Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
50 most recent check-ins by user stephan
|
2025-05-08
| ||
| 13:38 | Fix some doc hyperlinks, and remove some long-dead ones, as suggested in forum post 5022f7b600. The git-detached-head link in that post was not fixed because the Internet Archive link to it is not working for me. ... (check-in: 56e59a3c user: stephan tags: trunk) | |
| 11:25 | Replace a C++-style comment. No functional changes. ... (check-in: a68a7eac user: stephan tags: trunk) | |
|
2025-04-30
| ||
| 15:37 | Add a section to www/customskin.md about using skintxt2config.c instead of pasting changes into the UI. ... (check-in: 37e27051 user: stephan tags: trunk) | |
| 14:43 | Add CSS class md-table to TABLE elements created specifically from a markdown-format table. In response to forum post 3a5d04039553e5b4 ... (check-in: f0d11ab2 user: stephan tags: trunk) | |
|
2025-04-23
| ||
| 10:46 | When emitting the default password as part of the 'new' command, add the term 'remote-access' to it to clarify that it's only for remote use. Indirectly suggested by forum post 9dbd8e00ee. ... (check-in: 57276a51 user: stephan tags: trunk) | |
|
2025-04-15
| ||
| 10:22 | Slightly simplify the previous checkin. ... (check-in: 6a9c71f3 user: stephan tags: trunk) | |
| 10:15 | A proposed solution to the problem of /setup_ulist fails for repos with no subscriber table. Reported in forum post e2b0008592f6a776. ... (check-in: e50a5aac user: stephan tags: trunk) | |
|
2025-04-12
| ||
| 09:33 | Add a mention of /chat's recent improvements in the change log. ... (check-in: 7a85c54b user: stephan tags: trunk) | |
| 01:15 | Fix an age-old /chat bug which caused /chat-delete posts to misadjust the ajax-in-flight counter, leading to various UI widgets not being disabled when they should after a message is deleted. ... (check-in: 2827a38a user: stephan tags: trunk) | |
| 00:07 | Maintainability cleanups and docs in /chat. No (intended) functional changes. ... (check-in: 084001c7 user: stephan tags: trunk) | |
|
2025-04-11
| ||
| 22:48 | Internal doc updates in fossil.fetch(). Ensure that fossil.fetch()'s onerror()/ontimeout() handler do not propagate exceptions (a defensive measure, not a fix for a known bug). ... (check-in: 1d3db505 user: stephan tags: trunk) | |
| 20:29 | In /chat-generated notifications (typically error messages), add a button to the drop-down options to delete all notifications. That replaces the 'delete all poller notifications' button which previously only showed up only on those message types. Add a mention of the backoff timer to chat.md. ... (check-in: da1c351b user: stephan tags: trunk) | |
| 19:35 | Replace an a recurrent setInterval() timer in /chat's poll-connection error handler with a single-fire-as-needed setTimeout(). This saves some CPU and allows /chat to respond more quickly to non-timeout HTTP errors. ... (check-in: 1bfb06c7 user: stephan tags: trunk) | |
| 18:52 | Further refinements of the chat poll connection detection. The first N ignored errors are now spaced out unevenly. Use the server's configured chat-poll-timeout as the basis for calculating our client-side timeout time. ... (check-in: e8bbaf92 user: stephan tags: trunk) | |
| 16:09 | Minor cosmetic tweaks to the poll-in-distress indicator. Make it yellow in dark-mode skins, as red blends in too well. No functional changes. ... (check-in: 160d2692 user: stephan tags: trunk) | |
| 15:30 | Teach /chat to not be so verbose about connection errors. The first 3 will be subtly signaled via a tiny red line between the input field and message list, which will go away once the poller connection is re-established. After that, it will resort to the more verbose notifications. ... (check-in: e3eb8399 user: stephan tags: trunk) | |
|
2025-04-10
| ||
| 16:11 | Add a 'retry now' button to the /chat poller timeout notification widgets. Fix a bug which kept Chat.deleteMessageElem() from working if it was passed a MessageWidget instance. ... (check-in: 8c64fa26 user: stephan tags: trunk) | |
| 15:14 | Fix an = vs == bug in the /chat timeout error handling which only coincidentally worked that way. Update some /chat-internal docs. ... (check-in: 5fc2f2a0 user: stephan tags: trunk) | |
| 14:24 | Remove comments calling for a now-resolved TODO. No code changes. ... (check-in: a5521984 user: stephan tags: trunk) | |
| 14:22 | Teach /chat to behave better when the connection to the remote server goes down, using a back-off timer to throttle reconnection attempts instead of blindly sending one reconnection request per second. ... (check-in: 39b4bd9c user: stephan tags: trunk) | |
| 07:23 | When deleting a connection-restored notification, also delete the connection-broken notification it's associated with. Add a button to the message popup for connection-related notifications to delete all connection-related notifications. When running on localhost, reduce the /chat poller timeout to 15s. ... (Closed-Leaf check-in: 79932a52 user: stephan tags: chat-backoff-timer2) | |
| 06:12 | Improve an error message. ... (check-in: 1f97db77 user: stephan tags: chat-backoff-timer2) | |
| 06:08 | Remove some stray debug output. Make fossil.dom.remove() a bit more robust against passed-in elements which have already been removed from the DOM. ... (check-in: 040a7d45 user: stephan tags: chat-backoff-timer2) | |
| 06:02 | Restore the production-mode (not dev-mode) long-pole timeout. ... (check-in: 8da12695 user: stephan tags: chat-backoff-timer2) | |
| 06:01 | Rework the chat poll fallback considerably. The notification widgets now interact with other messages better and don't mis-fire when the long-poll connection times out. ... (check-in: 9516f7cd user: stephan tags: chat-backoff-timer2) | |
| 00:44 | Add a connection retry back-off delay to /chat when it cannot reach the server. (Edit: too many UI quirks. It needs reworking.) ... (check-in: 4443ef84 user: stephan tags: chat-backoff-timer2) | |
| 00:28 | Make (refine?) the back-off random delay adjustment to be a factor of the current delay. ... (Closed-Leaf check-in: f75fb1dd user: stephan tags: chat-backoff-timer) | |
| 00:16 | Get the /chat-fetch-one handler to lift the back-off delay on success. ... (check-in: d3baf9d5 user: stephan tags: chat-backoff-timer) | |
| 00:08 | Apply the connection-recovery reporting to the other AJAX commands, so that they can trigger the polling to continue if they determine that the connection is back up. Internal cleanups. ... (check-in: fd36f849 user: stephan tags: chat-backoff-timer) | |
|
2025-04-09
| ||
| 22:56 | Remove 'www.' prefix from www.fossil-scm.org links in the footers of several skins. ... (check-in: 67f8111e user: stephan tags: trunk) | |
| 22:37 | If /chat's poller cannot connect to the server, apply a back-off timer so that it does not keep hammering the remote every single second. It attempts to inform the user about outages and when reconnection has succeeded, but it's difficult to test the timing of the the UI elements thoroughly with a single pair of hands, so this is being checked in for dogfooding. ... (check-in: 2debc54e user: stephan tags: chat-backoff-timer) | |
|
2025-04-08
| ||
| 15:58 | Change the user-elevation notification subscription to a user-perms-changed notification, per /chat disucssion. ... (check-in: 36f72c08 user: stephan tags: trunk) | |
| 15:02 | In /setup_ulist, show the email address in the Alerts column so that accounts which have multiple notification subscriptions under different email addresses to not appear as visually-indistinguishable duplicates (they're distinguishable by the parameters for their subscription info link, but not to the naked eye). ... (check-in: a4f5a53c user: stephan tags: trunk) | |
| 12:49 | Ensure that ALL fopen() calls include 'b' in their modes. The git import/export fopen()s did not have this, which may (hypothetically, but unproven) lead to EOL conversion on Windows. ... (Leaf check-in: 13049ef2 user: stephan tags: fopen-mode-b) | |
| 12:43 | Change the message for the 'user elevation' notification to help clarify that it's not necessarily an elevation. Help/message text changes only - no functional changes. ... (check-in: eabccae1 user: stephan tags: trunk) | |
|
2025-04-01
| ||
| 12:16 | Add test-brotli command to get some basic compression measurements. Closing this branch, as brotli is simply too slow for what we want to do. ... (Closed-Leaf check-in: f5b559af user: stephan tags: brotli-compress) | |
|
2025-03-31
| ||
| 21:04 | Replace cgi.c:is_gzippable() with is_compressible() and add support for brotli compression to its decision-making process. There's still a lot to do before we can actually emit brotli compression. ... (check-in: 22e7b78d user: stephan tags: brotli-compress) | |
| 20:26 | Add a basic auto.def check for brotli compression support. The eventual goal of this branch is to add support for brotli-compressed http responses. ... (check-in: a90221b3 user: stephan tags: brotli-compress) | |
|
2025-03-28
| ||
| 14:47 | Fix a ULARGE_INTEGER struct initialization warning on Windows. ... (check-in: 1f0adaa7 user: stephan tags: trunk) | |
| 14:43 | Fix a ULARGE_INTEGER struct initialization warning on Windows. (Edit: moving to mistake branch because msys changed the exec bits of many files.) ... (Closed-Leaf check-in: 263acf16 user: stephan tags: mistake) | |
|
2025-03-27
| ||
| 12:42 | Merge trunk into the merge-info-html branch. ... (Leaf check-in: 1014ffb5 user: stephan tags: merge-info-html) | |
|
2025-03-26
| ||
| 12:41 | Include technote artifacts in the always-has-a-W-card-even-if-empty category in artifact_to_json(). ... (check-in: 655d6130 user: stephan tags: trunk) | |
| 12:30 | In artifact_to_json(), ensure that a P-card is added, independent of the artifact type, if the Manifest object has a P card. Also ensure that some "not quite correct" artifacts, like [45c52e842fc184] (a technote with out-of-order P/N cards), include their W-card. Problems reported in forum post 259a1bdda9bc9ff3. ... (check-in: 0d8c9136 user: stephan tags: trunk) | |
|
2025-03-25
| ||
| 18:13 | Remove +x bit from Makefile.in (not sure when it got added). ... (check-in: f75803fe user: stephan tags: trunk) | |
| 18:06 | Add -p short-form flag of --pretty to test-artifact-to-json. Restore the part of [8af0d05c66e6] which removes extraneous whitespace from artifact_to_json(). ... (check-in: f1847847 user: stephan tags: trunk) | |
| 17:29 | Add -p short-form flag of --pretty to test-artifact-to-json. Restore the part of [8af0d05c66e6] which removes extraneous whitespace from artifact_to_json(). (Edit: moving to "mistake" again. Despite having done "revert" on content.c, the content which should not have been committed was committed.) ... (Closed-Leaf check-in: 3e0d7b06 user: stephan tags: mistake) | |
| 17:18 | Add --json and --pretty flags to the artifact command, both of which emit json with --pretty running it through json_pretty(). Remove extraneous (read: human-friendly) space from artifact_to_json() output. (Edit: moved to "mistake" because this function is not only for emitting "artifacts," but for emitting any type of blob. ... (Leaf check-in: 8af0d05c user: stephan tags: mistake) | |
| 15:24 | Rename artifact_to_json()'s F-card "oldName" property to "rename", per /chat discussion (it's shorter, saving about 1kb of output with the current trunk). ... (check-in: d2db37a7 user: stephan tags: trunk) | |
| 15:13 | In artifact_to_json() add a W-card with a null value instead of eliding the W-card for empty (deleted) wiki pages and forum posts. Similarly, treat an empty H-card as JSON null instead of an empty string. ... (check-in: 87bdf7d1 user: stephan tags: trunk) | |
| 14:45 | In artifact_to_json(), restructure J-cards such that their optional '+' prefix is treated as a boolean 'append' property, not part of their name, analog to how T-card prefixes are handled. ... (check-in: 98396f47 user: stephan tags: trunk) | |