OpenBSD Journal

p2k17: Herzliche grusse vom Berlin (espie@ on mandoc, misc packages progress)

Contributed by Peter N. M. Hansteen on from the we remember puffy by the wall dept.

A new p2k17 hackathon report is in, from Marc Espie, who writes:
First time in Berlin.

I've been to germany lots of time, but never to Berlin. I'm actually old enough that my first german foray happen before 1989.

It was interesting. Very wide city, with lots of spaces everywhere.

Makes it slightly inconvenient for sight-seeing, as you need to walk a lot or take public transportation to see anything. Almost the opposite of Paris in that regard.

Got me thinking about the Wall, and now I will definitely have to look at historic maps, because the current network appears seamless, but it must have had to deal with the Wall disappearance at some point.

I arrived early with aja@ on the 1st, having taken the easy way (plane+bus and a bit of walk in drizzly rain). I know that some brave souls took the train, and jca@ actually took the bus all the way in, and had a few misadventures on the way... he finally got in 24 hours later.

Quite cozy room, the frogs took over the most comfy place near the entrance, of course. That plush sofa. <3

Most of my work this hackathon isn't visible yet. Right before I left Paris, some emails on the lists made me aware our mandoc hackers weren't very savvy about PostScript, so I shrank the generated output by over 50%, and I started really looking at it. So I wrote a quick and dirty AFM (adobe font metrics) parser so that I could get the output to look nicer. In case you're not aware of it, there's this thing called kerning, where you have to adjust spaces between characters to they looked nicer. For instance, you might put AV closer, and rt slightly farther apart, depending on the font. PostScript itself doesn't provide for that, you have to manually space the characters correctly.

I actually have a prototype that does everything correctly, and can use all the default font families in PostScript/PDF for output.

Now, we mostly need to figure out what to put in the mandoc directory (generated files, or afm files, which are under an acceptable licence, according to a2ps files), and I need to convince Ingo that yes, being able to choose the font family is desireable... probably the toughest part. In typical Ingo fashion, after I sent my first patch, the answer was along the lines: "I've got to learn some PostScript first, before I can get back to you".

That little side trip was mostly because everything else I wanted to work on was bound to be somewhat painful. I made some progress in getting the ports tree ready for lld -- which is already used on arm64, so fixing machine-independent issues on amd64 is less painful but also helps arm64. This might also give some breathing space to 32 bit architectures which are having lots of trouble keeping up with necessary linking space for browsers.

With a few commits to go, most of the ports tree is now in good shape for that. At least, it compiles. Doesn't mean it runs correctly -- One step at a time! Ironically, it means I was mostly immune to the samba binutils breakage, as lld is completely not affected.

Being in the same room and on the same timezone as other porters helps as well. jeremy@ prodded me into fixing a long standing buglet in pkg_add's scp mode, I helped naddy@ and afresh1@ on the ways of perl's time handling (short version: RTFM Time::HiRes), and sthen@ and I brought CCACHE back from the dead for dpb. landry@ also uncovered that sparc64 COMPILER handling had been broken for a few months. Doesn't seem like people care any more. Makes me sad, because the COMPILER stuff was actually done to try to keep things working for longer. That bug actually affected any platform trying to use clang from ports that doesn't have native clang in base. That got fixed, and about half the ports tree might be building again on sparc64.

There were also a few discussions about further developments, such as debug symbols in packages. And also figuring out ways to get ports development better integrated with dpb, as the privilege separation stuff tends to get in the way. But we've got to keep things to do, so that we get invited to other places.

The weather was fickle this week, so we didn't get very far from the hackroom on most days. Surprisingly cheap but good food was to be had, especially from the falafel/chawarma place at the end of the street, but we did manage to get some sight-seeing on sunday.

Between the open-air art exposition on the remaining wall (thx landry@), and the magnificent Pergamon museum. Kuddoes to aja@'s father for recommending that one. The exhibition is fairly impressive.

I got to practice my german a bit, which is still quite rusty. I generally can pass for a local when I limit myself to "Hallo" and shit, which usually gets me into trouble when the actual germans start talking really fast and using dialect in ways I don't understand.

Again, thanks for every organizer, and especially stsp@, for making this a great hackathon. Also thanks for the association who gave us this space graciously.

I gypped a few nice stickers, and had an epiphany looking at the NetBSD sticker.

Judging by the violet and orange color, this is obviously just a Halloween prank gone bad.

Until we meet again. Where will we go next time ?


Credits

Copyright © - Daniel Hartmeier. All rights reserved. Articles and comments are copyright their respective authors, submission implies license to publish on this web site. Contents of the archive prior to as well as images and HTML templates were copied from the fabulous original deadly.org with Jose's and Jim's kind permission. This journal runs as CGI with httpd(8) on OpenBSD, the source code is BSD licensed. undeadly \Un*dead"ly\, a. Not subject to death; immortal. [Obs.]