Older blog entries for yeupou (starting at number 63)

bind9 (1:9.2.4-1) unstable; urgency=high

* New upstream version. Closes: #269157 and others.
* Version debhelper build-dep. Closes: #262720

-- LaMont Jones <lamont@mmjgroup.com> Thu, 23 Sep 2004 09:11:37 -060

What the f+ck is #269157 about, what are "others"? One more time, to gain less than a minute, a debian maintainer decided that each user of the package he maintains will have to loose 5 minutes if he wants to know what is going on. It's really a pity, debian packaging is not at all homogenous: you have tremendous packagers, busy guys... and apparently people that absolutely refuses to do a trivial thing that would improve greatly user experience. And I do not see an end to that situation, as the debian docs that give example of good practices apparently failed to convince this last kind of maintainers.

nymia, I don't mean to be rude but whatever you wrote on a public website, like advogato.org, can be subject to discussion.

jamesh, reading details of your hard drive, even if it looks like good and fast hardware, were even not close to twice the perfs of most 7200rpm -- from my experience, it's no more than 15MB/s faster than usual 7200rpm hard drives. Not to mention that your hard drive is not a 7200rpm (a better comparison should have been done with 10000rpm hard drivers). Westerndigitals webpage also forgot to release noise-related information, while it seems important for a workstation.

Apart from that, I would say I'm usually reluctant to test brand new motherboards. Aren't you afraid to encounter tons of unrecognized hardware?

Last question: what system will you install on this x86-64 system? Debian?

Stevey, yes, dia is definitely a good piece of software. :-)

History: nymia, I'm afraid I do not get your point -- despite the fact that History is supposed to be my playground. Can you be more specific?

Music: I've just discovered the album "Guitar & Drums" of the Still Little Fingers and it definitely rocks.

Computing: I wonder what's the exact impact of the rj45 cables of my network on its speed. Its supposed to be 100mbps network but I'm not close at all to 25MB/s, I'm closer to 4MB/s or 5MB/s. I tested hard disk speed. On recent enough hardware, I usually got way more than 50MB/s so it cannot be the bottleneck (my older hard disk used on the network is a 3 year old Seagate). I heard that bad cables could create errors - I haven't such problems, only bandwidth issue. Or is it NFS?

Talking about harddisk, I recently bought an Hitachi 160GB IDE hard disk. All my other recent HD are Seagate "Barracuda" (40GB, 80GB). I have nothing to complain about this Hitachi HD performances. But it is way more noisy than the Seagate models -- while this harddisk is new, not 1 or 3 year old. This is no good.

nymia about your screenshots: the windows titles seems really hard to read; the panel around klipper looks ugly. I've seen cleaner KDE screenshots :)

I tested hevea, being told that it was a replacement for latex2html.

The test was not conclusive in favor of hevea. First of all, seing a function like -francais made me wonder why the hell do I need to pass such command line argument while this information is already provided within the LaTeX file? Afterward, it turned out that ps included images were simply not generated, without any link in the output anyway (I would not have been reluctant to run convert, if that would be the only thing to do).

It may be faster and cleaner than latex2html, it may be powerful along with macros. I have no time to write extra macros.

I'll probably retry at a later point but frankly it is not "plug and play" enough for me.

Today we posted at Gna! the first "hotspot". Let's hope we'll have the option to make numerous other "hospot", with interviewed persons as interesting as Philippe Gerum from the Adeos project.

Today, I realized the relation between the growing noise produced by my home server and the increase of the number of old hard disk inside it. I just realized that most of the noise come from two old hard 4.2 GB disks (Maxtor 90432D2, IBM-DCAA-34330).

One more thing I learned today: despite the fact it's never documented in the docs shipped with mainboards, mainboards do have IDE hard disk size limitations -- yes, it's not a big deal but it's always annoying to limit a 40GB hard disk to 32GB.

mathieu (given my name, it sounds to me to write this), you said "hearing someone say a language is better than another one just makes me want to yell".

It sounds a bit ultra-relativistic. I personally have the feeling that languages are not equally good to perform the same tasks. I base that impression on the fact that, to me, some languages have real disadvantages (PHP is slow, painy to debug, with no coherent function naming scheme) over others, even if they are not totally bad (otherwise we probably would not even know about their existence; PHP is easy to approach).

mathrick: The GNU Project recognize valid reasons not to use it, that why there is this Lesser GPL thing.

And, sure, it is not anyone obligation to love the GPL. But, the point is not here. Not anyone is writing software ; not anyone is writing software under a "free" license.

Cactus: imagemagick would not be portable and have so many dependancies? Well, I have no experience at all with imagemagick (apart using from time to time "import" to do screenshots and convert to do many kind of conversions within scripts), but according to packages.debian.org, it runs on the following architectures: alpha arm hppa i386 ia64 m68k mips mipsel powerpc s390 sparc. It is really not portable enough? And if you take a look at the dependancies, the strong depends you can find are: glibc (hum...), libbz2 and zlib1g (gzip, bzip2), libpng, libjpg, libjasper and libtiff (PNG, JPG and Tiff arent the most common image formats?) and finally a bunch of XFree packages you must have anyway once XFree is installed. So yes, you need XFree libs. But apart from that, the others depends are more than common -- and XFree is not exactly an obscure software known by only a few chosen ones. And, anyway, isn't your software depending on GTK+? If so, these depends are probably all met already by GTK+.

And in my experience, gdk-pixbuf is a bit slow, especially by comparison to imlib (well, I did not made any benchmarks, I just seen the difference once gqview dropped imlib).

I'm not saying you should not drop ImageMagick, you probably have good reasons. But saying it's not portable and have too many depends just sounds weird.

So pathetic to read "MIT licensed. The GPL can keep its dirty hands off my code"...