. You can find out by issuing this: bindkey -lL main If that returns bindkey -A viins main you're in vi-mode. Zsh chooses the default input mode by looking at $EDITOR and $VISUAL. If the value looks like vi or vim or something like that it chooses vi-mode per default. If you really want emacs mode, you can force that by doing this: bindkey -e In any case, you can find out the current binding of a key or key sequence in the main keymap by doing this: bindkey '^?' bindkey '^h' If you do want vi-mode, we have to look at its default bindings. The default binding in vi-mode is ^h - vi-backward-delete-char. There is no default binding for ^? in the viins table (at least not according to the manual). So, if your terminal sends ^? for backspace, you'd have to do this: bindkey -M viins '^?' vi-backward-delete-char I hope some of this helps. Regards, Frank -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
Re: FAO Thomas - code more carefully!
Jason, Not to butt into your argument, but this passive-aggressive behavior is disrespectful, not only to Thomas, but tmux community as a whole. The project is open-sourced to encourage contributions, attacking people on the mailing list discourages them. I'm sure if Nicholas agreed with your view that Thomas is hurting the project, he would not accept his changes into the main branch. If you don't like the recent changes, you can always fork your own version. The point of devel version is for testing new features, on EVERY project, not just tmux. If you want stable code, stick to the stable version. Thomas doesn't get paid for contributing, and he handled your attack a lot more tactfully than most others would. I personally look forward to the hooks support since that would allow much better integration of scripts with tmux core, making tmux more customizable. If you really feel that his recent contributions were of low quality, you can send him a personal email rather than broadcasting it to the entire mailing list. On Mon, Apr 22, 2013 at 7:59 AM, Jason Timrod jtim...@yahoo.com wrote: Hi, I've not herd anything from NicM yet. Does this mean my points here are being addressed? Jason - Original Message - From: Jason Timrod jtim...@yahoo.com To: Thomas Adam tho...@xteddy.org Cc: tmux-users@lists.sourceforge.net tmux-users@lists.sourceforge.net Sent: Sunday, April 21, 2013 6:29 PM Subject: Re: FAO Thomas - code more carefully! Hi, Well, I think I would like to hear from NicM (or the person in charge). The choose-tree -u fix I think should not go in to tmux. The more code ive seen from you, the worse things get because theres been lots of fixes on top of your own code from yourself and others. It shows poor programming. Jason From: Thomas Adam tho...@xteddy.org To: Jason Timrod jtim...@yahoo.com Cc: tmux-users@lists.sourceforge.net tmux-users@lists.sourceforge.net Sent: Sunday, April 21, 2013 5:16 PM Subject: Re: FAO Thomas - code more carefully! Hello, On Sun, Apr 21, 2013 at 05:02:50AM -0700, Jason Timrod wrote: Hi, Ive enjoyed using the devel version of tmux for ages - but lately ive had problems with it: I have to say that I am very surprised by this email, and am even more puzzled as to why you feel that I am solely responsible for you perceived problems. Yes, I may well have contributed to the areas you're referring to but you have to be a little more specific of my own personal involvement before you start to single me out as being harmful to this project. Why is there changes to choose-tree to fix -u? Do we need this? I do not think we do - I like the way it works now. Well I think you're in the minority. This is a bug-fix because the placement of the selected item is consistent with how the old choose-{session,window} commands worked in the past; highlighting those entries which were the current item. That this has now been folded in to a tree applies just as well when expanding/collapsing entries -- and hence by extension, using '-u'. Why do we need to support hooks? These have lots of huge changes - and most of the fixes in the devel of tmux are from Thomas's previous efforts to fix his mistakes. I can't say I follow you here. Hooks would be useful in so many different scenarios when different commands are run in tmux. It's something people have been wanting to do for a while now, and I've finally devised a solution to do it. It might also mean reducing a number of internal commands in tmux over time. As for these previous efforts you allude to, can you please show me where in the past I have done something very bad for this project? The choose-mode had a rewrite earlier this year - why? Thomas - did you break it when choose-tree came about and NicM fixed it? No -- internal improvements to different API parts are always on-going by myself, NicM, and other people who send in patches. What NicM and I were doing there was to simplify the fact that certain callbacks could be generalised, and hence reduce the code churn a little between the different choose-* commands. The git history also got messed up - and there was a thread about this on the mailing list you did not reply to? Why? I don't need to reply to everything. And now we have -u for choose-tree to fix it from today - I am worried Thomas that you are causing harm to tmux and not good. Can you please qualify this with something tangible? I notice that there was a fix added for missing -V as well. how did this happen? I mis-read the merge conflict, and chose the wrong one. Someone on IRC noticed this, and I fixed it. Why is this work not being reviewed? It is not good enough that you are creating all this badness for tmux. Does no one check your work? It would make tmux all the better for it!! My apologies to you and to anyone else who might feel my contributions to
Re: FAO Thomas - code more carefully!
Hi, Do you speak for tmux aswell? I dont feel i have disrespected anyone - i have pointed out weaknesses which need looking in to by NicM and no one has yet made this point. i am making that point. i am waiting for a plan of action. Jason From: Alexander Tsepkov atsep...@gmail.com To: Jason Timrod jtim...@yahoo.com Cc: Thomas Adam tho...@xteddy.org; tmux-users@lists.sourceforge.net tmux-users@lists.sourceforge.net Sent: Monday, April 22, 2013 9:47 PM Subject: Re: FAO Thomas - code more carefully! Jason, Not to butt into your argument, but this passive-aggressive behavior is disrespectful, not only to Thomas, but tmux community as a whole. The project is open-sourced to encourage contributions, attacking people on the mailing list discourages them. I'm sure if Nicholas agreed with your view that Thomas is hurting the project, he would not accept his changes into the main branch. If you don't like the recent changes, you can always fork your own version. The point of devel version is for testing new features, on EVERY project, not just tmux. If you want stable code, stick to the stable version. Thomas doesn't get paid for contributing, and he handled your attack a lot more tactfully than most others would. I personally look forward to the hooks support since that would allow much better integration of scripts with tmux core, making tmux more customizable. If you really feel that his recent contributions were of low quality, you can send him a personal email rather than broadcasting it to the entire mailing list. On Mon, Apr 22, 2013 at 7:59 AM, Jason Timrod jtim...@yahoo.com wrote: Hi, I've not herd anything from NicM yet. Does this mean my points here are being addressed? Jason - Original Message - From: Jason Timrod jtim...@yahoo.com To: Thomas Adam tho...@xteddy.org Cc: tmux-users@lists.sourceforge.net tmux-users@lists.sourceforge.net Sent: Sunday, April 21, 2013 6:29 PM Subject: Re: FAO Thomas - code more carefully! Hi, Well, I think I would like to hear from NicM (or the person in charge). The choose-tree -u fix I think should not go in to tmux. The more code ive seen from you, the worse things get because theres been lots of fixes on top of your own code from yourself and others. It shows poor programming. Jason From: Thomas Adam tho...@xteddy.org To: Jason Timrod jtim...@yahoo.com Cc: tmux-users@lists.sourceforge.net tmux-users@lists.sourceforge.net Sent: Sunday, April 21, 2013 5:16 PM Subject: Re: FAO Thomas - code more carefully! Hello, On Sun, Apr 21, 2013 at 05:02:50AM -0700, Jason Timrod wrote: Hi, Ive enjoyed using the devel version of tmux for ages - but lately ive had problems with it: I have to say that I am very surprised by this email, and am even more puzzled as to why you feel that I am solely responsible for you perceived problems. Yes, I may well have contributed to the areas you're referring to but you have to be a little more specific of my own personal involvement before you start to single me out as being harmful to this project. Why is there changes to choose-tree to fix -u? Do we need this? I do not think we do - I like the way it works now. Well I think you're in the minority. This is a bug-fix because the placement of the selected item is consistent with how the old choose-{session,window} commands worked in the past; highlighting those entries which were the current item. That this has now been folded in to a tree applies just as well when expanding/collapsing entries -- and hence by extension, using '-u'. Why do we need to support hooks? These have lots of huge changes - and most of the fixes in the devel of tmux are from Thomas's previous efforts to fix his mistakes. I can't say I follow you here. Hooks would be useful in so many different scenarios when different commands are run in tmux. It's something people have been wanting to do for a while now, and I've finally devised a solution to do it. It might also mean reducing a number of internal commands in tmux over time. As for these previous efforts you allude to, can you please show me where in the past I have done something very bad for this project? The choose-mode had a rewrite earlier this year - why? Thomas - did you break it when choose-tree came about and NicM fixed it? No -- internal improvements to different API parts are always on-going by myself, NicM, and other people who send in patches. What NicM and I were doing there was to simplify the fact that certain callbacks could be generalised, and hence reduce the code churn a little between the different choose-* commands. The git history also got messed up - and there was a thread about this on the mailing list you did not reply to? Why? I don't need to reply to everything. And now we have -u for choose-tree to fix it from today - I am worried Thomas that you are causing harm to tmux and not good.
Re: Make tmux.conf relative to sysconfdir
Ok makes sense, please try this. It is important that the default (without --prefix or --sysconfdir) stays as /etc not /usr/local/etc so I had to add a hack to configure.ac to do that, --sysconfdir will still override. diff --git a/Makefile.am b/Makefile.am index c131940..726582a 100644 --- a/Makefile.am +++ b/Makefile.am @@ -15,7 +15,7 @@ dist-hook: find $(distdir) -name .svn -type d|xargs rm -Rf # Preprocessor flags. -CPPFLAGS += @XOPEN_DEFINES@ +CPPFLAGS += @XOPEN_DEFINES@ -DTMUX_CONF=\$(sysconfdir)/tmux.conf\ # glibc as usual does things ass-backwards and hides useful things by default, # so everyone has to add this. @@ -240,9 +240,10 @@ endif # Build tmux.1 in the right format. tmux.1: tmux.1.in if test x@MANFORMAT@ = xmdoc; then \ - cp tmux.1.in tmux.1; \ + sed -e s|@SYSCONFDIR@|$(sysconfdir)|g tmux.1.in tmux.1; \ else \ - $(AWK) -fmdoc2man.awk tmux.1.in tmux.1; \ + sed -e s|@SYSCONFDIR@|$(sysconfdir)|g tmux.1.in| \ + $(AWK) -fmdoc2man.awk tmux.1; \ fi # Update SF web site. diff --git a/configure.ac b/configure.ac index 80cf126..590b9db 100644 --- a/configure.ac +++ b/configure.ac @@ -18,6 +18,9 @@ AC_PROG_CC AM_PROG_CC_C_O AC_PROG_INSTALL +# Default tmux.conf goes in /etc not ${prefix}/etc. +test $sysconfdir = '${prefix}/etc' sysconfdir=/etc + # Check for various headers. Alternatives included from compat.h. AC_CHECK_HEADERS( [ \ diff --git a/server.c b/server.c index 4bfa918..bd28d51 100644 --- a/server.c +++ b/server.c @@ -170,13 +170,13 @@ server_start(int lockfd, char *lockfile) cfg_references = 1; ARRAY_INIT(cfg_causes); - if (access(SYSTEM_CFG, R_OK) == 0) { - if (load_cfg(SYSTEM_CFG, cfg_cmd_q, cause) == -1) { - xasprintf(cause, %s: %s, SYSTEM_CFG, cause); + if (access(TMUX_CONF, R_OK) == 0) { + if (load_cfg(TMUX_CONF, cfg_cmd_q, cause) == -1) { + xasprintf(cause, %s: %s, TMUX_CONF, cause); ARRAY_ADD(cfg_causes, cause); } } else if (errno != ENOENT) { - xasprintf(cause, %s: %s, SYSTEM_CFG, strerror(errno)); + xasprintf(cause, %s: %s, TMUX_CONF, strerror(errno)); ARRAY_ADD(cfg_causes, cause); } if (cfg_file != NULL) { diff --git a/tmux.1.in b/tmux.1.in index 98bf957..7f783b8 100644 --- a/tmux.1.in +++ b/tmux.1.in @@ -122,7 +122,7 @@ Specify an alternative configuration file. By default, .Nm loads the system configuration file from -.Pa /etc/tmux.conf , +.Pa @SYSCONFDIR@/tmux.conf , if present, then looks for a user configuration file at .Pa ~/.tmux.conf . .Pp @@ -3705,12 +3705,12 @@ was renamed to .Ar name . .El .Sh FILES -.Bl -tag -width /etc/tmux.confXXX -compact +.Bl -tag -width @SYSCONFDIR@/tmux.confXXX -compact .It Pa ~/.tmux.conf Default .Nm configuration file. -.It Pa /etc/tmux.conf +.It Pa @SYSCONFDIR@/tmux.conf System-wide configuration file. .El .Sh EXAMPLES diff --git a/tmux.c b/tmux.c index 2916bbb..606c574 100644 --- a/tmux.c +++ b/tmux.c @@ -363,7 +363,7 @@ main(int argc, char **argv) if (pw != NULL) home = pw-pw_dir; } - xasprintf(cfg_file, %s/%s, home, DEFAULT_CFG); + xasprintf(cfg_file, %s/.tmux.conf, home); if (access(cfg_file, R_OK) != 0 errno == ENOENT) { free(cfg_file); cfg_file = NULL; diff --git a/tmux.h b/tmux.h index f0b9edf..fc5561b 100644 --- a/tmux.h +++ b/tmux.h @@ -39,10 +39,6 @@ extern char*__progname; extern char **environ; -/* Default configuration files. */ -#define DEFAULT_CFG .tmux.conf -#define SYSTEM_CFG /etc/tmux.conf - /* Default prompt history length. */ #define PROMPT_HISTORY 100 On Mon, Apr 22, 2013 at 03:18:38PM +0200, Dagobert Michelsen wrote: Hi Nicholas, Am 22.04.2013 um 15:09 schrieb Nicholas Marriott nicholas.marri...@gmail.com: I think this might be better than generating tmux.h, although I haven't tested it. You get the idea though :-). Thanks for the quick turnaround, this looks good, but does not change the manpage which I would consider a good idea. Best regards -- Dago -- You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process. - xkcd #896 -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter
Re: tmux on Raspberry Pi
Thanks! That produces four files whose names begin with strace.out.. There are around 20 open() failures across those files that all say No such file or directory. I suspect all or most of these are expected errors for optional files. Here they are: /etc/ld.so.nohwcap /tmp/tmux-1000/default /etc/tmux.conf /home/pi/.terminfo /etc/terminfo/x/xterm /var/run/nscd/socket /proc/0/cwd /proc/0/cmdline /etc/ld/so.preload /usr/lib/arm-linux-gnueabihf/libfm/bash /usr/local/sbin/bash /usr/local/bin/bash /usr/sbin/bash /usr/bin/bash /sbin/bash /usr/lib/arm-linux-gnueabihf/libfm/reattach-to-user-namespace (related to my Mac configuration) several other directories looking for reattach-to-user-namespace /usr/share/local/en_GB.UTF-8/LC_MESSAGES/bash.mo several other directories looking for bash.mo Does any of this give you a clue? On Mon, Apr 22, 2013 at 8:23 AM, Nicholas Marriott nicholas.marri...@gmail.com wrote: Try strace -ff -ostrace.out tmux new then look for an open() that fails in the output files. Send me the files if you can't figure it out and I'll have a look. On Mon, Apr 22, 2013 at 08:18:01AM -0500, Mark Volkmann wrote: I was able to install tmux on a Raspberry Pi running the recommended Wheezy Debian distribution. However, I wasn't able to start it. When I enter tmux new -s foo, it just outputs [exited]. If I run the same command with sudo, it works. Maybe I need to change permissions on some files. Any idea which and why I need to do that? -- R. Mark Volkmann Object Computing, Inc. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter ___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users -- R. Mark Volkmann Object Computing, Inc. -- Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter___ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users