From rworkman at slackbuilds.org Mon Jul 15 17:56:58 2013 From: rworkman at slackbuilds.org (Robby Workman) Date: Mon, 15 Jul 2013 12:56:58 -0500 Subject: [sbopkg-users] [Slackbuilds-users] nvidia-driver.info missing 64bit file info In-Reply-To: References: Message-ID: <20130715125658.0f714fef@home.rlworkman.net> On Mon, 15 Jul 2013 18:52:20 +0100 "Greg' Ar Tourter" wrote: > All the recently released/updated nvidia*-driver.info file are > missing the source files and associated md5sum for the 64bit version. > There are indeed common to bot the 32 and 64 bit version but need to > be specified for both. Currently sbopkg fails because it isn't > downloading the tar.xz files on 64bit machines since only the .run > file is specified. Hrm, my fault, but I'm not convinced that sbopkg is DTRT. I'm going to CC the sbopkg list to solicit feedback, but it's my point of view that sbopkg should *always* pull the things specified in $DOWNLOAD, and if on x86_64, pull the things specified in $DOWNLOAD_x86_64. Unless my memory is failing me, that is how we envisioned things working when we added the *_x86_64 variables to the .info file. -RW -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: not available URL: From rworkman at slackbuilds.org Tue Jul 16 05:52:10 2013 From: rworkman at slackbuilds.org (Robby Workman) Date: Tue, 16 Jul 2013 00:52:10 -0500 Subject: [sbopkg-users] [Slackbuilds-users] nvidia-driver.info missing 64bit file info In-Reply-To: References: <20130715125658.0f714fef@home.rlworkman.net> Message-ID: <20130716005210.45ecbdd5@home.rlworkman.net> On Mon, 15 Jul 2013 20:31:58 +0100 "Greg' Ar Tourter" wrote: > Hi Robby, > > Well looking through the repository, the nvidia-driver (and > avr8-gnu-toolchain, but that's only for the pdf manual) packages are > the only ones with files in common. So asking to download both > version seems rather a waste of bandwitdh Okay, fair point. > Moreover, for some packages (ucsc-blat, p4, p4v ), the source file > name is the same for 32 and 64 bit version but are downloaded from > different locations and have different checksum. in these case, wget > with append .1 to the second one and therefore making the slackbuild > use the 32bit version instead. Oh wow, I wasn't aware of that. Given all of this and erik's reply, I think it's probably a good change to make. Don't feel obligated to do this, as I don't mind doing it myself, but since I'm a bit tied up on another project right now, patches for all of the affected stuff would make this easier :) Also, since you pointed it out, you should be credited with the fixes anyway. Seriously, though, if you don't want to do it or don't have time, that's fine. -RW -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: not available URL: From rworkman at slackbuilds.org Wed Jul 17 04:25:56 2013 From: rworkman at slackbuilds.org (Robby Workman) Date: Tue, 16 Jul 2013 23:25:56 -0500 Subject: [sbopkg-users] [Slackbuilds-users] nvidia-driver.info missing 64bit file info In-Reply-To: <1626782.psE4A77K31@callisto> References: <20130715125658.0f714fef@home.rlworkman.net> <1626782.psE4A77K31@callisto> Message-ID: <20130716232556.3b9eb147@home.rlworkman.net> On Tue, 16 Jul 2013 11:26:01 +0200 Heinz Wiesinger wrote: > On Monday 15 July 2013 12:56:58 Robby Workman wrote: > > On Mon, 15 Jul 2013 18:52:20 +0100 > > > > "Greg' Ar Tourter" wrote: > > > All the recently released/updated nvidia*-driver.info file are > > > missing the source files and associated md5sum for the 64bit > > > version. There are indeed common to bot the 32 and 64 bit version > > > but need to be specified for both. Currently sbopkg fails because > > > it isn't downloading the tar.xz files on 64bit machines since > > > only the .run file is specified. > > > > Hrm, my fault, but I'm not convinced that sbopkg is DTRT. > > > > I'm going to CC the sbopkg list to solicit feedback, but it's my > > point of view that sbopkg should *always* pull the things specified > > in $DOWNLOAD, and if on x86_64, pull the things specified in > > $DOWNLOAD_x86_64. Unless my memory is failing me, that is how we > > envisioned things working when we added the *_x86_64 variables to > > the .info file. > > It's a little bit more complicated. The way it was meant: When > DOWNLOAD_x86_64 is empty, use DOWNLOAD (indicating source is the same > for both platforms), if filled (and not UNSUPPORTED/UNTESTED) use > DOWNLOAD_x86_64 (different source per platform). And then there is > also the rather rare case where DOWNLOAD could be > UNSUPPORTED/UNTESTED (app works only on x86_64). This is also more or > less how the website logic works when displaying download links to > the sources. Okay, that makes sense - thanks for the clarification. We should probably document that somewhere public if we haven't already. I should probably keep up better ;-) -RW -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: not available URL: From chess at chessgriffin.com Wed Jul 17 11:50:32 2013 From: chess at chessgriffin.com (Chess Griffin) Date: Wed, 17 Jul 2013 07:50:32 -0400 Subject: [sbopkg-users] [Slackbuilds-users] nvidia-driver.info missing 64bit file info In-Reply-To: <20130716232556.3b9eb147@home.rlworkman.net> References: <20130715125658.0f714fef@home.rlworkman.net> <1626782.psE4A77K31@callisto> <20130716232556.3b9eb147@home.rlworkman.net> Message-ID: <20130717075032.1ea926e5@slack.localdomain> On Tue, 16 Jul 2013 23:25:56 -0500 Robby Workman wrote: > On Tue, 16 Jul 2013 11:26:01 +0200 > Heinz Wiesinger wrote: > > > On Monday 15 July 2013 12:56:58 Robby Workman wrote: > > > On Mon, 15 Jul 2013 18:52:20 +0100 > > > > > > "Greg' Ar Tourter" wrote: > > > > All the recently released/updated nvidia*-driver.info file are > > > > missing the source files and associated md5sum for the 64bit > > > > version. There are indeed common to bot the 32 and 64 bit > > > > version but need to be specified for both. Currently sbopkg > > > > fails because it isn't downloading the tar.xz files on 64bit > > > > machines since only the .run file is specified. > > > > > > Hrm, my fault, but I'm not convinced that sbopkg is DTRT. > > > > > > I'm going to CC the sbopkg list to solicit feedback, but it's my > > > point of view that sbopkg should *always* pull the things > > > specified in $DOWNLOAD, and if on x86_64, pull the things > > > specified in $DOWNLOAD_x86_64. Unless my memory is failing me, > > > that is how we envisioned things working when we added the > > > *_x86_64 variables to the .info file. > > > > It's a little bit more complicated. The way it was meant: When > > DOWNLOAD_x86_64 is empty, use DOWNLOAD (indicating source is the > > same for both platforms), if filled (and not UNSUPPORTED/UNTESTED) > > use DOWNLOAD_x86_64 (different source per platform). And then there > > is also the rather rare case where DOWNLOAD could be > > UNSUPPORTED/UNTESTED (app works only on x86_64). This is also more > > or less how the website logic works when displaying download links > > to the sources. > > > Okay, that makes sense - thanks for the clarification. We should > probably document that somewhere public if we haven't already. > I should probably keep up better ;-) > > -RW I'm trying to follow this thread but apparently I haven't had enough coffee yet. ;-) Is there still an open issue for sbopkg? Something we need to look at? -- Chess Griffin From slakmagik at gmail.com Wed Jul 17 23:26:01 2013 From: slakmagik at gmail.com (slakmagik) Date: Wed, 17 Jul 2013 19:26:01 -0400 Subject: [sbopkg-users] [Slackbuilds-users] nvidia-driver.info missing 64bit file info In-Reply-To: <20130717075032.1ea926e5@slack.localdomain> References: <20130715125658.0f714fef@home.rlworkman.net> <1626782.psE4A77K31@callisto> <20130716232556.3b9eb147@home.rlworkman.net> <20130717075032.1ea926e5@slack.localdomain> Message-ID: <20130717232601.GA5797@slacktop> On 2013-07-17 (Wed) 07:50:32 [-0400], Chess Griffin wrote: > On Tue, 16 Jul 2013 23:25:56 -0500 > Robby Workman wrote: > > > On Tue, 16 Jul 2013 11:26:01 +0200 > > Heinz Wiesinger wrote: > > > > > On Monday 15 July 2013 12:56:58 Robby Workman wrote: > > > > On Mon, 15 Jul 2013 18:52:20 +0100 > > > > > > > > "Greg' Ar Tourter" wrote: > > > > > All the recently released/updated nvidia*-driver.info file are > > > > > missing the source files and associated md5sum for the 64bit > > > > > version. There are indeed common to bot the 32 and 64 bit > > > > > version but need to be specified for both. Currently sbopkg > > > > > fails because it isn't downloading the tar.xz files on 64bit > > > > > machines since only the .run file is specified. > > > > > > > > Hrm, my fault, but I'm not convinced that sbopkg is DTRT. > > > > > > > > I'm going to CC the sbopkg list to solicit feedback, but it's my > > > > point of view that sbopkg should *always* pull the things > > > > specified in $DOWNLOAD, and if on x86_64, pull the things > > > > specified in $DOWNLOAD_x86_64. Unless my memory is failing me, > > > > that is how we envisioned things working when we added the > > > > *_x86_64 variables to the .info file. > > > > > > It's a little bit more complicated. The way it was meant: When > > > DOWNLOAD_x86_64 is empty, use DOWNLOAD (indicating source is the > > > same for both platforms), if filled (and not UNSUPPORTED/UNTESTED) > > > use DOWNLOAD_x86_64 (different source per platform). And then there > > > is also the rather rare case where DOWNLOAD could be > > > UNSUPPORTED/UNTESTED (app works only on x86_64). This is also more > > > or less how the website logic works when displaying download links > > > to the sources. > > > > > > Okay, that makes sense - thanks for the clarification. We should > > probably document that somewhere public if we haven't already. > > I should probably keep up better ;-) > > > > -RW > > I'm trying to follow this thread but apparently I haven't had enough > coffee yet. ;-) Is there still an open issue for sbopkg? Something > we need to look at? > No, I think we're good. :) From woquendo at gmail.com Thu Jul 25 16:48:01 2013 From: woquendo at gmail.com (William Oquendo) Date: Thu, 25 Jul 2013 11:48:01 -0500 Subject: [sbopkg-users] [sbopkg] How to pass configure options to a build through a queue file? Message-ID: Hello everybody. I would like to know how to add a configure flag (not a compilation one) to a queue file. Specifically, it is for netcdf. I tried to rewrite the sqf file as: netcdf | --enable-cxx-4 but the flag is not read when the package is configured. Can anybody please tell me how to specify the flag? I have searched through google with no luck. -- Best regards / Cordialmente, William-Fernando Oquendo -------------------------- Este correo puede carecer de tildes o e?es debido al teclado. -------------- next part -------------- An HTML attachment was scrubbed... URL: From martin.matejek at gmx.com Tue Jul 30 22:38:49 2013 From: martin.matejek at gmx.com (Martin Matejek) Date: Wed, 31 Jul 2013 00:38:49 +0200 Subject: [sbopkg-users] Sbopkg logrotate Message-ID: <20130730223849.127990@gmx.com> Hi list, today while I was trying to determine why my system partition is running out of space, I found out that sbopkg build log's size is 124 MB, which is quite large for plain text file. Mine starts Thu Oct ?4 01:24:40 CEST 2012 and since then I've build and rebuild many packages (running -current). I know that on stable Slackware branch this log usually don't grow much, but if you are tracking current and don't reinstall slack with every new release, this log could grow to quite large file over the time. Have you thought about including logrotate config file for sbopg? It could be for long period of time like month or so, but it wouldn't hurt to rotate logs after some time. I already created one (http://webdev.fit.cvut.cz/~matejm14/linux/sbopkg), so feel free to use/modify it if you are interested in this idea. Martin From chess at chessgriffin.com Wed Jul 31 02:48:36 2013 From: chess at chessgriffin.com (Chess Griffin) Date: Tue, 30 Jul 2013 22:48:36 -0400 Subject: [sbopkg-users] Sbopkg logrotate In-Reply-To: <20130730223849.127990@gmx.com> References: <20130730223849.127990@gmx.com> Message-ID: <20130730224836.0d6787bd@slack.localdomain> On Wed, 31 Jul 2013 00:38:49 +0200 "Martin Matejek" wrote: > > Have you thought about including logrotate config file for sbopg? It > could be for long period of time like month or so, but it wouldn't > hurt to rotate logs after some time. > > I already created one > (http://webdev.fit.cvut.cz/~matejm14/linux/sbopkg), so feel free to > use/modify it if you are interested in this idea. > > Martin Nice, thanks, Martin. This might be a good candidate for the contrib/ directory. Cheers! -- Chess Griffin From chess at chessgriffin.com Wed Jul 31 02:52:18 2013 From: chess at chessgriffin.com (Chess Griffin) Date: Tue, 30 Jul 2013 22:52:18 -0400 Subject: [sbopkg-users] [sbopkg] How to pass configure options to a build through a queue file? In-Reply-To: References: Message-ID: <20130730225218.3a26b990@slack.localdomain> On Thu, 25 Jul 2013 11:48:01 -0500 William Oquendo wrote: > Hello everybody. > > I would like to know how to add a configure flag (not a compilation > one) to a queue file. Specifically, it is for netcdf. I tried to > rewrite the sqf file as: > netcdf | --enable-cxx-4 > but the flag is not read when the package is configured. Can anybody > please tell me how to specify the flag? I have searched through > google with no luck. > Sbopkg relies on the SlackBuild scripts for passing configuration options. Some SlackBuild scripts have this capability built in, in which case you can use the Options selection when in the package menu. ffmpeg is an example of this. For those that don't offer this, then you'll need to manually edit the SlackBuild (which you can do from within sbopkg) and then select the locally-edited version of the Slackbuild when building. -- Chess Griffin