• Added SaveTimestamp (Thanks to Markus Fix)

• Added UpcaseNewTags

• Added SkipThreshold Option (Thanks to Clint for the patch)

• Added Quiet option and added .pl file to SkipCGI (Thanks to Dan Heller for the patch)

• Support for MNG images (multiple image PNG's) added (Thanks to Scott Hoge for the patch)
• added a SkipFilter
• added second jpegsize method

• made the HTTP request be HTTP/1.1 with help from Robert Hood

• made it recognize the makebackup option
• added text to say if no values returned

• added Magnus' skip 1x1 image
• added text to say if tags inserted or updated

• Fixed seek problems in newgifsize - you can't seek with a socket (Duh!)
• Also fixed output record sep when it's not defined for File/Copy.pm

• replaced system(cp) with copy from File::copy; (Thanks HansS@Fenestrae.com)
• Fixes suggested by Davin Milun (milun@cs.Buffalo.EDU)
  • added fix for chmod check (Whoops!)
  • UsePerlCp - option for using perl cp or system cp (with -p flag)
  • IgnoreLinks - option to ignore symbolic links - This is default on

• This is now only for PERL 5 if you don't have it, get it from http://www.perl.org/
• Made it use strict and Socket
• Updated Socket code to get rid of not an number error
• fixed bug with mungper and a %
• Small change to the JPEG code

• added Greg's byte sex assumption fix to the png code.
• hacked perl 4 perl 5 compatability code to use eval. (I'm going to give up on perl4 soon)

• added if statement to work out which version of perl you're using. It's not right, but it's an indication of what I want to do
• added binmode to imgsize for PC - although I don't support PC's (unless it's running Linux :-)
• added it so that if you have ChangeIfThere set to no it woudln't even find the size of the file - make sense to me
• turned off newgifsize by default
• got rid of > in NeedsAlt regexp

• implemented hash table for looked up values.
• fixed clever way of calling perl.
• fixed way that config file is selected - due to change again soon..

• testing my new version printout -version command line argument
• Added usage C.L.O.
• added command line parsing so you can set any value
• file is only written if a change was made
• fixed change if there ask kind of code - it was a mess
• Added the following options
  • bool type to options array.
  • Proxies
  • Skip CGI
  • Script, InFilter and OutFilter as well as script
  • Makebackup, but didn't implement it yet.

• added version to top of file. Doh!

• got rid of all global variables - we'll apart from the option list
• it now runs with -w with no errors - no mean feat :-)
• make the image size getting routines return a 2 element array with either 0,0 or the width & height of the image
• added sys/socket.ph to the start so should work on solaris machines
• made stdout unbuffered so we can write bits at a time on the same line
• change the way the command line is used to support command line arguments in future
• now I only look at files with a length which look like text files
• added the following options (may still be unimplemented, but you can set them!)
  • UpcaseTags * upcase all HTML tags
  • TryServer * if you can't find the file try the server? needs base tag
  • QuoteNums * quote the numbers when you insert them
  • Munge% - if a % is there do you want to to get rid of it?
  • NeedAlt * complain if no alt tag present
  • Base * to specify the base to use if we can't find the file
• New output format fits each image on one line - looks nice - tells you what the tags were and will be
• It actually recognizes if you have a base tag - although does nothing with it :-)
• now it matches IMaGe's and FIGUREs (HTML3)
• regexp to match an IMG or figure is much better
• The chown code is much tider - proper use of stat
• new istrue and isfalse functions for checking options values
• replaces old attributes if they are there - keeping formatting - or inserts new attributes at start of TAG. It only insert width & height rather than rebuidling the line from $one -> $seven.
• the base tag is recognized and if present the whole page it fetched using HTTP - do you really want this - some browsers don't even support the base tag! (A la Lynx)
• If the file cannot be found it tries to contact the server to ask it for that file - this gets around server path translation problems

• Added png support (works with all the png's I know about!)

• Added {1+"$@"} to perl startup bit at the top apparantly this is better

• Added slash to my address alex\@ed STUPID ME! thanks to Michael Fischer

• Lot's of changes!
• Made chmod & chown stuff an option cos caused problems for some people.
• moved over to gnu indentaion levels (using emacs perl mode)
• added Configuration file reader (with help from weblint source!)
• global variables are now set via the options array

• Added code by Neal.McBurnett@att.com to help keep permissions & ownership the same as original file (this got rid of chmod setting)
• wwwimagesize skips symbolic links!

• changed rename to copy so that original file's permissions were the same.

• made it so SRC did not need quotes around it

• fixed stupid mistake with regexp on remote urls
• made small change to jpeg code (Thanks to Tom Lane for this one)

• support for jpeg's with lots of help from andrew
• support for xbms
• option to change backup extention
• option to overwrite backup if there (or ask?)
• option to MAYBE change width & height if they're already there
• option to chmod output file (override your umask)
• now calls perl in a clever way! (stolen from weblint, which is ace!)
• made code for urls same as for files using streams
• tried to make code robust - if it comes across something wrong it will now carry on and make no changes
• changed name from gifsize to wwwimagesize
• hoppefully never dies ungracefully
• added lots of horrible perl code (this is my first attempt at perl programming, please be kind!)

Andrew Tong, werdna@ugcs.caltech.edu wrote gifsize which is where WWWis started from and here is his original changelog

• Fixed some serious bugs in the regular expressions.
• gifsize will now tell you what it's working on as it chugs along
• Now, it won't run if it finds a .html.bak file already existing.

• some code based on "client", an example program from Programming Perl, pp 342-344