Older blog entries for movement (starting at number 83)

Just made the first faltering steps towards proper call-graph support in OProfile via gprof-format export.

Seems to work nicely :

               12.97   73.70       1/58          common_interrupt <cycle 1> [81]
               77.79  442.20       6/58          kmem_cache_free [9]
              661.24 3758.69      51/58          _pte_chain_free [12]
[10]     4.6  752.00 4274.59      58         page_remove_rmap [10]
             1596.54 1578.45     594/679         zap_pte_range [14]
              973.46  112.77     189/1594        do_wp_page [3]
                8.73    3.54      17/2542        zap_pmd_range [27]
                0.28    0.49       8/16796       handle_mm_fault [31]
                0.00    0.32       2/2           copy_one_pte [802]
Now I just need to do lots of cleanup, refactoring, analysis etc. But as it started off as basically completely experimental, I'm quite pleased the results are so sane from first impressions.
daniels> compressing the output down from the actual commands to "linking lib/kotext/bar.la"

Please please don't. By all means make it an option, it's very useful. But hiding what is actually happening by default is extremely ungracious. I've personally lost at least three hours of my life due to similar nastiness in the HotSpot build.

Sticky Minds review freeware testing tools, and find them to be "wild and woolly. There are few free tools that rival their commercial counterparts in terms of features, reliability, and support.".

I wonder if that tallies with the Advogato readership's experience.

21 May 2003 (updated 21 May 2003 at 01:17 UTC) »
dhess: I'm idly wondering how OProfile's template-based approach to command line parsing stacks up against those options. It's based on top of popt and was mostly written by phe.

You simply define an array of options like :

bool verbose;
vector<string> image_path;
double threshold;

popt::option options_array[] = { popt::option(verbose, "verbose", 'V', "verbose output"), popt::option(image_path, "image-path", 'p', "comma-separated path to search missing binaries","path"), popt::option(threshold, "threshold", 't', "minimum percentage needed to produce output", "percent"), };

It doesn't need to be an array, and you can easily add other options in source files. The above match templates which construct the correct values in the table used by popt. When you init you call parse_options(argc, argv, other); and the variables you passed in above are filled in.

It's a little bit rough around the edges but it seems to work pretty nicely. It handles --help and takes GNU and short options as appropriate (see man popt), and handles comma-separated options (with "\,"), which the similar yagol doesn't seem to.

Ankh: "parent directory naked" is certainly worthy of Disturbing Search Requests, I think.

This is probably old news to everybody except me, but someone pointed out something really cool.

23 Apr 2003 (updated 23 Apr 2003 at 18:07 UTC) »
apenwarr: whilst I agree with your underlying point, I'm afraid you picked a bad example: users certainly do care about pipelined HTTP, seeing as that is something that actually makes users feel that pages "load faster", as you mention.

Of course, they have absolutely no idea what it is, but that is beside the point. Similarly, I would suspect the Sherlock support integrates better on Mac OS, etc.

I can imagine normal users like being able to search through bookmarks (not that it works very well).

But yes, that list is pretty symptomatic of the old Mozilla preference for endless bloat and geek features. Galeon died a similar death. There is at least some hope that Phoebird won't repeat the same mistakes.

Addendum: Chandler

So the OSAF released the first package of their nascent PIM, Chandler. It's still obviously in a very primitive state, but I made some screenshots anyway.

Worryingly, File->Quit is already present.

Alphabet Soup is an interesting program that understands the shape of alphabetic glyphs, and can modify the shapes to produce plausible variations.

Random internet searches introduced me to the game of Nomic, a game in which changing the rules is a move. Unsurprisingly, it gets complicated quick.

Also unsuprisingly, the xfree86 "forum" mailing list descended into content poverty exceedingly quickly. Some people just love to be procmailed I suppose.

Does anyone know of a search engine that can handle symbols correctly ? Searching for, say, X$X on Google is not a pleasant experience.

In my continuing mission to trivialise the war, here are links to We Love the Iraqi Information Minister and The Poetry of D. H. Rumsfeld.

What is Mellel ?

"Mellel incorporates all the sophisticated options you would expect to find in an advanced word processor, and adds many innovations you cannot not find anywhere else."

Apparently.

They claim 10 unique features, but I can't for the life of me work out what they mean by "Indent tab and Combo tab". If it's as revolutionary as they claim, I wish they'd describe it better, so we can steal it. The fact they make a big deal about having relative sizes in their character styles suggests a little of the snake oil.

74 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!