Bad *ware day

So Evolution decided to go sideways yesterday. I think this was prompted by me changing my SMTP server password. But rather than prompt me for a new password, Evolution simply froze when sending mail. As in “by-all-appearances-totally-hosed-because-the-UI-hasn’t-been-even-redrawn-in-several-minutes” frozen. Attempts to make Evolution forget the old password don’t change this outcome. Ugh.

While it would be ever so slightly gratifying to report a bug on this issue, I figure the Evolution developers won’t be particularly interested since I’m using version 2.22 and the latest is 2.24. Hell, I’d be inclined to blow me off on that basis. And I figure, “The Fedora 10 Preview has Evolution 2.24, I can just update to that.” What could go wrong, right? Well, if you’re the least bit familiar with these things, you know lots can go wrong; and you’re probably figuring that this is where things really start to go south.

But it’s not. In defiance of the odds, that part went okay. And when it was done, the shiny new Evolution 2.24 installation would send mail just fine. There was only one niggling problem: attempting to compose a new message now froze Evolution. Okay, so it just went mostly okay.

Ugh. But at this point I’m inclined to blame myself. After all, I knew I was tempting fate by upgrading to a partial set of Fedora 10 Preview packages. So I figure there must be some poorly-connected dependency I need to upgrade; and I figure “Fuck it; just upgrade the whole damn thing.” And I proceed to torrent the Fedora 10 Preview DVD image.

Deluge pulls it down at around 1 MB/s. Nice. On my bottom-tier cable connection. And I proceed to burn the disc image.

Hmm… Can’t mount that one. Crap. Try again (with Nautilus), but at a slower burn speed.

Damnit. That one won’t mount either. WTF? Check the SHA1 sum. It’s good. Okay, install K3B and try it. Its interface is a mess; but at least it’s produced reliable results in the past.

Fuck. Number three won’t mount and won’t boot. Another coaster.

Or is it? Stick it in the MacBook Pro. Hm. That reads it fine. Hm. Now let’s try stile (which runs i386 Fedora 9). Well fuck me. stile mounts it and boots it as well.

So let me get this straight: my DVD burner on hinge will burn apparently-valid DVDs that the same damn burner can’t read.

Fuck you, Plextor. Fuck you real hard.

This afternoon I ordered a Samsung DVD burner from Newegg for $28. I think I paid around $150 for the Plextor drive three or so years ago.

Meanwhile, I seem to have updated enough packages over the Internet to make Evolution happy. I’m trying to avoid updating X and the kernel so that I can keep my Nvidia driver happy.

"N"ane

Nvidia has stuck libraries that their libGL needs into a subdirectory; a subdirectory that the linker doesn’t know about. So the customary -lGL is insufficient; one needs to add

-L/usr/lib/nvidia

as well.

So I’ve made a lame hack to AX_CHECK_GL that should accommodate this. (Well, it’s no more—or less—lame than what I had to do for Mac OS X a little while ago.) And there’s a new release.

Dylan

I have neglected this space for far longer than usual. Time has been short; and most of the time I have free I’d rather spend doing stuff than writing about what’s been done/happening. But there has been no shortage of happenings.

Back on July 14, my son was born.

Dylan, day 2

Dylan, day 2

Now he’s nearly three months old. I’ll post some more recent pictures later. He’s only gotten better looking. For now, Gina’s posted some videos to Vimeo.

Fading memory

The Tyan K8WE motherboard in hinge has always seemed to have flaky DIMM slots. I seemed to have to wiggle the DIMMs just right in order for the machine to recognize all of the memory. A recent graphics card upgrade nudged things and I was unable to nudge them back into an agreeable position. The board had about a month left on its three year warranty; so, back to Tyan it went.

So I am without my primary development machine for at least a week. It’s hard to believe I’ve had this thing for nearly three years. Last year I upgraded the CPUs to Opteron 285s from the 242s I originally installed; but still it doesn’t seem that old. Its predecessor was a dual Pentium 3 machine that lasted me five years

When hinge is put back together, I think I’ll go ahead and release OpenVRML 0.17.6. I still haven’t fixed that damn JPEG bug; but I think there have been enough less consequential changes to warrant a release.

Another change I’d hoped to work into a point release in the 0.17 series is the use of D-Bus to replace the existing IPC mechanism used in openvrml-xembed. Aside from being generally cleaner and nicer than the existing hacky IPC solution, I’m thinking that expressing the IPC interface in D-Bus should lead to an improved code organization that will in turn make it easier to convert the existing multiprocess approach into a multithread approach for Mac OS X and Windows. This is becoming more pertinent as it seems there may be folks interested in contributing code to a Windows plug-in.

An RSS parser in PHP for SourceForge feeds

I wrote this for parsing RSS feeds associated with a SourceForge project. It should be reasonably capable for that purpose; though I do not expect it to be generally robust for handling arbitrary feeds. Reed suggested that this might be generally useful; however, I don’t want to put it anywhere that I might feel compelled to maintain it. So this seems like a good spot.

//
// Copyright 2008  Braden McDaniel
//
// This program is free software; you can redistribute it and/or modify it
// under the terms of the GNU General Public License as published by the Free
// Software Foundation; either version 3 of the License, or (at your option)
// any later version.
//
// This program is distributed in the hope that it will be useful, but WITHOUT
// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
// FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
// more details.
//
// You should have received a copy of the GNU General Public License along
// with this library; if not, see .
//

class RSS2Image
{
    var $url = "";
    var $title = "";
    var $link = "";
    var $width = "";
    var $height = "";
}

class RSS2Item
{
    var $title = "";
    var $link = "";
    var $description = "";
    var $author = "";
    var $category = "";
    var $comments = "";
    var $enclosure = "";
    var $guid = "";
    var $pub_date = "";
    var $source = "";
}

class RSS2Channel
{
    var $title;
    var $link;
    var $description;
    var $copyright;
    var $last_build_date;
    var $generator;
    var $image;
    var $items;

    function RSS2Channel()
    {
        $this->title = "";
        $this->link = "";
        $this->description = "";
        $this->copyright = "";
        $this->last_build_date = "";
        $this->generator = "";
        $this->image = null;
        $this->items = array();
    }
}

class RSS2Parser
{
    var $parser;
    var $channel;

    var $in_channel, $in_image, $in_item;
    var $current_element;

    function RSS2Parser()
    {
        $this->in_channel = false;
        $this->in_image = false;
        $this->in_item = false;
        $this->parser = xml_parser_create();
        xml_set_object($this->parser, $this);
        xml_set_element_handler($this->parser,
                                "start_element",
                                "end_element");
        xml_set_character_data_handler($this->parser,
                                       "character_data");
    }

    function parse($data)
    {
        xml_parse($this->parser, $data);
        return $this->channel;
    }

    function start_element($parser, $name, $attribs)
    {
        $name = strtolower($name);
        if ($name == "channel") {
            $this->in_channel = true;
            $this->channel = new RSS2Channel();
            return true;
        }

        if ($this->in_channel) {
            if ($name == "image") {
                $this->in_image = true;
                $this->channel->image = new RSS2Image();
                return true;
            } elseif ($name == "item") {
                $this->in_item = true;
                $this->channel->items[] = new RSS2Item();
                return true;
            }
        }

        if ($this->in_image) {
            $this->current_element = &$this->channel->image->$name;
        } elseif ($this->in_item) {
            $this->current_element = &end($this->channel->items)->$name;
        }
        return true;
    }

    function end_element($parser, $name)
    {
        $name = strtolower($name);
        if ($name == "channel") {
            $this->in_channel = false;
        } elseif ($name == "image") {
            $this->in_image = false;
        } elseif ($name == "item") {
            $this->in_item = false;
        }
        unset($this->current_element);
        return true;
    }

    function character_data($parser, $data)
    {
        if (isset($this->current_element)) {
            $this->current_element .= $data;
        }
        return true;
    }
}

Exception-safe management of JNI local references

…or, solving the problem of how to delete local references when the execution context isn’t clear.

JNI includes notions of “local” and “global” references. Loosely speaking, local references correspond to those that are local to the scope of a function; and global references correspond to those that persist outside the scope of a function. When Java calls a native method, it provides the native code with at stack frame where local references can be stored. The native code can then proceed to make calls to JNI functions which, in general, return local references. These local references are automatically stored in the stack frame; and the Java runtime takes care of cleaning them up when it pops the stack frame upon leaving the native method implementation.

So far, so good. But not all JNI function calls occur in response to Java calling a native method implementation. In fact, if you’re starting up the VM via JNI, you probably end up calling JNI functions that return local references just the same. Only in this case, the Java runtime hasn’t provided you with a stack frame that it will magically clean up once your code is done executing. Instead, you’ll need to delete the local references explicitly.

Thus we have two very different ways local references must be handled, depending on the context of the JNI function calls. And the inevitable problem: in a utility function which might be called from either context, how should intermediate local references be handled? Consider a function that creates a java.net.URL instance:

jobject create_url(JNIEnv & env, const char * const url)
{
    const jstring url_string = env.NewStringUTF(url);
    if (!url_string) {
        env.ExceptionClear();
        throw std::runtime_error("failed to construct string for URL");
    }

    const jclass url_class = env.FindClass("java/net/URL");
    if (!url_class) {
        env.ExceptionClear();
        throw std::runtime_error("could not find java.net.URL class");
    }

    const jmethodID ctor_id =
        env.GetMethodID(url_class, "", "(Ljava/lang/String;)V");
    if (!ctor_id) {
        env.ExceptionClear();
        throw std::runtime_error("failed to get "
                                 "java.net.URL.URL(java.lang.String) "
                                 "constructor");
    }

    const jobject url_obj =
        env.NewObject(url_class, ctor_id, url_string);
    if (!url_obj) {
        env.ExceptionClear();
        throw std::runtime_error("could not create java.net.URL "
                                 "instance");
    }

    return url_obj;
}

The above code will work just fine when it is called from a native method implementation. But if it is called outside that context, it will leak the local references corresponding to url_string and url_class. (We can assume the caller has responsibility for the local reference corresponding to url_obj in both cases.)

So, let’s toss in the code to delete the local references. We need to be exception-safe, so let’s use ScopeGuard:

jobject create_url(JNIEnv & env, const char * const url)
{
    const jstring url_string = env.NewStringUTF(url);
    if (!url_string) {
        env.ExceptionClear();
        throw std::runtime_error("failed to construct string for URL");
    }
    scope_guard url_string_guard =
        make_obj_guard(env, &JNIEnv::DeleteLocalRef, url_string);

    const jclass url_class = env.FindClass("java/net/URL");
    if (!url_class) {
        env.ExceptionClear();
        throw std::runtime_error("could not find java.net.URL class");
    }
    scope_guard url_class_guard =
        make_obj_guard(env, &JNIEnv::DeleteLocalRef, url_class);

    const jmethodID ctor_id =
        env.GetMethodID(url_class, "", "(Ljava/lang/String;)V");
    if (!ctor_id) {
        env.ExceptionClear();
        throw std::runtime_error("failed to get "
                                 "java.net.URL.URL(java.lang.String) "
                                 "constructor");
    }

    const jobject url_obj =
        env.NewObject(url_class, ctor_id, url_string);
    if (!url_obj) {
        env.ExceptionClear();
        throw std::runtime_error("could not create java.net.URL "
                                 "instance");
    }

    return url_obj;
}

There. Now we can call the function outside a native method implementation. But in making that work, we’ve made the function unusable from within a native method implementation. Clearly we need to make the calls to JNIEnv::DeleteLocalRef conditional; and we’ll have to let the caller tell us what context the function is being called in.

jobject create_url(JNIEnv & env,
                    const char * const url,
                    const bool delete_local_refs)
{
    const jstring url_string = env.NewStringUTF(url);
    if (!url_string) {
        env.ExceptionClear();
        throw std::runtime_error("failed to construct string for URL");
    }
    scope_guard url_string_guard =
        make_obj_guard(env, &JNIEnv::DeleteLocalRef, url_string);
    if (!delete_local_refs) { url_string_guard.dismiss(); }

    const jclass url_class = env.FindClass("java/net/URL");
    if (!url_class) {
        env.ExceptionClear();
        throw std::runtime_error("could not find java.net.URL class");
    }
    scope_guard url_class_guard =
        make_obj_guard(env, &JNIEnv::DeleteLocalRef, url_class);
    if (!delete_local_refs) { url_class_guard.dismiss(); }

    const jmethodID ctor_id =
        env.GetMethodID(url_class, "", "(Ljava/lang/String;)V");
    if (!ctor_id) {
        env.ExceptionClear();
        throw std::runtime_error("failed to get "
                                 "java.net.URL.URL(java.lang.String) "
                                 "constructor");
    }

    const jobject url_obj =
        env.NewObject(url_class, ctor_id, url_string);
    if (!url_obj) {
        env.ExceptionClear();
        throw std::runtime_error("could not create java.net.URL "
                                 "instance");
    }

    return url_obj;
}

Well, there we are. It’s exception-safe and the caller can tell it the Right Thing to do. But…It sure does Suck.

  • It relies on the caller telling it to do the right thing; which means it’s pretty easy to use wrong.
  • We wind up having to check delete_local_refs at each point where we might (or might not) need to call JNIEnv::DeleteLocalRefs. Our need for exception-safety prevents us from consolidating this logic near the end of the function. Consequently, the code is interspersed with yet more error handling logic that gets in the way of understanding the function’s primary logic when reading the code.

So, is there a better way? Fortunately, yes.

JNI provides functions that allow us to create (and destroy) stack frames for local references, JNIEnv::PushLocalFrame and JNIEnv::PopLocalFrame. So, instead of relying on a stack frame that may or may not be there depending on the calling context, we can just create our own, regardless.

jobject create_url(JNIEnv & env, const char * const url)
{
    if (env.PushLocalFrame(3) < 0) { return 0; }
    scope_guard local_frame_guard =
        make_obj_guard(env, &JNIEnv::PopLocalFrame, 0);

    const jstring url_string = env.NewStringUTF(url);
    if (!url_string) {
        env.ExceptionClear();
        throw std::runtime_error("failed to construct string for URL");
    }

    const jclass url_class = env.FindClass("java/net/URL");
    if (!url_class) {
        env.ExceptionClear();
        throw std::runtime_error("could not find java.net.URL class");
    }

    const jmethodID ctor_id =
        env.GetMethodID(url_class, "", "(Ljava/lang/String;)V");
    if (!ctor_id) {
        env.ExceptionClear();
        throw std::runtime_error("failed to get "
                                 "java.net.URL.URL(java.lang.String) "
                                 "constructor");
    }

    const jobject url_obj =
        env.NewObject(url_class, ctor_id, url_string);
    if (!url_obj) {
        env.ExceptionClear();
        throw std::runtime_error("could not create java.net.URL "
                                 "instance");
    }

    return url_obj;
}

Once again, we use ScopeGuard; only now we’re using it to call JNIEnv::PopLocalFrame. Note that we’ve provided space for three local references in our frame, corresponding to url_string, url_class, and url_obj.

But the above code is horribly broken: when we return from the function, the local reference associated with url_obj gets cleaned up with all the others! So, how do we get this reference out of our local frame and return it, unbound, to the caller?

The solution is a temporary global reference. We can convert url_obj to a global reference for the remaining duration of the local frame; and then convert it back to a local reference before returning to the caller. Once more, ScopeGuard is our friend:

jobject create_url(JNIEnv & env, const char * const url)
{
    using boost::ref;

    //
    // We can safely run DeleteGlobalRef in the scope guard because
    // calling DeleteGlobalRef with 0 is a no-op.
    //
    jobject result_global_ref = 0;
    scope_guard result_global_ref_guard =
        make_obj_guard(env,
                       &JNIEnv::DeleteGlobalRef,
                       ref(result_global_ref));
    {
        if (env.PushLocalFrame(3) < 0) { return 0; }
        scope_guard local_frame_guard =
            make_obj_guard(env, &JNIEnv::PopLocalFrame, 0);

        const jstring url_string = env.NewStringUTF(url);
        if (!url_string) {
            env.ExceptionClear();
            throw std::runtime_error("failed to construct string for URL");
        }

        const jclass url_class = env.FindClass("java/net/URL");
        if (!url_class) {
            env.ExceptionClear();
            throw std::runtime_error("could not find java.net.URL class");
        }

        const jmethodID ctor_id =
            env.GetMethodID(url_class, "", "(Ljava/lang/String;)V");
        if (!ctor_id) {
            env.ExceptionClear();
            throw std::runtime_error("failed to get "
                                     "java.net.URL.URL(java.lang.String) "
                                     "constructor");
        }

        const jobject url_obj =
            env.NewObject(url_class, ctor_id, url_string);
        if (!url_obj) {
            env.ExceptionClear();
            throw std::runtime_error("could not create java.net.URL "
                                     "instance");
        }

        //
        // Create a global reference so that the new object will outlive
        // the local frame.
        //
        result_global_ref = env.NewGlobalRef(url_obj);
        if (!result_global_ref) { return 0; }
    }

    //
    // NewLocalRef does not throw any Java exceptions.
    //
    const jobject result = env.NewLocalRef(result_global_ref);
    if (!result) { throw std::bad_alloc(); }

    return result;
}

Ta da. We introduce an additional block scope for the duration of our local stack frame. We propagate the reference to the URL class instance out of the frame’s lifetime by creating an additional global reference. Once the frame is destroyed, this global reference is all that’s left. We then create a new local reference to return to the caller, and let a scope guard take care of cleaning up the global reference.

We've got Spirit

I have spent a large chunk of my free time over the last several months replacing OpenVRML’s ANTLR parsers with ones using Spirit. This was my first attempt to do something nontrivial with Spirit. So this comment from the author of Spirit absolutely made my day.

New Autoconf OpenGL macros

After nearly a year since pushing this project to Google’s project hosting, I’ve finally made a release tarball of my Autoconf macros for OpenGL. This was motivated mostly by significant changes that were precipitated by requirements for building on Mac OS X 10.5.

10.5 is a bit of a mixed bag relative to its predecessor as far as building with OpenGL is concerned. The good: they finally fixed the longstanding bug with the GLU tesselator callback function type being interpreted incorrectly. The bad: linking with X11 now requires some extremely goofy linker flags.

One nice thing, though, is that I was finally able to get rid of the annoying ‑‑with‑apple‑opengl‑framework option.

1997 called…

I have begun to wonder just what it would take to create an ActiveX control for OpenVRML.

It has been about a decade since I touched COM. At the time, I had just (barely) learned C++. And I was struck then that just about every Good Practice for C++ I was learning about was being flagrantly violated in COM.

I quickly lost an affinity for Windows programming for that and many other reasons. I’ve done development primarily on POSIXy platforms; but much of that has been sufficiently portable to Windows that I’ve maintained a superficial knowledge of Windows development. But I’ve done enough development and I have a sufficiently deep understanding of C++ that something like COM just isn’t intimidating anymore. Which is not to suggest that I relish picking it up; but, rather, my annoyance at it is overcome by both a desire to show something working and a morbid curiosity about Just How Bad It Is.

My first thought was to apply the “modern” technique for building COM components, ATL. ATL is a template library; though the Wikipedia entry that (as of this writing) claims that ATL was patterned after the C++ STL is quite far off the mark. Aside from being written in C++ and using templates, there isn’t much similarity between the two. For the most part, ATL seems to be a bunch of CRTP mixin class templates. I can see how I would value the legwork it saves me if I had 10 or 20 or 100 COM components to write. But as I have just one, the obfuscation implicit in ATL usage seems unlikely to pay off.

And it’s just as well. Apparently Microsoft no longer includes ATL with the Windows, née Platform, SDK. So if I want this thing to build with the Express version of the compiler, no ATL.

It’s a bit too early to say if this effort will actually pay off in terms of a usable OpenVRML ActiveX control. It could happen. It could also be the case that I become so put off with the API that I never consider doing this again.

Nokia N75

I’ve used Motorola mobile phones since I’ve had a mobile phone, starting with the V60i back in 2002 (as a customer of the original AT&T Wireless), followed with a V551 (with Cingular), which was in turn replaced with a RAZR V3 (upon its untimely demise).

The latter two phones had nearly identical software; but I was (and still am) a sucker for the RAZR’s form factor and construction.

  • I keep my phone in my pocket, so I like thin.
  • Antenna nubs are annoying. They’re pretty much a thing of the past on GSM phones; but when the RAZR was introduced, most clamshell phones had them.
  • There’s something satisfying about a cool-to-the-touch metal case.
  • The RAZR’s construction is solid. The hinge mechanism has a nice stiff feel to it with very little play. Mine has held up well through a couple of years of use.

The RAZR is not without its flaws, though. My chief annoyance was the weak phone book. This weakness is by no means unique to the RAZR. Most Motorola phones share a lot of the same software; and the phone book appears basically unchanged on models as recent as the KRZR. The number one, what-the-hell-were-they-thinking problem with the this phone book has to be: the single name field; i.e., no means of distinguishing between first and last names. So if you want to sort your phone book by last names, you need to put the last name first in each field. And then when you go to use synchronization software, you’re generally screwed because it doesn’t know about the scheme you’ve superimposed. A lot could be forgiven if this one problem were fixed. But there is plenty more wrong:

  • Rather than let you attach multiple phone numbers to a single entry, the Motorola phone book instead makes you create multiple entries with the same name, then gives you the option to visually merge the identically-named entries and treat one as primary. While this mostly works, it can at times be cumbersome to use.
  • That single name field is, on the RAZR at least, way too short.
  • The phone book accommodates storing phone number and e-mail addresses—nothing else.

So while I have preferred the Motorola clamshell phones, my wife has preferred Nokia candybar format phones. No question about it, Nokia’s UI is consistently superior—particularly its phone book. But there’s something about the ergonomics of opening a clamshell phone to answer it that I’ve always found attractive. And until relatively recently, Nokia’s suite of clamshell phones available in the US was quite weak.

That’s changed lately, which brings me to the N75. This phone uses a more recent version of the S60 software driving Gina’s 6682, so I was reasonably sure I’d be satisfied with the phone on this account. Indeed, the phone sports an extremely capable address book. It knows about first names and last names. It also knows about entries for which there should just be a company name. It can even store physical addresses—a feature I’ve been wishing for in a phone for quite some time.

The N75 has fared so-so in reviews. The major strikes against it seem to be that it is a battery hog and that it takes merely so-so pictures. My demands of a phone-bound camera are quite low; and coming from the RAZR’s VGA camera, even a mediocre 2 megapixel camera is a significant upgrade. I haven’t tried the camera much; but picture quality issues aside, I am quite pleased with the traditional camera-style ergonomics for the camera controls. Camera lenses on some phones seem to be designed to be obscured by a finger; that’s no problem here, even with my large paws. (The nice large keypad on this phone is also very accommodating of large fingers.)

As for the battery issue, I seem to be getting about two days on a charge. That may improve in the future, once I’m opening the phone to edit the address book less.

The construction of this phone feels reasonably solid; but it’s not up to the level of the RAZR in this regard. Nokia could learn a thing or two from Motorola with regard to hinge construction. Mind you, the N75 certainly isn’t bad in this regard. But the opening mechanism isn’t even as satisfying as my old V551; and it’s really nowhere near the nice feel of the RAZR. Apart from the hinge area, the body of the N75 is a matte black plastic. It seems likely to hide scratches reasonably well. But it doesn’t have the nice cool-to-the-touch feel of the RAZR’s metal case. Closed, the phone has almost the same footprint as the RAZR; but it’s about 30% thicker. While not terribly thick, the N75 isn’t really a “thin” phone by modern standards.

All in all, I’m pretty satisfied with the N75.

Return top