[Changes for 0.24 - 2005-08-17]

* Explicity use "Encode::EBCDIC" and "Encode::IBM" at Parse::AFP
  loading time, to reduce runtime loading penalty and make it easier
  for PAR to detect the dependency.  Reported by Sunny Chou.

[Changes for 0.23 - 2004-10-21]

* udcsplit.pl now always allocates constant memory.

[Changes for 0.21 - 2004-10-14]

* udc4pca.pl now supports RMI/RMB sections.
* In cp835, \x40\x40 is recognised as a character.
* Both afpmkfontdb.pl and udc4pca.pl now supports multiple font databases.
* "Missing font in database" is no longer a fatal exception for udc4pca. 
* If a file has no UDC chunks, don't bother joining individual pages, just
  copy the whole file.

[Changes for 0.20 - 2004-10-13]

* Apply the same logic fix to udc4pca.pl.
* In udcsplit.pl, do not bother to parse PTX once a UDC is found.
* Do not assume an external 'cat' in udc4pca.pl.
* Sort afp fragments before join them.

[Changes for 0.19 - 2004-10-13]

* The UDC lookahead logic was broken in udcsplit.pl; its offset needs
  to be 11 bytes, not 2 bytes.
* Add BNG (Named Page Group) handling in udcsplit.pl.

[Changes for 0.18 - 2004-10-13]

* udc4pca.pl was not writing IM sections.
* Use fork() to implement a two-pass algorithm for udc4pca, as suggested
  by Thomas Yeh.

[Changes for 0.17 - 2004-10-12]

* More udc4pca.pl optimization: only processes PTX chunks that has
  UDC in its TRN sequences; everything else is sent directly to output.

[Changes for 0.16 - 2004-10-12]

* Make the tight loop even faster using a dedicated tight_loop() contruct.
  Requires setting {output_file} in new(), like script/udc4pca.pl now does.

[Changes for 0.15 - 2004-10-12]

* On lazy evaluation with callback_members that includes '*', skip all
  noninteresting callbacks and simply write them into the output file.

[Changes for 0.14 - 2004-10-08]

* In lazy-loading mode, AFP records are now read in one-by-one.
* Remove RCS tags since we're now under SVK management.
* Correct README to use the Perl license, not only GPL.

[Changes for 0.13 - 2004-09-30]

* udc4pca: Respects font resolution, instead of hardcoding on 300.
* udc4pca: EBCDIC documents are respected.
* udc4pca: Space increments from FNO is now handled.
* New partial FNO parsing subclass.

[Changes for 0.12 - 2004-09-23]

* UDC detection logic in udc4pca.pl fixed with regards to the Big5 range.
* Update afpmkfontdb SQL schema to reflect that Bitmap should be BLOB.
* udc4pca: calculate DBCS increment on-the-fly instead of multiplying it later.
* udc4pca: use Black as the UDC color instead of 0200.

[Changes for 0.11 - 2004-09-22]

* In udc4pca.pl, respect SBI and BLN in setting baseline increments. 
* New PTX parsing subclasses for ( SIA SBI SVI STC BSU ESU RMB RMI RPS NOP ).
  Both contributed by Thomas Yeh.

[Changes for 0.10 - 2004-09-21]

* In udc4pca.pl, do not write out BII sections with the page has no UDC in it.
* New parsing subclasses for: ( BDM BDX BPM DXD EDM EDX EPM LNC LND )
  Contributed by Thomas Yeh.

[Changes for 0.09 - 2004-09-16]

* Fixed an infinite loop bug when parsing MCF1.
* afpdump.pl can now decode character sets using -e.
* We now depend on Encode::IBM to process various IBM-specific encodings.
* New utility, udc4pca.pl, that fills in UDC characters from fonts.db. 
* New API, ->spawn_obj, that constructs a new AFP record object.

[Changes for 0.08 - 2004-09-13]

* The HTML generated by afpdump.pl is now much more readable and pretty.
* Added the ability of detecting and displaying info on fixed width fonts
  to afpmkfontdb.pl and afpviewfontdb.pl.
* More correct parsing for PGD and STO.
* Fixed a spurious warning in afpconv.pl.

[Changes for 0.07 - 2004-09-09]

* Reduced memory usage for script/* tools.
* Depend on Parse::Binary 0.08 to further reduce memory usage.

[Changes for 0.06 - 2004-09-08]

* Corrected DBD::SQLite's BLOB usage in fontdb scripts.
* Cleaned up and speeded up afpmkfontdb.pl.
* Give sane defaults for the scripts.
* Correctly install all scripts instead of only the first one.
* Inhibit test warnings temporarily.

[Changes for 0.05 - 2004-09-08]

* New scripts: afpsplit.pl, afpmkfontdb.pl, afpviewfontdb.pl.
* Speeded up afpdump.pl and PTX parsing.
* Added CPC, FNC parsing.
* Raise dep to Parse::Binary 0.06, Perl 5.8.2.

[Changes for 0.04 - 2004-02-17]

* PTD1 and PGP1 were broken.
* Add backtrace sighandler to script/*.
* Switch to callback_members API.
* Add afpconv.pl.
* ->dispatch and ->dispatch_members, two utility method to ease filter writing.
  eventually we want it to become a full Parse::AFP::Filter class.

[Changes for 0.03 - 2004-02-13]

* Silence warnings during pack.
* Multimember support.
* Refactor per-member dump back to Parse::AFP.
* Start of variant packing
* Use Parse::Binary instead of Parse::AFP::FixedFormat.
* Adjust to the new accessor methods of Parse::Binary
* Move length functions to AFP side.
* Properly install afpdump.pl.

[Changes for 0.02 - 2004-02-10]

* Add afpdump.pl, an utility to dump afp files to HTML files.
* Raise dependency to Perl 5.8.1+.

[Changes for 0.01 - 2004-02-09]

* Initial CPAN release.