My thoughts on Slackware, life and everything

Category: Me (Page 18 of 27)

End of the week, end of the cycle?

Most of you will have seen the latest comment in the slackware-current ChangeLog:

Wed Sep 19 23:52:16 UTC 2012
Here we go one more time with Slackware 14.0 release candidate 5.
Really, this time it is not a drill!  Everything is in place and
ready to release at this point, and unless there’s some kind of
showstopper found (which doesn’t seem too likely after all the
testing that’s happened), the release can be expected soon.

It means that we could see a Slackware 14 release very soon. As always, it’s ready when it is ready… but it is not going to take weeks.

I took the opportunity to implement some ideas which I had been thinking about for my multilib packages. These are the results:

  • In the compat32-tools package, the “convertpkg-compat32” script will now by default add a “compat32” build tag to the converted package name. For instance, when the package “e2fsprogs-1.42.5-x86_64-1.txz” is converted, it becomes “e2fsprogs-compat32-1.42.5-x86_64-1compat32.txz”.
  • Also in the compat32-tools package, the “massconvert32.sh” script will now check the Slackware patches directory to see if there is an update for any package it needs to convert. For instance, when you run the script against a Slackware 13.37 package tree you would get the converted package “openssl-solibs-compat32-0.9.8x-i486-1_slack13.37compat32.txz” instead of “openssl-solibs-compat32-0.9.8r-i486-3compat32.txz” (which is part of the original 13.37 release).
  • The reason for adding a “compat32” build tag to all converted packages is to make system upgrades of a multilib Slackware easier. Up to now, if you were using slackpkg for the upgrades, you had to manually deselect all compat32 packages in the list which is produced by the command “slackpkg clean-system”. With the new scripts, you are able to blacklist all my multilib packages by just adding one line to the file “/etc/slackpkg/blacklist”:

# Blacklist all multilib ‘compat32’ packages:

[0-9]+compat32

  • During the slackware-current development cycle there were several upgrades of the gcc packages. At one time I had to fix my multilib rebuild of gcc. I want to keep my package build numbers in sync with Slackware’s original 64-bit packages to avoid confusion of the kind “do I have the correct multilib package installed“, so I decided to give the fixed gcc packages a build number of “1fix1_alien” instead of “1alien”. Unfortunately this broke the slackpkg blacklist line for my “alien” tag. The expression “[0-9]+alien” will not match packages with a “1fix1_alien” build number. So I decided to rename the multilib gcc packages and use the “1alien” build tag. This will be much friendlier for people who upgrade from Slackware 13.37 to 14.0. If you have been running -current all the time, you should be smart enough to understand my reasoning 🙂

If you are currently running Slackware 13.37 and want to profit from these enhancements, you can of course upgrade to my new compat32-tools package – even though I make it avalable in the “current” (and later on “14.0“) directory of my multilib repository. I took care not to break the compatibility with Slackware 13.37 when I updated the package during the past development cycle of slackware-current.

A note about the Slackware Documentation Project

We  (the editors) are steadily working on expanding the documentation wiki. I recently added an article about how to perform a Slackware system upgrade, to help people who are running Slackware 13.37 and want to upgrade to 14.0 when that is released. Check out “Upgrade Slackware to a New Release“.

We would like to welcome your contributions as well! If you had a problem in Slackware and found the solution, or if you have some particular knowledge which your fellow Slackers could profit from, feel free to visit the Wiki’s HOWTO’s page and create your own article there. Check the existing HOWTO articles to get a feel for what is possible.

If you do not want to write a new article, we still value your feedback. If you have any comments or suggestions about improvements for the site, we would like to hear from you.

Cheers, Eric

My Panasonic Lumix DMC-LX5

Back in february I bought the camera I had had my eyes on for over a year. It was a bit too expensive to consider it seriously, but then suddenly there was a discount of 75 euros and I grabbed the opportunity. A Panasonic DMC-LX5 was finally mine.

I had been using a relatively cheap Sony CyberShot DSC-P10 camera for years, after my first digital compact, a Canon PowerShot, broke down with the infamous E18 error (which earned its own web site http://e18error.com because so many people suffered by it). The Sony camera was meant to be temporary because I did not have money for a good replacement… but the money never came. I needed a better camera badly. What I really wanted was better light sensitivity and better wide-angle.

The LX-5 offers all of that and much more, with its f/2.0 24mm Leica lens, fast power-on and AF, full metal body and a battery that should last 400 photos on a single charge. Read more about its features, strengths and weaknesses (if any) at the PhotographyBLOG. I realize that this camera is not “new”, and lots of other types have entered the market and are competing with it, but I wanted a camera that is easy to carry, enables me to point-and-shoot while keeping the possibility for manual overrides, is sturdy and offers excellent picture quality in low-light circumstances. I don’t care for replaceable lenses or GPS geo-tagging, or superzoom lenses.

I discussed cameras with collegues and friends who shoot a lot of pictures, and the outcome was that I would best by either the LX5 or a Canon PowerShot S95… the latter because it has quite similar specs to the LX5 but also because there is custom firmware for the Canon line which lets you do a whole lot more with your camera: the Canon Hack Development Kit (CHDK).

In the end I decided against the Canon, the main reason being resentment about the E18 error which killed my first digital camera 1 day after end of warranty.

Why did it still take a year to buy the camera? I guess it was the price point. But I was following several photography related blogs and my enthousiasm was fueled by viewing the results of shooting with a LX5. That basically pushed me that final step. Special mention to Juha Haataja’s Light Scrape blog! I was fascinated by Juha’s stills, his comments about his daily encounters with the world and his overall philosophy with regard to what makes a good picture.

I took the camera with me (of course) during my holiday in Brittany, France. Back home and behind a computer I could evaluate the results. At first, I stuck mostly to the camera’s defaults (using Intelligent Auto or “iA”) because I still needed to get acquainted to it (for some reason, I never took the time to get to know the LX5 during those first months that I owned it – too busy with Slackware I guess).  But I quickly began to experiment, with the User Manual as a guide… luckily I downloaded that to my E-reader.

I don’t know if it is the result of my fiddling with the manual controls (or even fiddling too little) but I feel that some of the JPEGs, especially those taken in bright sunlight, seem a bit “ovet-processed”. Or perhaps that is just because I have been working with a mediocre camera for so long. I discussed this with Pat (who owns a Canon PowerShot and uses CHDK) and he suggested that I should try shooting JPEG+RAW and look what difference it would make when I do the image post-processing on my computer. So, that is what I am going to do in the next couple of weeks. I compiled an updated package of RAWtherapee, an open source RAW editor with support for the LX5’s RAW image format and I am going to play with that once I have shot a couple of RAW images.. If anyone knows equivalent or better alternatives to RAWtherapee (free software!) I would like to hear from you.

Here’s a couple of sample images from that first holiday batch (all JPEG). Let me state that I am very happy with the camera. The LX5 has a good grip, and the controls are easy to operate. The camera “radiates” quality. I could shoot pictures in awkward positions without fear of letting it slip. Its low light qualities are great, and I am curious to find out about all the possibilities of the manual controls. So far, it has enabled me to take pictures in situations where I would have failed with the old Sony. Despite my earlier comment about possible JPEG over-processing, I am impressed with the quality of the pictures.

Tell me what you thnk of them. Can the camera do better? Can I do better?

Eric

 

 

 

Busy days, not Slackware related

In the next days or weeks, I am going to try and rest and re-vitalize myself. It would be a waste of effort if I burnt myself out. But the shitty weather does not help. Who feels like it is summer in Europe? I have just emptied the buckets in the hallway which caught the water seeping through the crack in the roof… repair money is not available right now.

During the past week, there were many such distractions to keep me away from hacking at Slackware. This will likely remain unchanged during the next weeks. We had a fire in the kitchen, which caused damage (the stove and oven were destroyed), but luckily no one got hurt! It takes time to invite experts to assess the damage, arrange repairs and such. I als got hooked on some e-books I had loaded onto my Sony E-reader… reading is eating away precious time faster than you think!

Anyway, all this distraction had consequences for the package pipeline. With regular releases of KDE, LibreOffice, OpenJDK and VLC – all pretty big builds – my free time is increasingly limited to building updates for these aforementioned programs. It is frustrating at times that I have to make an advance planning because especially KDE and LibreOffice releases tend to come at roughly the same time. Building in several virtual machines at the same time hurts my server’s performance and it does not help the total build time per program…

Also, I really need to pick up the pace with my new but currently stalled ARM port of Slackware (which will have differences to the existing ARMedslack port). I am talking to some people about what would be a cool computer to own which really should be running Slackware. The outcome of those discussions I will reserve for a future post, because I will probably need assistance.

That is why I decided that I am not going to build packages for the upcoming KDE 4.9-rc2. I will wait for the final release of 4.9.0 instead.

I will try to get LibreOffice packages compiled for the 3.5.5 release which was announced today. It depends on the errors I encounter during compilation… I do not have time to hunt compilation issues down and fix them. So, fingers crossed!

End of rant.

Eric

Blog updated to 3.4, sqlite database fixed

I just upgraded my blog to WordPress 3.4.

My blog’s database backend is not MySQL but instead (for portability reasons) I am using PDO for WordPress, a plugin which uses PDO (a PHP data access abstraction layer) to allow the use a sqlite database; i.e. a simple file.

Unfortunately this sqlite backend is not 100% compatible with WordPress installations starting with version 3.3. You can not even create a new blog using the PDO plugin with WordPress 3.3 and higher. The blog has to be created using wordpress-2.2.x, and then upgraded to 3.3 or later.

It looks like the maintainer of the PDO plugin will not be able to give adequate support for the sqlite backend. Not because he does not want to, but because it is difficult to create the proper glue between WordPress SQL code and generic SQL backends (WordPress developers have stated that they are not interested in supporting anything else than MySQL).

The problems with the PDO plugin show when creating a new post… the blog will be “dead” for at least five minutes and the apache log full of SQL errors. I am considering to  move “back” to a supported backend and wanted to check if I could export the full blog so that I could import it to a fresh installation with a MySQL backend.

I quickly discovered that this was not possible. Somewhere along the line, WordPress has added a new table to the database (wp_commentmeta) and the PDO plugin apparently has not been able to add that to my sqlite database. An attempt to export the blog to an XML file (one of the standard tools in WordPress) would lead to a lot of errors in the apache log about “Problem preparing the PDO SQL Statement.  Error was no such table: wp_commentmeta“.

I compared my blog’s sqlite database to a fresh MySQL-backed blog database and thus found the new table’s properties. I upgraded the database on the fly using the sqlite3 command-line program as follows:

$ sqlite MyBlog.sqlite > CREATE TABLE wp_commentmeta (
 meta_id integer NOT NULL PRIMARY KEY AUTO_INCREMENT,
 comment_id integer NOT NULL DEFAULT '0',
 meta_key text DEFAULT NULL,
 meta_value text); > .quit

After that modification, exporting the blog to XML took only seconds to create a 4.3 MB file instead of the previous 400 KB partial dump which took painstakingly long. I still have not tried to import that XML dump into an empty blog. Also, this new post will be a test to see how the blog recovers… will it still take many minutes before it is available again after I press “Publish”?

Eric

End of may ramblings

It has been a while since my latest blog post, so I thought it would be good to talk a bit about the goings-on.

First about the Slackware web server. I know, people, that we have been without http://www.slackware.com/ for a few weeks now, but rest assured that the site will eventually be back. Either we put the old web site code on a temporary server or we wait a little longer and publish a new site based on a new CMS. It all pretty much depends on how much time we can volunteer for this – it is mostly handled by a few people in the coreteam. Pat should focus on Slackware and make sure we get a new stable release at some point – there is a lot to do still on that front.

And then about my own packaging activities.

There was a new release of LibreOffice yesterday, and I am currently building packages for that. If there are no issues with the build then you can expect those packages tomorrow.

KDE release team is in the process of releasing the first beta of KDE SC 4.9, and I intend to make packages for that. There was a heated discussion about this beta when the new release manager announced he was going to call off the beta1 release… but all issues have been resolved yesterday and new tarballs are going to be made available ahead of time to the packagers. There are quite a few changes compared to KDE 4.8.x so it will cost me a while to work out the updates to the KDE.SlackBuild framework.

I also hope that VLC releases their version 2.0.2 soon. It has been lingering just around the corner for a while, but a recent fall-out between several of the core developers threatened the whole project’s existence when their most important Linux developer quit the team out of frustration. That animosity has now subsided, the team is whole again, and development is progressing toward a new release.

I also want to thank the kind people who donated a few bucks after all the upheavals about the Slackware webserver’s outage. Although I work on Slackware in my spare time, because I like it, it’s my main hobby so to speak, getting some funds enabled me to buy a new and faster build box last year, and now I ordered a Raspberry-Pi (finally…) and I am also looking for a good tablet which allows the installation of a different OS than only Android… so that I can put my unfinished port of Slackware to modern ARM architectures on it and finally release that. A release does not make sense if it runs on only one device (the Trim Slice). Perhaps I will buy some more ARM hardware too. Even Pat was becoming excited about this ARM port.

Eric

« Older posts Newer posts »

© 2025 Alien Pastures

Theme by Anders NorenUp ↑