The beginning of this week followed a different plan:

• Practice for driver's test, which is on Monday (7th).

I passed! I know most of you will find this weird because it seems like I entered the game nearly 5 years late, but you really don't need a car (or to know how to drive) when you live in New York City. With a gigantic network of subways, b-u-s-e-s-s, and even a railroad, many teenagers here have actually preferred to not drive. And my lack of driving skills haven't come back to haunt me in college because thankfully there's also a bus system there.

The last two weeks have been filled with potpourri. As my day job ended, hell ensued, which was coupled with setting down project continuity arrangements. (During that period, I somehow got myself sick too, making matters even more difficult.)

I started working the log_reader plugin, which would transparently read (for viewing) logs created by other instant messaging programs. The scope of this second project is to finalize MSN pre-Live and Trillian reading, and possibly also MSN Live.

I also filed a request to purl.org to create the /ns domain, for use with XML namespaces. Either it got declined or it never went through.

This week was one of the most difficult weeks thus far.

I completely rewrote the unified logging specification and then the schema. I had writer's block about 75% of the time. I even recruited a friend to help proofread the spec.

In the end, it was a very fun experience, but I really need to get better at this.

Attached are the specification as of now. Since the schema can't be attached, you'll have to use the link that will bring you to the latest available version. Of course, 0.4-02 is already in the works as this is being written. =)

I haven't been keeping tabs about what I've been doing for the last two weeks. Things are slowing down because I'm actually getting quite fatigued for no apparent reason. I blame it from weird the feedback loop that I have. See, people tend to sleep more if they're very tired. Unfortunately, more sleep leads to being even more tired for me, giving me less overall energy to work with.

It's a problem I've tried to remedy previously in many different ways but I haven't had that much luck.

For the week of July 2nd, I cleaned up the code that I had added and changed everything to use GIOChannels.

This week I'm going to concentrate on:

• generalizing the event reporting system in Gaim. It'll consist of using an enum of supported events types, and a fallback to create events that are not specifically listed by Gaim.
• continuing writing the specification of course. I'm now following the XHTML2 working draft because it notes current issues and has elements/attributes/values tables.

On the side, I'd like to look into writing another set of gaim_log_common_writers that ensures a well-formed XML version of the log is outside of user space (it'll use Glib I

You mustn't send the extra blank char in the HTTP or HTTPS header or MSN protocol.
or else you will encounter a very strange problem!
I will try to illustrate why in protocol description

A reminder for those developer, especially for me,
I have spent about 1 weeks for that useless debugging:(

Have been on MSNP13 Protocol Validation Process for about two weeks
Because MSN don't give the official MSN protocol leashed,what I should do is to unleash and validate the unofficial protocols' corectness
I use the Following website for reference
http://msnpiki.msnfanatic.com/index.php/MSN_Protocol_Version_13
and
http://www.hypothetic.org/docs/msn/index.php

My Problem is:
follow the login and validating procedure,I can login the server and resolve the contact list.
But other MSN contact can't see my client's status as online.
So I can't continue the development,such as chat with others, Transfer file,...

This past week has been very hectic for me. Of the five days this week, I had to stay 4 hours late for my day job 3 times. For those who wonder, we're paid by the week and not by the hour so there's no overtime.

I've been finding myself concentrating more and more on the side projects instead of my actual goals. That might be a problem. I've yet to contact my mentor to discuss whether or not it's the goals I meet or the number of things I get accomplished that will matter in the end. This is evident in this week's list.

• xml log writing now always flushes (from user-space) a well-formed XML file. After having a discussion in #gaim about using glib IO, I decided there was too much sentiment against introducing glib IO that can threaten my code being merged back into the tree at the end of SoC.

I've decided to outline my short-term goals for what I need to do. A friend asked me about it online, and I think it was a pretty clear breakdown of what I'm trying to achieve (code-wise). Before I begin, I'll describe it on a higher-level.

One of the overall goals is to allow multiple loggers to log in Gaim. Currently, Gaim keeps a single logger identifier selected, and uses a struct containing function pointers to corresponding logging activites - opening, writing, close, retriving. I will assume logging both conversations and system events (buddy sign ons, going aways, etc) are enabled. When a new logging entity is created (a conversation window opens), a GaimLog structure becomes associated with the entity until its destruction (the conversation window closes). Instead of having a single static reference to a logger for the whole duration of its life, I can either (1) let the association happen dynamically - when a logging activity is called, I'll check the prevailing setting, thus allowing users to change logging format mid-conversation - or (2) still have the reference be dynamic, but have it be a terminated linked list of these GaimLogLogger structures. Gaim would simply traverse these all these various functional pointers, and since they're all sane (e.g., if you try to write to a non-existent logfile, the file will be created), this is totally possible.

I'm typing this report earlier because I've completed a milestone yesterday. So...

• The functions in the log viewer had lots of calls to functions that return static data, and other functions that really relied on the timestamped format of the log filenames. It just took some time to pinpoint which areas of the code were operating on static data (I learned that a return type of const char * usually means that it returns static data), and bandaging up sections of code that parse the filenames to accept both old a new formats.
• It took me three days to get it, but I wrote a unified logging format XML transformer to gaim XHTML for viewing, so now xml_logger has a working read() function. Messages are totally supported. Events need work; I'm either going to change to the events system so I can log the metadata in XML or do lots of sscanf's.