I needed to find and operate on the most recently modified file in a directory, but from the Windows XP command line rather than a Linux shell script. Knowing how unexpectedly powerful (but so very arbitrary and convoluted!) the Windows command line can be, I figured it was possible with a bit of help file reading.

And lo, it was good:

 @echo off  :: This temp batch file is used to execute a subroutine. set TMP_RECENT_BAT=%TEMP%\tmp_recent.bat  :: Reset the "output" env var in case something goes wrong set MOST_RECENT_FILE=  :: Create the temp batch file. :: :: This subroutine lists all files in the current directory, :: sorts it in descending last-modified-time order, :: and strips out the topmost file (i.e. most recently modified). :: :: Directories, system files, and hidden files are not considered. :: :: A side-effect of this routine is that the filename has a prefix of "[1]" echo @echo off > %TMP_RECENT_BAT% echo dir /o-d /a-d-h-s /b %1 ^| find ^"^" /v /n ^| find ^"[1]^" >> %TMP_RECENT_BAT%  :: Execute the subroutine above, and grab the output in an environment var. :: We are assuming the subroutine above correctly returns only one line. for /f "usebackq delims=="%%i in (`%TMP_RECENT_BAT%`) do set TMP_RECENT=%%i  :: Strip the first 3 chars of the return value of the subroutine, i.e. "[1]" :: :: If the subroutine didn't return a good value, don't try to set the var, :: b/c we'll get a string like "~3" instead of a good value. if "%TMP_RECENT%" == "" goto skip  set MOST_RECENT_FILE=%TMP_RECENT:~3%  :: This line not really necessary, since script that use this can just access :: the environment variable named "MOST_RECENT_FILE". echo %MOST_RECENT_FILE%  :skip set TMP_RECENT=  :: Delete the temp file del %TMP_RECENT_BAT% 

I visit my blog this morning for some admiration and/or self-critcism, and I discovered that I had votes for my blog entries! Yay, people read and respond to my blog! Except… they all voted “No”, that the entries were not interesting or helpful… for several entries.

It turns out the voting plugin I made for my blog had a major oversight: search engine spiders click on them, too. Oops. And for some reason I can’t fathom, they all prefer the “No” vote.

So there you have it, even software thinks my blog is dull. Normally, I’m a fan of listening to people express their opinion, especially if it runs contrary to mine, but I think this time I’ll have to do something about it.

I decided to remove the calendar from my blog’s sidebar, and replace it with a “recent posts” list. I realized that it’s not very helpful, since readers aren’t so much interested in when I posted as what I posted about.

There was no pyblosxom plugin to do this, so I rolled up my sleeves and learned some more Python, referring to the Pyblosxom API documentation along the way. The resultant plugin is simple, although I had to fill in the gaps in the API docs through trial and error.

The code I wrote is below. Note that I hardcoded the flavour type (1024px), but I believe there’s an value in the entry dictionary I could have gotten it from.

 from Pyblosxom import tools from Pyblosxom import pyblosxom import time, re 

 def cb_prepare(args):   request = args["request"]   data = request.getData()   entry_list = data["entry_list"]   config = request.getConfiguration()   root = config["datadir"]   s=""     txt_dict = {}   for txt in tools.Walk(request, root, pattern=re.compile('.*\.txt$')):     timetuple = tools.filestat(request, txt)     txt_dict[time.strftime("%y%m%d%H%M",timetuple)]=txt    keys = txt_dict.keys()   keys.sort()   keys.reverse()   url_list = []   for k in keys:     path = re.sub(root, "", re.sub(".txt$", "", txt_dict[k]))     title = pyblosxom.blosxom_entry_parser(txt_dict[k],request)['title']     url_list.append("<li><a href='" + path + ".1024px'>" + title + "</a>")    url_list = [url_list[x] for x in range(0,19)]   for url in url_list:     s += url    data['recentPosts'] = s 

Around the same time, I had the idea of adding a “voting” mechanism to allow people who don’t have anything to say (comment) to still indicate which entries are interesting. I figured an easy approach would be to add voting buttons that triggered an AJAX-style HTTP request to special URLs, which would show up in the Apache logs and could later be parsed out. Being unfamiliar with Python, I could just focus on reading data in the plugin and save on code to tally votes (I was feeling lazy at the time).

Shortly thereafter, I got a blog comment:

Name: Gunfus
Entry URL: http://blog.astradele.com/life/Why-blog-.2008-02-01-14-50

I don’t think your yes/no is working to well. If you click it it dissapears (this is fine), if you reload the page it shows again (this is aceptable), once you reload, the counter is at 0/0 (this is expected?)

I thought I tested it properly, but… So I got into it again and wrote the solution completely in Python. At least I’m getting more familiar with the language. The solution still uses a custom URL format that don’t resolve to an actual page, but it’s the Pyblosxom plugin that notes the URI, checks to see if it’s similiar to an existing entry, creates files to track the IP addresses of votes on a per entry basis, and generates tallies.

I considered trying to devise a way to auto-refresh the tally instead of just hiding the voting “bar”, but upon consideration I wasn’t up to the effort it would take on the Python side.

The code:

 import sys, StringIO, os, os.path, re 

 def verify_installation(request):     return 1  def include(req, filename):     """     This takes in one argument--a filename.  It opens the file, reads     in the contents, and returns them as the value of this variable.      (It actually takes in two, but one of them is provided by     PyBlosxom.)     """     if not filename.startswith(os.sep):         data = req.getData()         basedir = data["root_datadir"]         if data.has_key("bl_type") and data["bl_type"] == "file":             basedir = os.path.dirname(basedir)          filename = os.path.normpath(basedir + os.sep + filename)      try:         f = open(filename, "r")         lines = "".join(f.readlines())         f.close()         return eval_python_blocks(req, lines)     except:         return ""  def cb_logrequest(args):   import os, time   http = args['request'].getHttp()   config = args['request'].getConfiguration()   request_uri = http['REQUEST_URI']   remote_addr = http['REMOTE_ADDR']   entry_name = re.compile("^/vote_(yes|no)(/.*)\..*$").sub(r"\2",request_uri)   if (os.path.exists(config['datadir'] + entry_name + ".txt")):     try:       os.makedirs(config['datadir'] +           "/votes" + re.compile("^(.*)/.*$").sub(r"\1", entry_name))     except:       pass      if (re.compile("^/vote_yes").match(request_uri)):       vote_file=config['datadir'] + "/votes" + entry_name + ".yes"     else:       vote_file=config['datadir'] + "/votes" + entry_name + ".no"      is_dupe=False     if (os.path.exists(vote_file)):       f=open(vote_file, "r")       for line in f:         is_dupe = is_dupe or re.compile(remote_addr).match(line)       f.close()      if (not is_dupe):       f=open(vote_file, "a")       f.write(remote_addr + "\n")       f.close()  def cb_story(args):   """   """   entry = args["entry"]   request = args["request"]   config = request.getConfiguration()   http = request.getHttp()    entry_filename = entry.getId()   yes_votes_file = re.compile(".txt$").sub(".yes",re.compile("^" + config['datadir']).sub(config['datadir'] + "/votes",entry_filename))   no_votes_file = re.compile(".txt$").sub(".no",re.compile("^" + config['datadir']).sub(config['datadir'] + "/votes",entry_filename))    entry['vote_yes_url'] = "/vote_yes/" + entry['file_path'] + "." + entry['flavour']   entry['vote_no_url'] = "/vote_no/" + entry['file_path'] + "." + entry['flavour']    request_uri = request.getHttp()["REQUEST_URI"]   remote_addr = request.getHttp()["REMOTE_ADDR"]    count=0   if (os.path.exists(yes_votes_file)):     f = open(yes_votes_file, "r")     for line in f:       count += 1     f.close()   entry["yes_votes_count"]=str(count)    count=0   if (os.path.exists(no_votes_file)):     f = open(no_votes_file, "r")     for line in f:       count += 1     f.close()   entry["no_votes_count"]=str(count)