Fossil

Timeline
Login

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

50 most recent check-ins by user george

2022-05-25
20:52
Clarify the documentation: say explicitly that Clone capability, "g" provides read access to tickets, wiki and forum; and that Write capability, "i" implies that changes to tickets, wiki and forum are also accepted from clones upon syncronization. ... (Leaf check-in: 6e5337e8 user: george tags: trunk)
2022-05-24
20:32
If neither TICKET nor TICKETCHNG table has regular "mimetype" column then generated "mimetype" columns of these two tables are queried and used during the extraction of backlinks from ticket change artifacts. See forum post 40c1208a0f84 for rationale. ... (check-in: 6fb642c0 user: george tags: trunk)
2022-05-18
22:52
Fix a couple of invokations of backlink_extract() that were overlooked in the previous check-in. ... (Closed-Leaf check-in: 33c67c28 user: george tags: generated-tkt-mimetype)
20:21
Add support for generated "mimetype" column in the TICKET table. Introduce integer mimetype codes and refactor backlink_extract() accordingly. Make the overall handling of the generated "mimetype" columns a bit more clear. ... (check-in: e39f7790 user: george tags: generated-tkt-mimetype)
2022-05-14
20:52
Handle a possible existence of the additional checkbox introduced in [29a24941ed9bf4]. ... (Leaf check-in: 76844c35 user: george tags: wcontent-subsets)
19:12
A tiny optimization inside of ticket_insert(): rearrange two db_finalize() calls in order to avoid additional malloc() for zMimetype. ... (check-in: 563c6b9e user: george tags: generated-tkt-mimetype)
18:27
If "mimetype" column of the TICKETCHNG table is a generated column, than use these generated values for extraction of backlinks from the ticket-changes. ... (check-in: 5ab73ab8 user: george tags: generated-tkt-mimetype)
17:42
Fix a subtle bug in ticket_insert() which may lead to redundant rows in the BACKLINK table. The bug appeared in [7c13a57358ae]. ... (check-in: 3b42738e user: george tags: generated-tkt-mimetype)
2022-05-13
23:35
If TICKETCHNG table has TKT_USER column then make the values in that column match to the corresponding U-cards. Add TKT_USER to the default schema. See forum thread 5593538afb77 for rationale. ... (check-in: cf00d07c user: george tags: trunk)
2022-05-11
21:23
Ensure that $requested_page is always initialized. This is a follow-up to [c68fa2edd76d90] that fixes a bug in [b05a6c6bc826d3]. ... (check-in: a44e3c73 user: george tags: trunk)
2022-05-09
21:22
Supply TH1 interpreter with two auxiliary variables: $requested_page (holds the first segment of the requested PATH_INFO) and $canonical_page (holds the canonical name of a web-page being served). Also amend default HTML header: add two corresponding classes to the BODY element and move <meta charset="UTF-8"> to the top. ... (check-in: b05a6c6b user: george tags: trunk)
2022-05-08
20:40
Rename proposed column to tkt_user which seems more natural. Amend default schema to create this column (and also drop a nearby comment). Add a tiny optimization to getAllTicketFields(). ... (Closed-Leaf check-in: 6d95ae4f user: george tags: tkt-ucard)
2022-05-01
21:00
Add support for optional tkt_ucard column of the TICKETCHNG table. If this column is present then its value is populated from the U-card of the corresponding artifact. ... (check-in: d681e1dc user: george tags: tkt-ucard)
20:53
Minor code clean-up inside of ticket_insert() function. ... (check-in: 3e4ba24e user: george tags: tkt-ucard)
2022-04-23
21:32
Extend Markdown with footnotes support. See known limitations and the corresponding forum thread. ... (check-in: 3990518b user: george tags: trunk)
17:23
Change signature of add_inline_footnote() in order to move away from returning of unreliable pointer. Amend a few comments. Fix a couple of minor issues that fuzzer complains about. ... (Closed-Leaf check-in: 0850862e user: george tags: markdown-footnotes)
15:56
Fix a possible heap-buffer-overflow in parse_htmlblock() introduced by [1e919d601f774fdb]. This is not related to footnotes but was revealed by fuzzing (case 80cbb6b185807e98a953426af7b1f802c9d13957). ... (check-in: bc4c5b63 user: george tags: markdown-footnotes)
2022-04-22
12:49
Remove redundant assert() that fails for the case when the content of a span-bounded inline footnote is rendered into a void. This is a corner case that was revealed via fuzzing. ... (check-in: cab8a586 user: george tags: markdown-footnotes)
2022-04-21
22:31
Add a test case for fragment-bounded footnote that contains markup within the corresponding text fragment. ... (check-in: d38ec43d user: george tags: markdown-footnotes)
21:13
Fix another use-after-realloc bug in handling of inline footnotes which was discovered during fuzzing. Also fix a few other issues revealed via fuzzer. ... (check-in: c5456211 user: george tags: markdown-footnotes)
13:36
Revert Makefile.in to normal builds (leave a few comments about fuzzing). ... (check-in: 94077966 user: george tags: markdown-footnotes)
13:16
Fix a use-after-free bug in handling of nested inline footnotes. The bug was discovered by fuzzing with '-fsanitize=fuzzer,undefined,address -DFOSSIL_FUZZ' appended to TCCFLAGS in Makefile.in. It's noteworthy that the ',undefined,address' part was essential to find the bug (otherwise just 'double-free' was reported). Many thanks to Stephan for documenting the fuzzing procedures and support. ... (check-in: 31e5df5f user: george tags: markdown-footnotes)
2022-04-19
15:25
Remove unnecessary field from the auxiliary union 'bitfield64_t' and amend the corresponding comments. Also add comment about FOOTNOTES_WITHOUT_URI macro. ... (check-in: cf1e9691 user: george tags: markdown-footnotes)
2022-02-23
12:33
Count overnesting as the fourth type of the footnote-related issues and report accordingly. ... (check-in: ae297bb6 user: george tags: markdown-footnotes)
09:45
Minor refactoring. Move the definition of BLOB_APPEND_LITERAL() macro from markdown_html.c to blob.c so that it could be used outside of markdown_html.c. Also rename it to lowercase for consistency with other API. Within markdown.c use that newly available macro instead of blob_append_string(). Within markdown_html.c use it for footnotes-relevant code. Other invocations of BLOB_APPEND_LITERAL() within markdown_html.c are left intact (they use an alias) in order to simplify the potential merge with the trunk. ... (check-in: c8a8d0c9 user: george tags: markdown-footnotes)
08:21
Minor refactoring. Move the definition of BLOB_APPEND_BLOB() macro from markdown_html.c to blob.c so that it could be used outside of markdown_html.c. Also rename it to blob_appendb() for consistency with blob_appendf() and other API. Within markdown.c use that newly available macro where appropriate. Within markdown_html.c use it for footnotes-relevant code. Other invocations of BLOB_APPEND_BLOB() within markdown_html.c are left intact (they use an alias) in order to simplify the potential merge with the trunk. ... (check-in: 33a681eb user: george tags: markdown-footnotes)
07:36
Fix handling of user-provided classes for unreferenced, joined and overnested footnotes. In all these cases the tokens of user-provided classes are rendered as plain-text and no special classes are added anywhere. ... (check-in: 875472a8 user: george tags: markdown-footnotes)
2022-02-21
05:14
Add a comment for append_footnote_upc(). Also substitute a variable of zero value with just "0" constant. No functional changes. ... (check-in: ae8a3dd5 user: george tags: markdown-footnotes)
04:29
Impose a limit on the depth of nesting of inline footnotes. Also add a few test cases: for depth limiting and HTML hijacking. ... (check-in: f4ff013a user: george tags: markdown-footnotes)
2022-02-20
23:00
If there are issues with footnotes then set TH1 variable $footnotes_issues_counters to a space separated list of integers that count for "misref", "unref" and "joins". This eliminates the need for JavaScript for the case when a custom skin wants to warn about issues with footnotes in the header of a page.
Also fix counting of "joins": count the number of unique labels that have multiple definitions (and not the number of such definitions).
... (check-in: 773cef5c user: george tags: markdown-footnotes)
2022-02-19
01:59
Fix hyperlinks on the /winfo page. These were broken when a page was accessed through /info/HASH or /winfo/HASH aliases. ... (check-in: 356a4845 user: george tags: trunk)
01:16
Parse inline footnotes even if a renderer does not define a callback for rendering of footnote markers. This seems more correct even though the current implementation of backlink processor does define such callback as an empty function. ... (check-in: e06c12d1 user: george tags: markdown-footnotes)
01:00
Handle some corner cases more thoroughly: dismiss empty footnotes, passthrough (more carefully) user-provided classlist if the token is not followed by a blank character or if a footnote's text consists just of such token and blank characters. Also simplify a little bit a few places inside of is_footnote() function. ... (check-in: fe315780 user: george tags: markdown-footnotes)
2022-02-18
01:33
Add --lint-footnotes option to the test-markdown-render command. If this flag is given and footnotes in the input have issues, then print to stderr the counters of "misrefs", "strays" and "split-defs" and exit with error. This should partially address a concern raised at the forum. ... (check-in: 1f525713 user: george tags: markdown-footnotes)
2022-02-17
22:09
If a footnote's text starts with a token of the special form then use this token to derive a set of CSS classes that are added to that footnote and its references. This enables users to style elements of a particular footnote provided that the administrator provisioned and documented some special CSS classes in a custum skin. Default skin does not provide any of such special classes which makes this feature an "opt-in". ... (check-in: 92516ced user: george tags: markdown-footnotes)
00:17
Clean-up and rephrase some comments. ... (check-in: a62c8768 user: george tags: markdown-footnotes)
2022-02-16
23:08
Make parsing slightly faster and fix a comment. No changes in functionality. ... (check-in: a36dd09d user: george tags: markdown-footnotes)
22:11
Include REQUEST_URI into footnotes' hyperlinks. This should make links work even if base href (in a page's header) is not consistent with the REQUEST_URI. If FOOTNOTES_WITHOUT_URI macro is defined while compiling src/markdown_html.c then bare "#fragment" hyperlinks (without REQUEST_URI) are generated. ... (check-in: 2c1f8f35 user: george tags: markdown-footnotes)
2022-02-15
21:35
FIXME-comments in the code did not indicate issues and are now rephrased to be mere reminders. ... (Leaf check-in: d253ece0 user: george tags: base-href-fix)
02:26
Do not override base href for wiki pages. This needs testing. ... (check-in: 89dd3451 user: george tags: base-href-fix)
00:58
When setting $webpagename pass the value of g.zPath though escape_quotes(). This is unclear how g.zPath can contain quotes but let this strange case be handled in a safe way. ... (check-in: 288fe34c user: george tags: base-href-fix)
00:14
Move <meta charset="UTF-8"> to the begining of the default header. Supply BODY element with a class that derives from $webpagename. ... (check-in: 6d135904 user: george tags: base-href-fix)
2022-02-14
23:32
Minor code refactoring: rename a temporary variable and utilize matching_bracket_offset() one more time. No changes in functionality. ... (check-in: 5b845a07 user: george tags: markdown-footnotes)
23:06
Do not export g.zRelReqURI to TH1 interpreter because getParameter proc can retrieve PATH_INFO and QUERY_STRING. Instead export g.zPath (as $webpagename) since that is typically needed in the TH1 headers/footers of custom skins. ... (check-in: ff4c7ed6 user: george tags: base-href-fix)
22:43
Make style_set_base_href_suffix() safe for misuse: if the resulting suffix contains unescaped quotes then escape them. $base_href_suffix is intended for interpolation inside of the quoted href attribute. This check-in should address the case when a user of malfunctioning browser (which mishandles quoting) is tricked by an adversary to visit a specially crafted hyperlink. ... (check-in: d97752f3 user: george tags: base-href-fix)
2022-02-13
19:29
Fix parsing of "free-standing" footnotes that was (slightly) broken by the previous check-in. ... (check-in: 23c3e0b2 user: george tags: markdown-footnotes)
17:54
Rename variable g.zUrlSuffix to g.zRelReqURI (Relative Request URI). Provide it to TH1 interpreter as $relrequri. ... (check-in: 05e3fa76 user: george tags: base-href-fix)
16:00
Fix hyperlinks on the /winfo page. These were broken when a page was rendered through /info/HASH of /winfo/HASH aliases. ... (check-in: 5ce372ce user: george tags: base-href-fix)
01:35
Cherry-pick from branch 'cgi-compliance' (and thus back out [5bb921dd0893a548]). Adapt the computation of g.zUrlSuffix in set_base_url() accordingly. ... (check-in: 5c649c7e user: george tags: base-href-fix)
2022-02-12
21:51
Add an overlooked call to style_set_base_href_suffix() in src/info.c and add a comment of why not to do so for /ext ... (check-in: 87fba731 user: george tags: base-href-fix)