Re: [Fink-devel] Re: fink/perlmod/Fink ChangeLog,1.518,1.518.2.1 Engine.pm,1.156,1.156.2.1 PkgVersion.pm,1.209,1.209.2.1 SourceItem.pm,1.1,1.1.2.1
On Sun, Jan 11, 2004 at 06:28:49PM -0500, Daniel Macks wrote: > On Sun, Jan 11, 2004 at 11:53:41PM +0100, Max Horn wrote: > > Am 11.01.2004 um 23:31 schrieb Ben Hines: > > >On Jan 11, 2004, at 2:22 PM, Daniel Macks wrote: > > > > > >>Switch to an object-oriented and consistent way of handling source > > >>tarballs. Seems to slow down indexing quite a bit and make fink.db up > > > > > >If there is anything we don't need it is to slow down indexing. I > > >don't care how much cleaner it makes the code, we need to index > > >constantly. It needs to be fast. > > > > I agree with Ben. Indexing is already now terribly slow. We spent a lot > > of effort to get it as fast as possible (using Storable, at least) - a > > noticeable slow down is not acceptable, IMNSHO. > > I also agree, which is why I put it in a branch:) > > It started out as some changes that didn't have a timing effect, but > the more I worked on it... Okay, I just checked the things I did in HEAD. For 'fink index': Benchmark: timing 10 iterations of now, was... now: 169 wallclock secs ( 0.00 usr 0.04 sys + 146.45 cusr 13.56 csys = 160.05 CPU) @ 250.00/s (n=10) was: 154 wallclock secs ( 0.00 usr 0.02 sys + 136.48 cusr 13.73 csys = 150.23 CPU) @ 500.00/s (n=10) Depending on how you want to spin those number, it's either only 1.5 seconds slower on my not terribly recent G4, or it's 10% slower:) I could probably bring that down a little but not much. Undo it all? dan -- Daniel Macks [EMAIL PROTECTED] http://www.netspace.org/~dmacks --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Re: fink/perlmod/Fink ChangeLog,1.518,1.518.2.1 Engine.pm,1.156,1.156.2.1 PkgVersion.pm,1.209,1.209.2.1 SourceItem.pm,1.1,1.1.2.1
On Sun, Jan 11, 2004 at 11:53:41PM +0100, Max Horn wrote:
> Am 11.01.2004 um 23:31 schrieb Ben Hines:
> >On Jan 11, 2004, at 2:22 PM, Daniel Macks wrote:
> >
> >>Switch to an object-oriented and consistent way of handling source
> >>tarballs. Seems to slow down indexing quite a bit and make fink.db up
> >
> >If there is anything we don't need it is to slow down indexing. I
> >don't care how much cleaner it makes the code, we need to index
> >constantly. It needs to be fast.
>
> I agree with Ben. Indexing is already now terribly slow. We spent a lot
> of effort to get it as fast as possible (using Storable, at least) - a
> noticeable slow down is not acceptable, IMNSHO.
I also agree, which is why I put it in a branch:)
It started out as some changes that didn't have a timing effect, but
the more I worked on it...
Suffice to say, it's not good for HEAD and likely never will be, but
it included some ideas I might try again later so I committed it.
Things like making more uniform fields and reducing the use of our
ever-changing collection of private $pkg->{_thing}.
dan
--
Daniel Macks
[EMAIL PROTECTED]
http://www.netspace.org/~dmacks
---
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
___
Fink-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Re: fink/perlmod/Fink ChangeLog,1.518,1.518.2.1 Engine.pm,1.156,1.156.2.1 PkgVersion.pm,1.209,1.209.2.1 SourceItem.pm,1.1,1.1.2.1
Am 11.01.2004 um 23:31 schrieb Ben Hines: If there is anything we don't need it is to slow down indexing. I don't care how much cleaner it makes the code, we need to index constantly. It needs to be fast. -Ben On Jan 11, 2004, at 2:22 PM, Daniel Macks wrote: Switch to an object-oriented and consistent way of handling source tarballs. Seems to slow down indexing quite a bit and make fink.db up I agree with Ben. Indexing is already now terribly slow. We spent a lot of effort to get it as fast as possible (using Storable, at least) - a noticeable slow down is not acceptable, IMNSHO. Bye, Max --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
[Fink-devel] Re: fink/perlmod/Fink ChangeLog,1.518,1.518.2.1 Engine.pm,1.156,1.156.2.1 PkgVersion.pm,1.209,1.209.2.1 SourceItem.pm,1.1,1.1.2.1
If there is anything we don't need it is to slow down indexing. I don't care how much cleaner it makes the code, we need to index constantly. It needs to be fast. -Ben On Jan 11, 2004, at 2:22 PM, Daniel Macks wrote: Switch to an object-oriented and consistent way of handling source tarballs. Seems to slow down indexing quite a bit and make fink.db up --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Heureka! (Was: Fink-0.18.0 doesn't compile)
This could be easily tested with a disk image, for example. Mount a read/write disk image, get info, choose 'ignore permissions' (looks like it is default) and try to install fink. It does indeed fail: This first test is designed to die, so please ignore the error message on the next line. # Looks like your test died before it could output anything. ./00compile.ok ./Base/initialize...ok ./Base/paramok ./Base/param_booleanok ./Command/cat...ok ./Command/chowname..ok 1/0# Failed test (./Command/chowname.t at line 27) # got: 'ben' ./Command/chowname..NOK 2# expected: 'nobody' ./Command/chowname..ok 4/0# Failed test (./Command/chowname.t at line 34) ./Command/chowname..NOK 5# got: 'unknown' # expected: 'nobody' ./Command/chowname..ok 7/0# Failed test (./Command/chowname.t at line 43) ./Command/chowname..NOK 8# got: 'ben' # expected: 'nobody' # Failed test (./Command/chowname.t at line 44) # got: 'unknown' # expected: 'nobody' ./Command/chowname..NOK 9# Looks like you failed 4 tests of 9. ./Command/chowname..dubious Test returned status 4 (wstat 1024, 0x400) DIED. FAILED tests 2, 5, 8-9 Failed 4/9 tests, 55.56% okay ./Command/commands..ok ./Command/exports...ok ./Command/failure...ok ./Command/touch.ok ./Config/exportsok ./Config/failureok ./Config/load_save..ok ./Config/optionsok ./Config/param..ok ./Config/verbosity_levelok ./FinkVersion/exports...ok ./FinkVersion/version...ok ./Mirror/exportsok ./PkgVersion/get_perl...ok Failed Test Stat Wstat Total Fail Failed List of Failed --- ./Command/chowname.t4 1024 94 44.44% 2 5 8-9 Failed 1/20 test scripts, 95.00% okay. 4/482 subtests failed, 99.17% okay. make: *** [test] Error 1 ### execution of make failed, exit code 2 compiling fink-0.18.0.cvs-20040111.2118 failed -Ben On Jan 10, 2004, at 6:20 AM, David R. Morrison wrote: We need our testing guru to come back and help us out here... --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] building The GIMP 2.0pre1
On Jan 9, 2004, at 10:16 AM, Martin Costabel wrote: The install script for freetype2 moves lib/libfreetype.la into a lib/freetype2/ directory, but that doesn't get installed from %i into %p so there appears to be a problem with that package. It gets installed into %p when you install the freetype2-dev splitoff. I can't comment on what is necessary for gimp, haven't looked at it. Nope... in 10.3 anyway, the .dylib symlink is still hidden away in /sw/lib/freetype2/. This is on purpose, because we want fink packages to use apple's freetype2 by default. Fink's freetype2 is 'hidden away' for packages that specifically need that version. ./sw/lib/freetype2/lib/libfreetype.dylib -> /sw/lib/libfreetype.6.dylib -Ben --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] fink-maintainer-gaim :: gaim >= 0.71 please
David R. Morrison wrote: Fink 0.6.2 was released on 17 November 2003, less than two months ago. You are right, my first message to fink-devel where I asked that qt3 in the bindist for 10.3 should be replaced by something that actually works on 10.3 dates from 6 November, not much more than 2 months ago (version was 0.6.1 at the time). It seems much longer :-( As I have mentioned on this list before, I have been working steadily to implement this for quite a while. To be sure that the dependencies don't get messed up, I've been working on getting the updates into stable and only then creating a kind of "supplementary bindist" which will be put into 0.6.2-current (but only for 10.3 users). Why does this have to be subject to such a heavy procedure? Replacing an obviously broken package (*) by anything else is always an improvement. This should be possible on a case-by-case basis and be done quickly. This question is independent of the creation of a complete complementary bindist for Panther. The release of the supplementary bindist should be less than a week away at this point. Very good. -- Martin (*) Note that the failure of qt3 made the whole kde part of the bindist non-functional on Panther. --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] CVS conflict detection
Am 11.01.2004 um 00:21 schrieb Ben Hines: Fink should be able to realize what this is: WARNING: Unable to parse the line "<<< libbonobo2.info" in "/sw/fink/dists/unstable/main/finkinfo/gnome/libbonobo2.info". WARNING: Unable to parse the line "===" in "/sw/fink/dists/unstable/main/finkinfo/gnome/libbonobo2.info". WARNING: Unable to parse the line ">>> 1.5" in "/sw/fink/dists/unstable/main/finkinfo/gnome/libbonobo2.info". WARNING: Unable to parse the line "<<< libbonobo2.info" in "/sw/fink/dists/unstable/main/finkinfo/gnome/libbonobo2.info". WARNING: Unable to parse the line "===" in "/sw/fink/dists/unstable/main/finkinfo/gnome/libbonobo2.info". WARNING: Unable to parse the line ">>> 1.5" in "/sw/fink/dists/unstable/main/finkinfo/gnome/libbonobo2.info". WARNING: Unable to parse the line "<<< libbonobo2.info" in "/sw/fink/dists/unstable/main/finkinfo/gnome/libbonobo2.info". We could have fink notify the user of a possible CVS conflict and offer to remove the file and selfupdate-cvs again. We could also suggest that if they want to keep the changes they should say no and resolve the conflict by hand. Sounds like a good idea. Maybe throw in a ref to an appropriate FAQ item, too= Cheers, Max --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] fink-maintainer-gaim :: gaim >= 0.71 please
Martin Costabel <[EMAIL PROTECTED]> wrote: [snip] > I have been trying for ages to suggest to use 'current' for replacing > packages which are known to be broken. There are several packages in > 0.6.2 that are known not to work on 10.3 and IMHO it absolutely stinks > that after almost 3 months Fink 0.6.2 was released on 17 November 2003, less than two months ago. > we still have to tell people every few days: > Well, you know, if you look at page 3 of the news on Fink's home page, > it tells you that not all binary packages of 0.6.2 are actually working > on Panther; the thing you are trying to run (qt3, for example) is one of > the exceptions; you have the choice between compiling them yourself (but > for qt3, this will take 3 hours and usually does not work the first time > you try) and getting them from bbrauns inofficial unstable bindist. > As I have mentioned on this list before, I have been working steadily to implement this for quite a while. To be sure that the dependencies don't get messed up, I've been working on getting the updates into stable and only then creating a kind of "supplementary bindist" which will be put into 0.6.2-current (but only for 10.3 users). The release of the supplementary bindist should be less than a week away at this point. -- Dave --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] fink-maintainer-gaim :: gaim >= 0.71 please
Ben Hines wrote: On Jan 8, 2004, at 7:24 PM, Max Horn wrote: We can't get anything into 0.6.2 - that's a fixed release :-). Not totally correct as you know, we do have an 'current' dir which we can throw packages into which point release stable users can access. For occasionally totally broken packages such as 'bidwatcher' at times (which can break when ebay changes formats), or gaim (which can break when MS changes formats) , it is a good use of it IMO. I have been trying for ages to suggest to use 'current' for replacing packages which are known to be broken. There are several packages in 0.6.2 that are known not to work on 10.3 and IMHO it absolutely stinks that after almost 3 months we still have to tell people every few days: Well, you know, if you look at page 3 of the news on Fink's home page, it tells you that not all binary packages of 0.6.2 are actually working on Panther; the thing you are trying to run (qt3, for example) is one of the exceptions; you have the choice between compiling them yourself (but for qt3, this will take 3 hours and usually does not work the first time you try) and getting them from bbrauns inofficial unstable bindist. -- Martin --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: FinkCommander Feedback ([Fink-devel] librep-shlibs-0.16.2-11)
hi I just did run into the same problem as Colin. (ibrep-shlibs-0.16.2-11 failed because of missing symbols Dl_info) After fink selfupdate, I tried to update-all packages Reading your suggestion, I upgraded to XCode 1.1 (otherwise, I thought my dev tools are ok) but the error is still there. I just got the error: I had another dlfcn.h installed in /usr/local/include which was found before /usr/include (as it is supposed to be unfortunately). Also I can't remember who did install it there. (hm, having sub-dirs "ogg vorbis" and "FL" I suspect some sound tools I did install) OK. So Colin, so if your tools are correct, check for a duplicate dlfcn.h in /usr/local/include Matthias Ringwald On Jan 10, 2004, at 2:43 AM, Benjamin Reed wrote: Colin Alworth wrote: unix_dl.c: In function `rep_find_c_symbol': unix_dl.c:466: error: `Dl_info' undeclared (first use in this function) if you don't have these, your dev tools are incomplete. Try reinstalling XCode and make sure you install the BSD SDK. -- Benjamin Reed a.k.a. Ranger Rick -- http://ranger.befunk.com/ gpg: 6401 D02A A35F 55E9 D7DD 71C5 52EF A366 D3F6 65FE "Just try to imagine a world where e-mails are sent by your brain before they are written, and are ready before they arrive by people you have never even met in countries you have never even heard of!" --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
