Re: Runtime error on Linux (undefined symbol)
On 10/04/09 01:41, Mun Johl wrote: Hi Tony, Please see my comments below. On Tue, Apr 07, 2009 at 09:19 PM PDT, Tony Mechelynck wrote: [...] TM http://vim.wikia.com/wiki/Setting_the_font_in_the_GUI Great write-up, thanks! It answered my main point of confusion which was why on my Solaris box using GTK I can specify the fonts with X-server names (e.g. 9x15) but on my Linux box using GTK2 I can't. If I read your tip correctly, you imply that GTK2 font specification differs from GTK, and I assume GTK uses the older X-server fonts. Yes, GTK1 uses what I called other-X11 'guifont' (also known as XLFD), which is incompatible with the GTK2 'guifont'. It's kind of unfortunate GTK2 doesn't support both styles, but such is life I suppose. The change happened before I came to Vim (and I was still mostly on Windows at first), so yes, I suppose in this case we should roll with the ball. Thanks very much for all the help, Tony. My pleasure. Best regards, Best regards, Tony. -- Every four seconds a woman has a baby. Our problem is to find this woman and stop her. --~--~-~--~~~---~--~~ You received this message from the vim_dev maillist. For more information, visit http://www.vim.org/maillist.php -~--~~~~--~~--~--~---
Re: Suggestion: Redefine \Uxxxxx in double-quoted strings
On Tue 07 Apr 2009 at 12:42:08 +0200, Bram Moolenaar wrote: This has two advantages: 1. It's backwards compatible. 2. Avoids accidentally typing the wrong number of hex digits. 3. Allows typing a hex digit next as a separate character. Eh, _three_ advantages. Nobody expects the spanish inquisition! -Olaf. -- ___ Olaf 'Rhialto' Seibert-- You author it, and I'll reader it. \X/ rhialto/at/xs4all.nl -- Cetero censeo authored delendum esse. --~--~-~--~~~---~--~~ You received this message from the vim_dev maillist. For more information, visit http://www.vim.org/maillist.php -~--~~~~--~~--~--~---
Re: Revising the zOS/EBCDIC port
Ralf Schandl wrote: after several years I had the chance to compile Vim on zOS Unix. That's nice to hear. I like supporting small groups of users. I already fixed some problems, but some are left: 1) test 11 fails, because gzip is not available on zOS Unix (at least not by default). Perhaps this test could be rewritten to use a program, that is guaranteed to be available. Xxd would be a candidate. See attached test11a.tar.gz Sounds like a good idea. 2) test24 fails, as it is ASCII dependent The ASCII dependencies doesn't seem to be the only problem here. Still investigating... 3) test 42 fails (as expected), as multi_byte is disabled Well, in fact the test result equals the test42.ok, but the zOS diff can't handle files containing nul-chars. Also the two files are identical doesn't mean that the test was successful, because the files are no valid multibyte files after transferring them to zOS, due to LATIN1-EBCDIC conversion. 4) test 55 fails, as the expected sort result of lists are ASCII dependent. test 57 fails, as the expected sort result is ASCII dependent 5) test 58 (spell check) ends in a SigSegV (spell.c:7474) Not investigated. As I need Vim for administrative tasks, spellcheck has no priority for me. I guess the spell checking only works for ASCII compatible encodings, thus not for EBCDIC. You can perhaps try if it works at all. If not it's better to disable, so that Vim doesn't crash when someone tries to use it. 6) QUOTESED in Makefile Is: QUOTESED = sed -e 's/[\\]/\\/g' -e 's/\\//' -e 's/\\;$$/;/' On zOS Unix all backslashes have to be escaped: QUOTESED = sed -e 's/[]//g' -e 's///' -e 's/;$$/;/' Any idea how to solve this in a portable way? You need to adjust the configure script for this. 7) eval.c: The table with the functions is sorted by ASCII value. So binary search doesn't work on an EBCDIC system. I fixed this by sorting the table in a function called from eval_init. Is this the right place? Should be. Another solution is to skip the binary search and use linear search. I will try to prepare a patch next week. Any ideas or comments? I'm not so worried about the tests. The more of the real problems you can fix the better. -- ** Hello and Welcome to the Psychiatric Hotline ** If you are obsessive-compulsive, please press 1 repeatedly. If you are co-dependent, please ask someone to press 2. If you have multiple personalities, please press 3, 4, 5 and 6. If you are paranoid-delusional, we know who you are and what you want - just stay on the line so we can trace the call. If you are schizophrenic, listen carefully and a little voice will tell you which number to press next. If you are manic-depressive, it doesn't matter which number you press - no one will answer. If you suffer from panic attacks, push every button you can find. If you are sane, please hold on - we have the rest of humanity on the other line and they desparately want to ask you a few questions. /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net \\\ ///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\ \\\download, build and distribute -- http://www.A-A-P.org/// \\\help me help AIDS victims -- http://ICCF-Holland.org/// --~--~-~--~~~---~--~~ You received this message from the vim_dev maillist. For more information, visit http://www.vim.org/maillist.php -~--~~~~--~~--~--~---
[PATCH] configure: $x_includes used even if not set
Hi, I am the maintainer for vim in MacPorts. The configure script triggered a linking error for _Xsetlocale for us on Mac OS X. Here is a patch which adds a conditional to make sure this test is not run if the x_includes variable is actually not set. Regards, Rainer --~--~-~--~~~---~--~~ You received this message from the vim_dev maillist. For more information, visit http://www.vim.org/maillist.php -~--~~~~--~~--~--~--- --- configure.in.old2009-04-11 00:05:39.0 +0200 +++ configure.in2009-04-11 00:06:09.0 +0200 @@ -2952,7 +2952,7 @@ if test $enable_multibyte = yes; then cflags_save=$CFLAGS ldflags_save=$LDFLAGS - if test -n $x_includes ; then + if test -n $x_includes test $x_includes != NONE; then CFLAGS=$CFLAGS -I$x_includes LDFLAGS=$X_LIBS $LDFLAGS -lX11 AC_MSG_CHECKING(whether X_LOCALE needed)
Re: [PATCH] configure: $x_includes used even if not set
On 11/04/09 00:14, Rainer Müller wrote: Hi, I am the maintainer for vim in MacPorts. The configure script triggered a linking error for _Xsetlocale for us on Mac OS X. Here is a patch which adds a conditional to make sure this test is not run if the x_includes variable is actually not set. Regards, Rainer Couldn't you run test just once? Maybe something more or less like if test -n $x_includes -a $x_includes != NONE Just my sense of aesthetics, I'm not on a Mac. Best regards, Tony. -- Those who express random thoughts to legislative committees are often surprised and appalled to find themselves the instigators of law. -- Mark B. Cohen --~--~-~--~~~---~--~~ You received this message from the vim_dev maillist. For more information, visit http://www.vim.org/maillist.php -~--~~~~--~~--~--~---
Re: [PATCH] configure: $x_includes used even if not set
On Fri, Apr 10, 2009 at 6:43 PM, Tony Mechelynck wrote: Couldn't you run test just once? Maybe something more or less like if test -n $x_includes -a $x_includes != NONE Just my sense of aesthetics, I'm not on a Mac. Usually, yes - but lore tells of shells where test isn't POSIX compatible, where -a and -o don't behave predictably but and || do... *shrug* ~Matt --~--~-~--~~~---~--~~ You received this message from the vim_dev maillist. For more information, visit http://www.vim.org/maillist.php -~--~~~~--~~--~--~---
Re: [PATCH] configure: $x_includes used even if not set
On 11/04/09 04:16, Matt Wozniski wrote: On Fri, Apr 10, 2009 at 6:43 PM, Tony Mechelynck wrote: Couldn't you run test just once? Maybe something more or less like if test -n $x_includes -a $x_includes != NONE Just my sense of aesthetics, I'm not on a Mac. Usually, yes - but lore tells of shells where test isn't POSIX compatible, where -a and -o don't behave predictably but and || do... *shrug* ~Matt Hm. What about the shells available on the Mac, and in particular on Mac OS X ? Best regards, Tony. -- Then here's to the City of Boston, The town of the cries and the groans. Where the Cabots can't see the Kabotschniks, And the Lowells won't speak to the Cohns. -- Franklin Pierce Adams --~--~-~--~~~---~--~~ You received this message from the vim_dev maillist. For more information, visit http://www.vim.org/maillist.php -~--~~~~--~~--~--~---
Re: [PATCH] configure: $x_includes used even if not set
On Fri, Apr 10, 2009 at 10:53 PM, Tony Mechelynck wrote: On 11/04/09 04:16, Matt Wozniski wrote: On Fri, Apr 10, 2009 at 6:43 PM, Tony Mechelynck wrote: Couldn't you run test just once? Maybe something more or less like if test -n $x_includes -a $x_includes != NONE Just my sense of aesthetics, I'm not on a Mac. Usually, yes - but lore tells of shells where test isn't POSIX compatible, where -a and -o don't behave predictably but and || do... Hm. What about the shells available on the Mac, and in particular on Mac OS X ? Changing the configure script would change things wherever ./configure is used. And, besides, it looks prettier is hardly a reason to accept behavior that might be broken somewhere. If we can agree that test ... -a ... and test ... test ... are functionally equivalent, but that the latter might work somewhere the former doesn't, it's pretty clear that we should use the latter. ~Matt --~--~-~--~~~---~--~~ You received this message from the vim_dev maillist. For more information, visit http://www.vim.org/maillist.php -~--~~~~--~~--~--~---
Re: [PATCH] configure: $x_includes used even if not set
On 11/04/09 05:02, Matt Wozniski wrote: On Fri, Apr 10, 2009 at 10:53 PM, Tony Mechelynck wrote: On 11/04/09 04:16, Matt Wozniski wrote: On Fri, Apr 10, 2009 at 6:43 PM, Tony Mechelynck wrote: Couldn't you run test just once? Maybe something more or less like if test -n $x_includes -a $x_includes != NONE Just my sense of aesthetics, I'm not on a Mac. Usually, yes - but lore tells of shells where test isn't POSIX compatible, where -a and -o don't behave predictably but and || do... Hm. What about the shells available on the Mac, and in particular on Mac OS X ? Changing the configure script would change things wherever ./configure is used. And, besides, it looks prettier is hardly a reason to accept behavior that might be broken somewhere. If we can agree that test ... -a ... and test ... test ... are functionally equivalent, but that the latter might work somewhere the former doesn't, it's pretty clear that we should use the latter. ~Matt OK. The former may run marginally faster due to the fact that the whole expression is computed in the same place (even on shells where test isn't a builtin), but that's hardly a reason to risk an easily avoidable malfunction. Best regards, Tony. -- The Preacher, the Politician, the Teacher, Were each of them once a kiddie. A child, indeed, is a wonderful creature. Do I want one? God Forbiddie! -- Ogden Nash --~--~-~--~~~---~--~~ You received this message from the vim_dev maillist. For more information, visit http://www.vim.org/maillist.php -~--~~~~--~~--~--~---
