Re: Possible targets for ad-hoc ports operations?
On Mon, Sep 26, 2005 at 10:58:47PM -0400, Brad Ely wrote:
The way bsd.port.mk is written makes it fairly easy to implement
the following for-{all,build,run}-depends target(s) to do ad-hoc
operations with the ports infrastructure. Just pass in a statement
to be executed within the context of the makefile target.
For example:
(cd to port dir)
make for-all-depends='make fetch || true'
make for-all-depends='make checksum'
make for-all-depends='make clean=packages'
make for-all-depends='make show=PKGNAME MULTI_PACKAGES FLAVORS'
make for-all-depends='make show=FULLPKGNAME MAINTAINER' | xargs -L2
make for-all-depends='make print-package-signature'
find `make for-all-depends=pwd` ! -path '*CVS*' -type f -mtime -10
make for-all-depends='printf %s\t%s\t%s\n $$dir $$multi $$flavor'
Any good?
Brad Ely
Not really. You can use the output of the *-dir-depends inside shell scripts
that would do more or less what you propose.
There's not any real need to add more infrastructure, this is what we
usually do.
e.g.,
make all-dir-depends|tsort|while read i; do cd /usr/ports; SUBDIR=$i make
show=PKGNAME; done
Re: opera 8.50
On Tue, Sep 27, 2005 at 12:18:24AM +0200, frantisek holop wrote: so how is $SUBJECT working for others? dumping core on me every once in a while... a ulimit issue? anybody seeing this? (i am staff login class) on some sites in particular, it crashed on me regularly before the update. with 8.50 it seems to crash on said sites more frequently. i don't really use it that much though, just to get by when websites i need to go to are 100% flash or make firefox gobble all of the cpu. another issue: a page with a lot of images: from a certain point in the page the images are not loaded. and they are not broken of course. i haven't seen this yet.
Re: NEW: print/latex-beamer
On Mon, Sep 26, 2005 at 09:22:22PM +0100, Laurence Tratt wrote: OK, I wasn't very good on this point - sorry about that. There are old versions in TeTeX, and they don't work that well for me personally [in fact, they don't appear to be installed properly and I couldn't even get simple examples to compile with what's in TeTeX.] This is nonsense. Beamer in teTeX-3 in OpenBSD 3.7 works fine. It works on simple and not so simple examples for me. Besides, it works on all examples from http://www.tug.org/tex-archive/macros/latex/contrib/beamer/examples/ except the Chinese text in example4 that expects CJK package which is apparently not a part of the standard teTeX distribution. I bet if I download CJK from CTAN that it would work. But I really have no use for CJK. Also it's not old at all. We have beamer-3.01. The latest is 3.05-beta. We have pgf version that seems older (0.65) than the new one (0.95). Doesn't affect me frankly. Most of the people use EPS, PNG, or JPG files anyway. xcolor hasn't been changed in more than a year. It's still at version 2.0 that is in our teTeX distribution. teTeX is also missing the fairly vital documentation and examples without which one is pretty much stuck. Is it so difficult to visit CTAN for docs and other stuff? http://www.tug.org/tex-archive/help/Catalogue/entries/beamer.html The ports I've posted are up to date. There are no incompatibilities since the ports install in texmf-local, thus preserving whatever was in textmf-dist. I believe this is useful to you personally if you prefer to have the latest and greatest features. I personally, and most of other people can live with the standard teTeX distribution. It's like OpenBSD. It doesn't have all the fancy stuff some other OSs have, yet I still use it. :-) This does perhaps raise a question of what we should do when there are semi-duplicates with teTeX... Especially with something like beamer has a much quicker release schedule than the somewhat monolithic teTeX. We shouldn't do anything. Keep teTeX as it is. Six months is not a long wait for the new package release. If you really need the latest and greatest stuff, your personalized packages, fonts, and what not, drop it in $HOME/texmf tree and enjoy. At one point beamer development will slow down (sooner than later) and your port would be deleted from the ports tree. I think it should remain your personal port. Unless of course you persuade us with a precise list of simple examples that do not work in the current teTeX setup, rather than just a vague claim that it didn't work for you. Because it works for me. Regards, Zvezdan Petkovic
Re: update: math/octave
On Sat, Sep 24, 2005 at 12:56:15AM +0200, steven mestdagh wrote: hmm yes, something like 12 out of 1200 tests fail. however, when issuing 'gmake check' in the test directory, only 1 test fails. the exp() function does not return the expected (by the test scripts anyway) result: value expected exp (Inf) NaN Inf exp (-Inf) NaN 0 any ideas? I'll try to figure out why those other tests fail. so, those other 11 tests are failing because HOME is set to something nonexistent. is there a good way to fix such things? these tests are still doing the same, but i got some inspiration from other ports to fix other things - added MODULES=gcc3 (octave only compiles with gcc3) - does anyone want to test such architectures? - fixed the shared library names, get rid of symlinks there - used GNU_ARCH within packing list - patched configure to not test for runtest (it is included in REGRESS_DEPENDS) Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm octave.tgz Description: application/tar-gz
Re: NEW : sane-backends-1.0.16
Also please @comment out all those lib/sane/libsane-*.so.1 files. They doesn't make sense, right? Alek -- - Mamo, naprawdę nie musisz być taka zaambarasowana. - Zaambarasowana? W jakich zapleśniałych notatkach jakiego profesora historii języka znalazłeś takie słowo? Zapewniam cię, że nigdy, nawet w najgorszych koszmarach nie przyszło mi na myśl, żeby być... zaambarasowaną. -- Orson Scott Card, Teatr Cieni
Re: NEW : sane-backends-1.0.16
Antoine Jacoutot [EMAIL PROTECTED] wrote: c) isn't gphoto FLAVOR SHARED_ONLY=Yes candidate? I mean, will it work on architectures without shared libs? I have not tried it, but why would'nt it ? gphoto FLAVOR adds two shared libraries, right? So how is it supposed to work on archs without shared libs? Alek -- Greeyaz. Powinnam pamiętać, że trzeba kombinować od dupy strony, żeby przewidzieć, co wymyślą bachory ze Szkoły Bojowej. -- Orson Scott Card, Cień Hegemona
Re: Nmap - Core Dumps on AMD64
On Tue 2005.09.27 at 01:32 +0200, [EMAIL PROTECTED] wrote: Dear Portmaintainer, nmap crashs (core dump) on AMD64. I used nmap this way: nmap -sV -P0 -oN report.txt 194.55.223.0 was this command run as a regular user or as root? (nmap acts differently depending on priv) can you also give me the output with --debug? I did not compiled it with debugging-informations but here is the GDB-Output: none of that really helped... OpenBSD 3.8 (installed using a snapshot, then I upgraded using the SRC from the CVS (-rOPENBSD_3_8)). It's an dual Opteron (64Bit-Mode) with bsd.mp. can you try with just a snapshot? (not compiled by you) did it also coredump with bsd (not bsd.mp)? i don't have any mp machines, so i can't tell what happens. did you use the nmap snapshot pre-compiled package, port or compiled yourself from source? Maybe I did something wrong (?) but I guess it's a bug. Btw: Is it planed to adapt the new nmap-Version to 3.8? you mean 3.93? it was just updated.
Re: Nmap - Core Dumps on AMD64
Appears it was fixed in 3.93, only reproducible with the old version.
new port: Spreadsheet::WriteExcel
A port of the Perl module Spreadsheet::WriteExcel is available from: http://sebastian.foriru.co.uk/~sams/OpenBSD/p5-Spreadsheet-WriteExcel.tgz please test/commit/mock as appropriate. it works on i386. Sam -- Loneliness: If you find yourself struggling with loneliness, you're not alone. Yet you are alone. So very alone
Re: NEW : sane-backends-1.0.16
Hi... Once again, a new revision of my sane-backends port. Changelog : - add missing entries in PFRAG.gphoto2 (thanks to A. Piotrowski) Please, test and comment... Thank you all. Antoine sane-backends.tar.gz Description: application/gzip
Re: opera 8.50
frantisek holop wrote... so how is $SUBJECT working for others? dumping core on me every once in a while... a ulimit issue? anybody seeing this? (i am staff login class) another issue: a page with a lot of images: from a certain point in the page the images are not loaded. and they are not broken of course. Seems to be pretty stable for me, I'm now using it as my main browser instead of firefox (mainly because of the flash plugin) Some pages it has problems with, I go to firefox and they work fine, but they are mainly sites Designed for Internet Explorer. -- dorqus
Re: Possible targets for ad-hoc ports operations?
On Tue, 27 Sep 2005 08:18:41 +0200
Marc Espie [EMAIL PROTECTED] wrote:
On Mon, Sep 26, 2005 at 10:58:47PM -0400, Brad Ely wrote:
The way bsd.port.mk is written makes it fairly easy to implement
the following for-{all,build,run}-depends target(s) to do ad-hoc
Any good?
Brad Ely
Not really. You can use the output of the *-dir-depends inside shell scripts
that would do more or less what you propose.
There's not any real need to add more infrastructure, this is what we
usually do.
e.g.,
make all-dir-depends|tsort|while read i; do cd /usr/ports; SUBDIR=$i make
show=PKGNAME; done
Yeah, I have a similar script (even posted it here a while ago).
The nice thing about having the target(s) is that you get the
dependent environment (PKGPATH, SUBPACKAGE, and FLAVOR) already set
up for you. Pulling that from the make side into the shell script
isn't quite as fast. Besides, I didn't think you liked shell
wrappers that much. So sometimes its Ok to treat the ports
infrastructure like a black box? :-P
--
Brad Ely
Seeking a reputable source for information on HTML coding .. please read (September 27, 2005)
Hello, Some of my clients are searching online for a reputable source for information on HTML coding/programming. My job is to find one place to work with. I'd like to discuss an arrangement with you. Please contact me at your earliest convenience. I will be in today (Tuesday) from 8:00 AM PST to 5:00 PM PST. You'll need to be at your computer when you call me. Please call when you have a few minutes and I will a demonstrate how you would benefit from what we do and b answer any and all questions you have. Best Regards, Richard Meyers Business Segment Analyst, Star Position Phone (US): 877.732.5128 ext 2005 Outside US: 858.217.4934 PS: It is my job to send my clients looking from the search engines to one place at the exact moment they are looking. We have over 17,400,000 people on our network, and growing. The users on our Network are the clients I am referring to.
[NEW]: new port submission for syspatch
Greetings, I've made available a port of the syspatch utility. SPMS/syspatch is 'yet another way to binary patch OpenBSD boxes'. Some more infos on syspatch/SPMS may be found There: http://sf.net/projects/syspatch . The current release is syspatch-0.5. The port for syspatch-0.5 may be found there: . CVS module 'ports' from http://sourceforge.net/cvs/?group_id=146473 . Else HTTP tarball http://metawire.org/~olem/syspatch/download/syspatch-port.tgz Port has be tested ok on 3.8/i386 but should be ok for other Arch as no arch-dependend things in it. Best Regards, -- Olivier LEMAIRE SIEMENS BUSINESS SERVICES S.A.S. / ORS-PMO-OCI 39-47 Bvd Ornano - 93527 Saint Denis Office: 01.49.22.84.69 / Cellular: 06.12.72.49.31
Re: qemu propolice problem?
On Tuesday 27 September 2005 00:44, Peter Valchev wrote: CVSROOT:/cvs Module name:ports Changes by: [EMAIL PROTECTED] 2005/09/26 14:47:56 Modified files: emulators/qemu : Makefile emulators/qemu/patches: patch-Makefile_target Log message: o add patch to permit osx x86 guest to run o fix Makefile.target to add -fno-stack-protector for op.o for all archs, noted, debugged, and tested by Antoine Jacoutot on macppc, thanks! What's the story with this? Adding -fno-stack-protector left and right is obviously not acceptable as a final solution, are you looking at the problem in depth? I believe using -fno-stack-protector is correct in this case. Note that this is in Makefile.target, which is used to compile 'opcodes' for the emulated processor. And propolice interferes with that (even if it works, it would slow the emulation by doing its checks for every 'instruction') The rest of the qemu is compiled with propolice. Can
Re: qemu propolice problem?
I believe using -fno-stack-protector is correct in this case. Note that this is in Makefile.target, which is used to compile 'opcodes' for the emulated processor. And propolice interferes with that (even if it works, it would slow the emulation by doing its checks for every 'instruction') The rest of the qemu is compiled with propolice. Propolice only instruments functions which contain what it thinks look like strings. (A variety of issues make it a bit more complicated than being able to say just strings).
