Older blog entries for lkcl (starting at number 670)

lamson list archive viewer started

Eating Dogfood, one Web app at a time: Pyjamas List Archive

well, i took the Mail example and finally did something with it that i've been meaning to do for a while: actually got it to really read real email messages. the other application which does that is http://beammail.org which thomas did, a year ago, and i must update it. thomas actually did send, login and even read contacts and TODO information, so it's a worthwhile exercise to update it.

but the List Archive reader is slightly different. firstly, i decided to read the pre-processed json data, not the Maildir stuff. this will reduce the load on my happy server. if lamson has already gone to the trouble of splitting out the headers, splitting out the attachments and so on, why do it again?!


and the code is at:

    git clone gitolite@pyjs.org:pyjamaslamson

Syndicated 2010-07-02 12:01:39 from blog/lkcl

working pyjamas git jsonrpc service

Up and Running!


amazing - a JSONRPC-based git web service, along with a happily bare but functional front-end. it's… wow! i kinda love it. it doesn't look like much, but it's actually one of the first real functional development tools i've done with pyjamas, other than "demos".

a mailing list archive viewer is next, then i can drop groups.google.com.

if anyone wants the source code, it's at

   git clone gitolite@pyjs.org:pyjamasgitweb

Syndicated 2010-06-29 22:27:41 from blog/lkcl

jsonrpc and pyjamas git web service

New Project: Pyjamas Git Web Server

git clone gitolite@pyjs.org:pyjamasgitweb

of course, it has to be a jsonrpc service: i'm going for a service based on the python XMLRPCServer at the moment, just for kicks, but will be setting it up as a WSGI service soon enough.

the pyjamas front-end of course had to be done - there's really no point in having a web site http://pyjs.org and then installing perl (except for joey's ikiwiki) or php or even a wickedly-fast cgit web service, because they all serve HTML. and the point of the pyjamas project is to duh make every application used in the development of pyjamas work as both a desktop and a web app. so, it could be the best, the most lovely, the most fastest most stonkingly brilliant development tool on the planet, loved by everyone from linus to redmond: if it ain't a pyjamas app, it ain't gonna get onto http://pyjs.org. except joey's ikiwiki :)

ultimately the plan is to put in place a text editor (even if it's a TextArea for now) so that the developers themselves can entirely use pyjamas tools - editing *in* the web browser (or desktop app). that'll be fun, because personally i'll need to recreate vim, including syntax colour highlighting, in python, before i'll use it. deep joy.

and the neat thing about this is that it's not just useful for pyjamas: it could be used by anyone, on any project.

Syndicated 2010-06-28 17:35:30 from blog/lkcl


redoing the pyjamas website

i want the pyjamas web site to be a practical showcase, managing a project using its own technology. i don't want pyjamas hosted on sourceforge, groups.google.com and code.google.com or anywhere which doesn't actually use pyjamas and python. so, what do i do? install ikiwiki :)

but seriously: the plan is to create a pyjamas mailing list archive reader and run python-only mailing list software; create a pyjamas wiki editor and (if convenient!) run off the back of ikiwiki even if it means just using the git repository and the ikiwiki commit hooks; write a bugtracker etc. maybe even a git web tree and code browser in pyjamas, that would be truly horrendous.

Syndicated 2010-06-28 01:18:16 from blog/lkcl

FontForge and Python

well, i've been asked to create some demonstrations of the use of python with fontforge - not just "little scripts" but entire dialog boxes which will do the same job as the existing menu options. ultimately, the goal is to replace every single menu option with an equivalent. ok - the goal is for me to prove that even one dialog can be replaced, such that other people can realise it's possible. and, importantly, whilst replacements for the menu options are being developed, font developers can continue to use either the replacements or the existing code.

why do this at all?

well, fontforge's widget set was developed from scratch at a time when widget sets didn't handle unicode (which fontforge has to) and many more things which are listed on the fontforge FAQ. however, as time has progressed, _some_ features of widget sets - not all - have caught up. for example, the FF FAQ states that GTK can only do one type of RGB bitmap but FF requires several, and so on.

but, some of the fontforge input boxes are... archaic: for example there is no mouse-wheel-driven input widget, so font developers are forced to type in numbers (which often involves them looking down at the keyboard, not at the screen).

to replace such widgets and to generally make any kind of UI changes is... tiresome, as the code - including its widget set - is entirely written in c.

so, to make the job easier, the experiment is to implement an example menu option in several different python-based widget sets, and see which one flies. personally, i like pyjamas, not least because - believe it or not - its use means that fontforge could turn into a client-server application (!) using JSONRPC to get at the data, and the front-end GUI could be run in a modern web browser.

but, to be a proper experiment, several widget sets will be tried. so far i have a pygtk / cairo experiment up and running, and also a pyjamas (gwtcanvas) one. all really quite exciting. ultimately, it is hoped that it will be possible to use fontforge-nox with python bindings, and drop the main fontforge window, replace it with pygtk / pyjamas / pyqt4 / pyclutter and... done!

Glade UI Generator for Pyjamas

yaay! kees bos has started a User-Interface generator, based on the glade concept for GTK, for pyjamas. called PyJSGlade.

in essence, it's very simple: an editor, allows addition of panels and widgets and the editing of their properties, save the layout as an XML file, and then a second stage once you're done throws out a .py file which represents the application.

pretty straightforward, and very very cool.

what will be particularly interesting at some stage will be how to include HTML files and link widgets to existing bits of DOM. that's gonna be fun and games, and something that GTK-Glade _can't_ do.


well, surprisingly, the pyjamas-dev list is getting very active, all of a sudden, with what... five or six ongoing conversations a day. it's kinda nice to see that there's signs of the project taking off. i've got an upcoming article and also europython 2010 so it might get a little more hectic, which would be nice. web site needs a redesign though...

26 May 2010 (updated 26 May 2010 at 16:13 UTC) »
CT-PC89E or ST-PC89E as it's now known

hooray! we have the schematics of the 200-pin SO-DIMM and the netbook's motherboard! exactly how my friend adam managed to get these i have no idea, but it is very very good news. it's now possible, for example, to create an upgraded 1ghz ARM board (e.g. with an S5PC110 or the very expensive OMAP3530) for what is otherwise a slightly "lacking" machine.

the problem with the current S3C6410 system is that the software developers (mid-fun.com) are GPL violators. seatron, the hardware developers, are having difficulties convincing mid-fun to release the source code, and the hardware design is intimately tied with the [gpl violated] linux kernel.

so, by having the schematics, many of those gpl violations problems go away; additionally, by making a replacement 1ghz ARM SO-DIMM board the linux kernel GPL violations _also_ go away because we would have full knowledge of the design...

25 May 2010 (updated 25 May 2010 at 17:22 UTC) »
apathy and its antidote

well, i went down to the guildford library, took photos of the 50 or so pages of the 1948 statute reform act... over two weeks ago. the photos are still, unexamined, on my SD card.

in other words: out of sheer apathy, i'd left the Treason investigation behind me.

so receiving a stalinist "invitation" to attend an "assessment" was like a 50kV taser shot to the nuts, throwing me into _completely_ the opposite direction intended by the fascist/ignorant/stalinist decision made by some as yet unknown person in the Surrey Police Force.

i'm meeting with my MP on friday. i've written a letter to the Queen, explaining my outrage and concern that i may have to seek asylum outside of the EU. i've made a complaint to the IPCC and will be asking my MP to contact them on my behalf, in order to add teeth to the enquiry. a friend of mine has contacted the slightly unnerving UK Column on my behalf, and someone from UK Column just _happens_ to be visiting and lobbying members of the House of Lords today...

all in all the _complete_ opposite effect intended by the stalinistic approach, which is presumably intended to either call into question the credibility of the person reporting the crimes, or to scare them off.

as people who actually bother to do their research about me tend to know, i don't take too kindly to being threatened by "powerful" people: it tends to piss me off and i start looking for metaphorical two by fours to go metaphorically smacking them about the head with.

question: has anyone heard of the "Fixation Unit"? it's apparently a unit that is tasked with referring citizens who report crimes committed by politicians to "mental health assessments".

such delicious irony: my apathy was overcome by the accusation of "fixation". who's _actually_ "fixated" - me or this fucking "Fixation Unit"?

22 May 2010 (updated 22 May 2010 at 13:11 UTC) »
Contacting the Independent Press Complaints Commission

I researched and reported two crimes of Treason at my local police station; references PNNNNNNN and PNNNNNNNN.

today i was shocked to receive a letter from Spelthorne Mental Health Community Team "inviting" me to a "mental health assessment".

to be absolutely honest scared shitless of the implications and extremely angered by what are effectively accusations by Police Officers that i am "mentally unwell", and of the implied breakdown of trust by those Police Officers passing on my name and address without my consent.

i reported these crimes in good faith.

if i had known that i would be "accused of being mentally unwell" then do you _seriously_ think that i would have reported the crimes? what are the implications there: do you think that people will report serious crimes if they are at risk from being "locked away as mentally ill"?

i am now extremely concerned for my safety, as a direct consequence of the actions taken by Police Officers. god knows what these idiot Social Workers are going to do, now. my partner marie is here on a 2-year Visa, and we have a 14 month old daughter. is the Home Office going to have access to this information through the NHS "Spine", without my consent, and "refuse to issue a new visa" when we come to renew it? if Marie is deported, will I ever see my daughter again?

so you can see - the decision by the Police Officers to pursue "mental health allegations and accusations" against me has serious, serious implications and consequences.

one thought occurred to me which should shock you into investigating and taking this seriously: there is an incredible similarity between what is being done to me and what Stalin used to do to his "political enemies". whilst Stalin himself was sufficiently mentally ill to "invent" enemies and get people executed (which is not going to happen in 21st Century England), many intelligent people in Russia in the 50s were simply too afraid to talk openly, for fear of accusations of "being mentally ill".

my only form of defense against this kind of attack is to make everything that i say public. that includes this report, and it includes contacting newspapers.

Thank you, your complaint has been received by the IPCC and will be allocated to a Casework Manager. We try to respond to complaints as soon as possible, however the allocation time does depend on the number of complaints received. In the meantime if you have any queries please contact us, quoting your reference number on : 08453 002 002 Kind regards IPCC

661 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!