Daniel Veillard | 01791d5 | 1998-07-24 19:24:09 +0000 | [diff] [blame] | 1 | |
Daniel Veillard | ad8f99d | 2000-01-15 14:20:03 +0000 | [diff] [blame^] | 2 | TODO for the XML parser and stuff: |
| 3 | ================================== |
Daniel Veillard | 01791d5 | 1998-07-24 19:24:09 +0000 | [diff] [blame] | 4 | |
Daniel Veillard | ad8f99d | 2000-01-15 14:20:03 +0000 | [diff] [blame^] | 5 | URGENT: |
| 6 | ======= |
| 7 | - Support for UTF-8 and UTF-16 encoding |
| 8 | => added some convertion routines provided by Martin Durst but I didn't |
| 9 | try to glue them in. I plan to keep everything internally as UTF-8 |
| 10 | this is slightly more costly but more compact, and recent processors |
| 11 | efficiency is cache related. The key for good performances is keeping |
| 12 | the data set small, so will I. |
| 13 | => the new progressive reading routines call the detection code which |
| 14 | need to be enabled, then thest the ISO->UTF-8 stuff, and add more |
| 15 | charset conv routines. |
| 16 | |
| 17 | TODO: |
| 18 | ===== |
| 19 | |
| 20 | - Tools to produce man pages from the SGML docs. |
| 21 | |
| 22 | - Finish XPath |
| 23 | => attributes addressing troubles |
| 24 | => defaulted attributes handling |
| 25 | => namespace axis ? |
| 26 | |
| 27 | - Add Xpointer recognition/API |
| 28 | |
| 29 | - Add Xlink recognition/API |
| 30 | => started adding an xlink.[ch] with a unified API for XML and HTML. |
| 31 | |
| 32 | - Implement XSLT |
| 33 | => seems that someone volunteered ?!? |
| 34 | |
| 35 | - Implement XSchemas |
| 36 | |
| 37 | - O2K parsing; |
| 38 | => this is a somewhat ugly mix of HTML and XML, adding a specific |
| 39 | routine in the comment parsing code of HTML and plug the XML |
| 40 | parsing one in-there should not be too hard. Key point is to get |
| 41 | XSL to transform all this to something decent ... |
| 42 | |
| 43 | - Add regression tests for all WFC errors |
| 44 | => did some in test/WFC , not added to the Makefile yet. |
| 45 | |
| 46 | - Optimization of tag strings allocation. |
| 47 | |
| 48 | - Language identification code, productions [33] to [38] |
| 49 | |
| 50 | - Conditional sections in DTDs [61] to [65] |
| 51 | => should this crap be really implemented ??? |
| 52 | |
| 53 | |
| 54 | - Allow parsed entities defined in the internal subset to override |
| 55 | the ones defined in the external subset (DtD customization). |
| 56 | => This mean that the entity content should be computed only at |
| 57 | use time, i.e. keep the orig string only at parse time and expand |
| 58 | only when referenced from the external subset :-( |
| 59 | Needed for complete use of most DTD from Eve Maler |
| 60 | |
| 61 | - maintain coherency of namespace when doing cut'n paste operations |
| 62 | => the functions are coded, but need testing |
| 63 | |
| 64 | - function to rebuild the ID table ? |
| 65 | |
| 66 | - extend the shell with: |
| 67 | - edit |
| 68 | - load/save |
| 69 | - mv (yum, yum, but it's harder because directories are ordered in |
| 70 | our case, mvup and mvdown would be required) |
| 71 | |
| 72 | - Parsing of a well balanced chunk |
| 73 | |
| 74 | - Add HTML validation using the XHTML DTD |
| 75 | - problem: do we want to keep and maintain the code for handling |
| 76 | DTD/System ID cache directly in libxml ? |
| 77 | |
| 78 | - Add a DTD cache prefilled with xhtml DTDs and entities and a program to |
| 79 | manage them -> like the /usr/bin/install-catalog from SGML |
| 80 | right place seems $datadir/xmldtds |
| 81 | |
| 82 | - turn tester into a generic program xml-test installed with xml-devel |
| 83 | |
| 84 | - Add output to XHTML in case of HTML documents. |
| 85 | |
| 86 | - dynamically adapt the alloc entry point to use g_alloc()/g_free() |
| 87 | if the programmer wants it |
| 88 | |
| 89 | Done: |
| 90 | ===== |
| 91 | |
| 92 | - External entities loading: |
| 93 | - allow override by client code |
| 94 | - make sure it is alled for all external entities referenced |
| 95 | Done, client code should use xmlSetExternalEntityLoader() to set |
| 96 | the default loading routine. It will be called each time an external |
| 97 | entity entity resolution is triggered. |
| 98 | - maintain ID coherency when removing/changing attributes |
| 99 | The function used to deallocate attributes now check for it being an |
| 100 | ID and removes it from the table. |
| 101 | - push mode parsing i.e. non-blocking state based parser |
| 102 | done, both for XML and HTML parsers. Use xmlCreatePushParserCtxt() |
| 103 | and xmlParseChunk() and html counterparts. |
| 104 | The tester program now has a --push option to select that parser |
| 105 | front-end. Douplicated tests to use both and check results are similar. |
| 106 | |
| 107 | - Most of XPath, still see some troubles and occasionnal memleaks. |
| 108 | - an XML shell, allowing to traverse/manipulate an XML document with |
| 109 | a shell like interface, and using XPath for the anming syntax |
| 110 | - use of readline and history added when available |
| 111 | - the shell interface has been cleanly separated and moved to debugXML.c |
| 112 | - HTML parser, should be fairly stable now |
| 113 | - API to search the lang of an attribute |
| 114 | - Collect IDs at parsing and maintain a table. |
| 115 | PBM: maintain the table coherency |
| 116 | PBM: how to detect ID types in absence of DtD ! |
| 117 | - Use it for XPath ID support |
| 118 | - Add validity checking |
| 119 | Should be finished now ! |
| 120 | - Add regression tests with entity substitutions |
| 121 | |
| 122 | - External Parsed entities, either XML or external Subset [78] and [79] |
| 123 | parsing the xmllang DtD now works, so it should be sufficient for |
| 124 | most cases ! |
| 125 | |
| 126 | - progressive reading. The entity support is a first step toward |
Daniel Veillard | 260a68f | 1998-08-13 03:39:55 +0000 | [diff] [blame] | 127 | asbtraction of an input stream. A large part of the context is still |
| 128 | located on the stack, moving to a state machine and putting everyting |
| 129 | in the parsing context should provide an adequate solution. |
Daniel Veillard | ad8f99d | 2000-01-15 14:20:03 +0000 | [diff] [blame^] | 130 | => Rather than progressive parsing, give more power to the SAX-like |
| 131 | interface. Currently the DOM-like representation is built but |
| 132 | => it should be possible to define that only as a set of SAX callbacks |
| 133 | and remove the tree creation from the parser code. |
| 134 | DONE |
Daniel Veillard | 01791d5 | 1998-07-24 19:24:09 +0000 | [diff] [blame] | 135 | |
Daniel Veillard | ad8f99d | 2000-01-15 14:20:03 +0000 | [diff] [blame^] | 136 | - DOM support, instead of using a proprietary in memory |
| 137 | format for the document representation, the parser should |
| 138 | call a DOM API to actually build the resulting document. |
| 139 | Then the parser becomes independent of the in-memory |
| 140 | representation of the document. Even better using RPC's |
| 141 | the parser can actually build the document in another |
| 142 | program. |
| 143 | => Work started, now the internal representation is by default |
| 144 | very near a direct DOM implementation. The DOM glue is implemented |
| 145 | as a separate module. See the GNOME gdome module. |
| 146 | |
Daniel Veillard | 01791d5 | 1998-07-24 19:24:09 +0000 | [diff] [blame] | 147 | - C++ support : John Ehresman <jehresma@dsg.harvard.edu> |
| 148 | - Updated code to follow more recent specs, added compatibility flag |
Daniel Veillard | 260a68f | 1998-08-13 03:39:55 +0000 | [diff] [blame] | 149 | - Better error handling, use a dedicated, overridable error |
| 150 | handling function. |
| 151 | - Support for CDATA. |
| 152 | - Keep track of line numbers for better error reporting. |
| 153 | - Support for PI (SAX one). |
Daniel Veillard | ad8f99d | 2000-01-15 14:20:03 +0000 | [diff] [blame^] | 154 | - Support for Comments (bad, should be in ASAP, they are parsed |
| 155 | but not stored), should be configurable. |
| 156 | - Improve the support of entities on save (+SAX). |
Daniel Veillard | 01791d5 | 1998-07-24 19:24:09 +0000 | [diff] [blame] | 157 | |