This page lists the bugs fixed in each release of AscToHTM. For a list of functionality changes see the AscToHTM documentation
- Fixed bug that caused the output filename to be reset if you changed policy files.
- Fixed bug whereby a file could be converted over itself with unexpected results, e.g. file.html would become file.html. This could happen when converting wildcards that allowed any (or no) extensions and the output files are placed in the same directory as the input files.
- Some drop-down boxes (such as Script File) wouldn't accept values wider than the data entry box.
- After selecting multiple files the select policy file dialog would have the wrong defaults in it.
- Fixed long-standing limitation whereby the status window would stop updating ad could only save 32,676 charatters to a log file. Now the window keeps updating, and it can all be saved to the log file, although you may not be able to scroll back through all the text.
- When generating frames with .htm extension, the contents list was .html
- When loading a policy with "Use .htm extension" set into a session with .htm generation already selected the "repeated policy" warning message is now suppressed
- Fixed bug that wouldn't correctly resolve certain relative file names in INCLUDE statements.
- Changed algorithm that reverses URL encoding to only do so on the domain part of a URL.
- Fixed bug that could cause the second of two underlined headings to be enclosed in a pair of <HR> tags.
- Fixed bug whereby a single blank line after a code sample or other preformatted region wouldn't be enough to cause a single line to be placed inside paragraph tags.
- Fixed bug whereby a delimited section such as BEGIN_PRE...END_PRE immediately following an underlined heading would leave the first line out of the formatting.
- TABLE_LAYOUT wasn't working as advertised
- Fine-tuned the detection of whether or not a file has an in-situ contents list
- Corrected text on Settings Tab
- Fixed bug in colour samples when loading policy with a "#" in it.
- Some policies weren't taking effect the first time they were changed.
- Stop "<" becoming a bullet
- Corrected confusion between left and right margins
- Fixed bug for handling short lines (introduced in version 3.3)
- Fixed emphasis bug whereby "* * *" would end up marking some white space as <STRONG>.
- Fixed bug that caused "Heading Colour" policy not to be implemented unless a font was selected or CSS markup was suppressed
- Fixed bugs in table analysis when "Expect sparse tables" was enabled. Fixed a number of table analysis bugs.
- Fixed bug if created file had invalid characters in it's filename.
- Program now handles duplicate headings correctly in the contents list by adding an "instance count" to the anchor name.
- Fixed a bug that lead to non-unique anchor names for duplicate headings.
- Console version didn't terminate correctly if the a mass conversion ran out of disk space.
- Fixed bug where line breaks were inserted before HTML entities like á causing unwanted spaces before special characters, especially those used in foreign languages.
- Changed the save policy dialog to add the extension ".pol" if omitted
- Fixed bug that caused colour values with a '#' in them to be incorrectly displayed.
- Changed to not allow ":" in filenames. This could cause conversion into many files fail when the underlined heading contained a ":".
- Fixed bug in naming duplicate files when splitting a document at underlined headings.
- Improved hyperlink parsing. In particular the domain name is checked to ensure it's plausible using improved algorithms.
- Improved URL and email address validation. In particular I now check the domain name is viable (.com, .edu, .co.uk etc) and handle better.
- Stopped "double reporting" of key policy decisions.
- Improved error handling when writing to full device like a RAM disk or a floppy disk.
- A small change is made to limit the maximum possible page width detected, and a bug related to mail headers is fixed.
- Changed detection of new paragraphs after hanging paragraphs. This fixes an analysis error.
- The Windows help file in version 3.0 wasn't accessible via the F1 key. This is now fixed.
- INCLUDE statements now report any failure to open the file, and insert directory paths where necessary to resolve the file locations (Windows)
- Fixed various bugs so that the output filename and directory can be changed. Previously this didn't work too well.
- So many others, I can't remember them all :-)
- Numbers in the text will now only convert to hyperlinks if the exact section number can be found in the document. This improves on previous versions that only did approximate range checks.
- Underlined headings may now be split into files, and duplicate filenames are now avoided
- URLs separated by semi-colons were not being correctly converted
- email addresses with two "@"s are now rejected, and no longer accept "%" at start of email address
- No longer treat ___some_text___ as underlined
- Contents list generation failed for documents with "Section: n" as the headings
- Fixed bug that could wrongly right-align some columns in a table if there was a big gap.
- Fixed various bugs related with emphasis phrases that spanned multiple lines. We now no longer allow detection of "nested" emphasis, nor is emphasis allowed to cross blank lines.
- Improved file handling. Now get error messages when files aren't found, and the program is better at locating files under Windows. The program now supports the use of relative filenames for include files etc.
- In a usenet reference such as news://comp.os.vms the "news:" was being left outside the display part of the generated hyperlink
- Repeated Link definitions on the same line weren't being substituted correctly
- Improved handling of user-delimited contents lists using the BEGIN_CONTENTS ... END_CONTENTS directives
- General Protection Fault was occurring on the "Analyse file" button if a file had not yet been converted. This bug was introduced in V2.3
- No longer inserts <BR> before first heading if the first line in the file is a heading
- "headings" that were underlined, but which were in fact part of a table are no longer added to generated contents lists.
- Underlined/capitalised headings now get correct hyperlinks in contents list lists when using DOS filenames
- Fixed bug that caused contents lists immediately following underlined headings to not be recognised
- URLs in heading no longer lead to invalid anchor tags
- Various URL changes; "alt." and "news." no longer treated as a newsgroup, https:// pages now recognised
- Properly reject "indent" of 1. Previously this was reported as rejected, but was actually accepted.
- "Expect code samples" policy was being ignored in table calculations.
- The TABLE_HEADER_COLS directive only worked when there were header rows as well.
- Use of emphasis inside a TABLE cell was not being detected at all. Now it is detected if held on a single line. Phrases that are emphasised over several lines inside a table cell may still not be detected.
- Fixed a bullet numbering problem. Previously the bullet numbering
wasn't being reset correctly if two sets of bullets were separated
by text with a larger margin than the bullets.
- Emphasis at end of line now properly recognised
- Fixed bug that caused emphasis markup near pre-formatted sections of
- Small correction to line-splitting algorithm to prevent <TAG>s being
spilt when they're near the natural break.
- Pre-processor command $_$_TITLE was being ignored if the "Use first heading as title" policy was set.
- Added a "Preform simple conversion" tick box on the front panel. This does exactly the same as the "Keep it simple" policy.
- Improved the Headings dialog to allow headings policies to be more easily edited now.
- Pre-processor document sections now working.
- Wildcard support has been added (see 188.8.131.52).
- Major re-writing of contents list generation has occurred (see 3.4.2). Includes new "Use any existing contents list" and "Generate external contents file". More changes are expected here in later versions.
- New Directory Page feature. Supporting policies include:-
"Show file titles in Directory"
"Indent headings in Directory"
"Directory return hyperlink text"
"Directory Script file"
"Directory header file"
"Directory footer file"
- New "Minimum TABLE column separation" policy and "TABLE_MIN_COLUMN_SEPARATION" pre-processor command to allow some tuning of table analysis.
- New "Use first heading as title" policy
- New "Use first line as title" policy
- New "Recognised USENET groups" policy
- New "Automatic centring tolerance" policy
- New "Use <P> markup for paragraphs" policy to allow choice of either <P> or <BR> markup to be used for paragraphs.
- New "Default table width" policy and "TABLE_WIDTH" pre-processor command to allow table widths to be specified as percentages
- New pre-processor command "HTML_LINE"
- End effects now fixed.
- Various emphasis features eliminated.
- In-situ contents lists weren't getting the right file names in their hyperlinks when the file was being split.
- Right justified numbered sections weren't being detected correctly past section 9.
- No longer break long URLs over two lines. This occasionally led to hyperlinks that didn't work.
- No longer generate files for underlined sections when document has numbers sections as well, and is being split into files.
- No longer detect pre-formatted text inside BEGIN_HTML ... END_HTML section.
- Fixed tab conversion bug.
- Fixed "Minimum automatic <PRE> size". Previously didn't work at all like advertised.
- Colour samples in Windowed version were being shown as gray on initial draw and on re-draw.
- Anchor points added to generated contents list had their </A> missing.
- Fixed occasional "Invalid indent -1" error
- Loads of bugs in parsing user PRE sections (sorry Dennis!).
- < and > characters inside a PRE section caused characters to be lost off the end of lines
- URL-parsing improved
- Contents list file links back to main file if no other section links generated
- Newsgroups in headings no longer converted
- Several hyperlink parsing errors fixed. Previously there were problems with punctuation around links, email addresses with protocols (e.g. MX%"email@example.com") and newsgroups with the word "news:" in front e.g. news:uk.jobs.
- improved output of pre-formatted text. "<" characters were getting confused, and the pre-formatted lines were being broken in two.
- fixed some errors that occur when directory paths are included in the filenames. Probably still more changes required in this area, particularly with a view to supporting multiple file drag'n'drop under Windows.
- improved detection of pre-formatted regions.
- Fixed lots of memory leaks
- Improved cross-reference detection. These are now range checked against the observed section numbers. This will reduce the likelyhood of DirectX 3.0 and Windows 95 becoming links to chapters 3 and 95.
- Contents list generation for documents with chapters and no subsections now works.
- Improved Contents list detection.
- Fixed bug that caused links to underlined or capitalised heading with very long names to sometimes fail.