Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
204 check-ins using file libfossil_VC6.dsp version 3f2f94c84d
|
2021-02-26
| ||
| 06:03 | Some doc touchups, enum renaming, and removed some dead code. check-in: 473fd8718c user: stephan tags: trunk | |
| 04:57 | A workaround for the plink count mismatch between libf/fossil crosslinking. This gets the plink counts correct (fixing fossil's ability to run 'blame'), but why this workaround is needed is unclear. A couple char-to-bool API conversions. check-in: a9332967a7 user: stephan tags: trunk | |
|
2021-02-25
| ||
| 10:14 | Removed a TODO and applied upstream change [fossil:dad521bb06757a7f]. check-in: 998358d8fe user: stephan tags: trunk | |
| 08:25 | Renamed another enum. check-in: 9470d94b1b user: stephan tags: trunk | |
| 08:22 | Some API renaming. Modified how fsl_deck_parse() determines the deck's RID/UUID. Introduced fsl_repo_blob_lookup(). check-in: 467b9e1c18 user: stephan tags: trunk | |
| 07:14 | Removed fossil-content.h - was logically removed a week or so ago. check-in: 95a93ff3b3 user: stephan tags: trunk | |
| 02:48 | Removed some dead code. check-in: b024494e7e user: stephan tags: trunk | |
|
2021-02-24
| ||
| 20:41 | Removed a now-incorrect comment. check-in: ce331bf419 user: stephan tags: trunk | |
| 19:34 | fsl_appendf() %j now emits "null" (always without quotes) when passed a NULL pointer. check-in: b1e8da4fa7 user: stephan tags: trunk | |
| 09:28 | Removed some dead code. Added %j (JSON string) encoding to fsl_appendf() using a streaming approach (no allocation). check-in: a008179c2d user: stephan tags: trunk | |
| 06:00 | f-timeline: renamed event to technote. check-in: ab9b5ace37 user: stephan tags: trunk | |
| 05:42 | Renamed enum fsl_file_perm_t to fsl_fileperm_e and fsl_rc_t to fsl_rc_e. check-in: eb5c621f98 user: stephan tags: trunk | |
| 05:38 | Minor internal cleanups in the crosslinking code. check-in: adefc301a3 user: stephan tags: trunk | |
| 04:04 | Brought checkin crosslinking up to date wrt fossil (whew) and it appears to work: deleting the mlink/plink entries and running f-parseparty -c seems to do the right thing. Fixed a couple mis-handled cases of cherrypick crosslinking and mis-setting of the fsl_card_Q::type field in fsl_deck_parse(). check-in: 4dde2be3b0 user: stephan tags: trunk | |
| 00:03 | Removed some C99-related config options which are no longer needed since we require C99. check-in: 9ffc72c6a3 user: stephan tags: trunk | |
|
2021-02-23
| ||
| 23:55 | Internal doc tweaks. check-in: c7e7b6cf73 user: stephan tags: trunk | |
| 19:45 | Hardened deck validation for certain special cases, in particular at output-time. check-in: a392b75868 user: stephan tags: trunk | |
| 18:09 | Internal doc/#if block cleanups. Stop checking for NULL deck arguments on fsl_deck_XXX() APIs. Changed error code for a few deck-related cases. check-in: 2896b543be user: stephan tags: trunk | |
|
2021-02-22
| ||
| 05:21 | Moved [DbFunctions] wiki page to an embedded doc. check-in: abdf234a5c user: stephan tags: trunk | |
| 05:20 | More char-to-bool conversions. check-in: 6360bb6d8f user: stephan tags: trunk | |
| 04:53 | Removed an extraneous #undef and an unused #define. check-in: beaad1d76d user: stephan tags: trunk | |
| 04:41 | As a precaution against changes made in [7800a90d1699] and [ddf4ac0df508], fsl_content_deltify() is now explicity a no-op if asked to deltify a phantom. check-in: 3affea2904 user: stephan tags: trunk | |
| 03:48 | Added new file missing from [5a3d4bceb3] and enabled search index updates during crosslinking: the mysterious 'SQL logic error' message disappeared after a bit of table-dropping and rebuilding. check-in: 4cbbf66211 user: stephan tags: trunk | |
| 03:35 | Renamed fsl_phantom_t to fsl_phantom_e. check-in: 4880e2e2ac user: stephan tags: trunk | |
| 03:34 | Technote counterpart fix of [7800a90d16]. check-in: ddf4ac0df5 user: stephan tags: trunk | |
| 03:12 | Fixed upstream bug [fossil:ecb705359a58ac32]. check-in: 7800a90d16 user: stephan tags: trunk | |
|
2021-02-21
| ||
| 21:06 | Tinkering with the C++ Deck-to-fsl_deck conversion operator signatures. check-in: e0e62d6f21 user: stephan tags: trunk | |
| 19:07 | Removed an unused var. check-in: f4e8a2184d user: stephan tags: trunk | |
| 18:27 | More enum type renaming. Updated C++/s2 bindings for other recent renamings. check-in: 9c1abb2e95 user: stephan tags: trunk | |
| 18:12 | When reparenting, fail if the given artifact type may not have a P-card. check-in: 1278c47f0a user: stephan tags: trunk | |
| 17:56 | Ported over the crosslinking bits which reparent artifacts via the 'parent' tag. check-in: e7e12b008d user: stephan tags: trunk | |
| 16:52 | Ported in fsl_search_doc_touch() from fossil and call it at the appropriate points in the crosslinking steps, but it's disabled because one of the queries mysteriously fails with 'SQL logic error' when run from this context (but works fine in the fossil shell). Some more char-to-bool conversions. check-in: 5a3d4bceb3 user: stephan tags: trunk | |
| 15:58 | Removed libfossil-DS201x.* project files (MSVC?) because they're out of date and long-since unmaintained. check-in: ce3338ffd3 user: stephan tags: trunk | |
| 15:10 | Renamed FSL_CATYPE... (CA=Control Artifact) to FSL_SATYPE... (SA=Structural Artifact) so that they correspond to the terms as set out by the fossil file format docs. check-in: 8728796601 user: stephan tags: trunk | |
|
2021-02-19
| ||
| 00:05 | Ported in post-dephantomization bits, but they are untested for lack of phantoms (which generally only appear via the sync process). check-in: 2ab8c86b3d user: stephan tags: trunk | |
|
2021-02-18
| ||
| 05:11 | Replaced the extraneous fsl_double_t typedef with double. check-in: b359548620 user: stephan tags: trunk | |
|
2021-02-17
| ||
| 19:43 | Doc additions concerning overriding timeline comment text during crosslinking. Added support for the C-card for wiki timeline updates (noting that that card is never emitted for wiki artifacts by fossil, though it does permit them). check-in: edaf56e488 user: stephan tags: trunk | |
| 03:16 | Replaced a stale fossil-wiki-format link with a markdown one. check-in: fcfc2ae105 user: stephan tags: trunk | |
|
2021-02-15
| ||
| 19:44 | Added FSL_CATYPE_BRANCH_START artifact pseudo-type to give fsl_sym_to_rid() a way to report the start of a branch rather than its most recent RID, in order to match a feature from the main fossil repo (which appears to be unused, but better safe than sorry). check-in: 128adb8515 user: stephan tags: trunk | |
| 19:16 | Resolved (haha) issue with the last round of sym-to-rid updates which left 'trunk' unresolvable. check-in: ae343e941c user: stephan tags: trunk | |
| 18:38 | Updated the symbolic-name-to-rid code to bring it more in line with fossil's and ported in some closely-related behaviors. check-in: 55f8d321d2 user: stephan tags: trunk | |
| 04:13 | Renamed mf.c (mf==manifest) to deck.c. check-in: 799c43f9c5 user: stephan tags: trunk | |
| 04:11 | Fixed the unsetting of the B-card on a manifest when the checkin processes decides that it's not useful as a delta manifest and switches to a baseline manifest. Bug was in the B-card unsetter logic. check-in: 4a015bdd47 user: stephan tags: trunk | |
| 03:24 | Renamed src/fsl_XYZ.c to src/XYZ.c. f-checkin vfile dump now includes renames. check-in: ed3183170a user: stephan tags: trunk | |
| 00:21 | Reimplemented fsl_config_file_reset() to account for schema prefixes. fsl_file_size() now returns type fsl_int_t instead of fsl_size_t, to simplify the -1-on-error return checking. Added fsl_config_global_preferred_name(), which performs the global config file lookup algo used by fossil, and changed fsl_config_open() to use that. check-in: 8837671e83 user: stephan tags: trunk | |
|
2021-02-14
| ||
| 13:17 | Sorted out the broken autosetup: apparently autosetup removed its lib/ in some version, moving those files up one dir, and i didn't notice when upgrading. check-in: 9bc3eb88ca user: stephan tags: trunk | |
| 13:14 | Again autosetup-test-tclsh +x bit: fossil is creating a checkin for this but is not including this file. Seems to be that permissions-only changes, though recognized by status, are causing the checkin to not be build properly. Forced a change in that file to force fossil to add it this time. check-in: 67eed95e19 user: stephan tags: trunk | |
| 13:10 | autosetup-test-tclsh +x bit again. (Wha?) check-in: d793563cae user: stephan tags: trunk | |
| 12:36 | Started replacing some of the char-as-bool with C99 bool. check-in: 4da3578966 user: stephan tags: trunk | |
| 02:41 | Added f-parseparty --skip-unknown flag and have crosslink failure honor the fail-fast setting. check-in: a8af06d834 user: stephan tags: trunk | |
| 02:26 | Improved (slightly) an error report when crosslinking a checkin fails because the content for the P-card artifact cannot be found or is a phantom. check-in: 88947dfafc user: stephan tags: trunk | |
| 01:38 | Renamed fsl_tag_rid() to fsl_tag_an_rid() to help avoid confusion with fsl_tag_id(), which does something much different. Split the crosslinkers into core/timeline-update bits. Fixed a reallocation bug in the crosslink callback list. Added --crosslink option to f-parseparty. Enabled output of empty P-card on checkin manifests with no parents (initial checkin) to please f-mfparse tests (empty P's are harmless, just slightly unsightly). check-in: 63893e6c12 user: stephan tags: trunk | |
| 01:34 | Minor build tweaks. check-in: 06aac2e747 user: stephan tags: trunk | |
|
2021-02-13
| ||
| 23:44 | Minkering/testing with replacing the checkin timeline crosslinker. Moved some code around. check-in: c42cfd8c10 user: stephan tags: trunk | |
| 23:43 | Corrected f-checkin's vfile dump to only include changed files. check-in: 345ab3e1c1 user: stephan tags: trunk | |
| 21:49 | Changed crosslink listener interface and started porting the parts which update the timeline into user-overrideable listeners, which will allow clients to 100% customize timeline content without having to deal with the rest of the crosslinking schema updates. Honor fossil's forbid-delta-manifests setting. Fixed fsl_deck_output() usage in f-apps, as called for by previous commit [ac9b034f79]. check-in: 911afa1bcc user: stephan tags: trunk | |
| 13:03 | Made the fsl_deck object in fsl_deck_output() and crosslink listeners non-const, the former to avoid forcing the client to do much up-front work and the latter because it's necessary for such code to be able to traverse F-cards. That had a bit of a rat's tail of side effects to clean up. Edit: reminder: this caused a sync problem because we added an attachment artifact file which is itself an export of an artifact from the db, leading to a unique constraint violation when pushing. Deleting the attachment table entries, syncing, then rebuilding, resolved it. Hopefully this duplicate artifact won't cause any similar downstream problems. check-in: ac9b034f79 user: stephan tags: trunk | |
| 05:52 | Updated attachment crosslinking to account for technote attachments. check-in: 08ecc42a5f user: stephan tags: trunk | |
|
2021-02-12
| ||
| 22:21 | Started removing the many now-extraneous (fsl_id_t) casts in variadic calls. check-in: 11e5173970 user: stephan tags: trunk | |
| 22:21 | Wha??? doc/presentation-fossil-data-model.md was demonstrably removed (moved to main fossil tree) in [9e08d879fb0b4722], and not re-added since then. How did it re-appear? Edit: i may (not certain) have(???) inadvertently re-added it when crosslinking random manifests during testing. check-in: edc97b0a3a user: stephan tags: trunk | |
| 22:09 | libfossil is now C99, rather than C89, per discussion at [https://fossil.wanderinghorse.net/r/libfossil-forum/forumpost/e02ff0e8d0]. fsl_int{16,32,64}_t have been replaced by the standard uint{16,32,64}_t, FSL_xxx_T_PFMT/SFMT replaced with PRIx32 and friends, and auto.def now checks for -std=c99 capability (but that check could probably be done differently/improved). fsl_int_t and fsl_uint_t are retained, as they're more about documentation/intent than fixed sizes. check-in: ee070353b9 user: stephan tags: trunk | |
| 15:22 | Implemented crosslink begin/end steps, minus still-pending ticket/wiki-specific features. Crosslinking wiki pages updated to use fossil's newer wiki timeline comment format and updated f-timeline to recognize it as well. check-in: 31c8bddf22 user: stephan tags: trunk | |
| 13:00 | Ported in mlink schema fix from [https://fossil-scm.org/home/info/1f5af80047be4589]. check-in: 9cc40bf1be user: stephan tags: trunk | |
| 12:20 | checkin crosslinking: added missing Q-card (cherrypick) handling, corrected a parent-count case when dealing with Q-cards, and generally brought it up to date with regards to fossil(1). check-in: 4dcdf5a8f7 user: stephan tags: trunk | |
| 07:41 | Header file reorgs. check-in: c1681af460 user: stephan tags: trunk | |
| 07:02 | Implemented forum artifact crosslinking and restructured fsl_deck_crosslink() into internal type-specific routines. check-in: 1817557652 user: stephan tags: trunk | |
| 05:03 | Removed java/tcl include paths from the build, as those bits are long-since untested/unmaintained. check-in: 72c2f60034 user: stephan tags: trunk | |
| 04:37 | Fix 'uninitialized use' of rc variable when default switch is hit. check-in: 32d545fd62 user: mark tags: trunk | |
| 03:38 | Removed presentation-fossil-data-model.md - moving it to the main fossil tree. check-in: 9e08d879fb user: stephan tags: trunk | |
| 03:37 | Minor doc updates. check-in: 85c5517be8 user: stephan tags: trunk | |
|
2021-02-11
| ||
| 19:13 | Completely reimplemented ownership semantics of fcli argument/flag memory, greatly simplifying app-side management of those. Now fcli sticks all of them in fcli_fax() so that their lifetime is predictable and managed, so clients no longer have to free them. check-in: 4307886eab user: stephan tags: trunk | |
| 18:38 | Moved around and tinkered with some of the hash policy code. Added f-new --hash option to set the policy when creating a new repo. check-in: dec0670fb2 user: stephan tags: trunk | |
| 13:01 | Replaced a hard-coded UUID in test.cpp with a ref to rid:1. check-in: 9e6df31e88 user: stephan tags: trunk | |
| 12:56 | Fixed a missed link in the fossil-wiki-to-markdown port. check-in: f2f497c658 user: stephan tags: trunk | |
| 12:49 | Checked in the generated fsl_schema_*_cstr.c files to make it easier to port the build to non-Unix systems. check-in: eeac3c6fe0 user: stephan tags: trunk | |
| 12:42 | Moved C++ and s2 bindings to bindings/{cpp,s2}. Got the C++ bindings compiling for new gcc breakage and new, stricter T-card validation. check-in: fdc5e1014e user: stephan tags: trunk | |
| 12:25 | Moved f-tools wiki page to f-apps/index.md. check-in: 827f5c8bc3 user: stephan tags: trunk | |
| 11:29 | Manifest parsing: moved the guess-card-type logic to the end, after the cards have had a chance to apply the type themselves. Ported in some undocumented(?) but code-implemented corner-case constraints, such as not allowing tag names which look like hex hashes. check-in: 4b7e3d1ac6 user: stephan tags: trunk | |
| 06:38 | fsl_deck_parse(): when we know that the current card can only possibly belong to a certain deck type, assign the deck type immediately, rather than delaying it to the "guess" phase (which may actually have already deduced the type, anyway). This provides the same results as before but should be a bit more robust. Removed some dead code. check-in: 92a5e7fee4 user: stephan tags: trunk | |
| 06:14 | Improved fsl_mf_parse() artifact type guessing to eliminate an over-zealous type classification. Implemented the technote swapped P/N-cards bug workaround from the main fossil tree. f-parseparty now includes UUIDs of failed parses and can now parse the whole core fossil (and forum) artifact collection without any errors. check-in: 88060bfb9b user: stephan tags: trunk | |
| 04:42 | Added f-parseparty --types=... flag to select types of artifacts to test. check-in: c0a6cc5b55 user: stephan tags: trunk | |
| 04:41 | Moved some at-exit cleanup code. check-in: d6afb451a0 user: stephan tags: trunk | |
| 03:26 | Merged in make-compdb branch and added test for that feature in auto.def. check-in: 2b7680dd90 user: stephan tags: trunk | |
| 00:16 | Added f-parseparty, which parses all of the artifacts in a repo, and uncovered a serious manifest parsing bug in fossil(1) as a side effect. Fixed a memory ownership bug when manifest parsing fails, which led to a double-free. Related adjacent touchups. check-in: 0163960ce0 user: stephan tags: trunk | |
|
2021-02-10
| ||
| 19:37 | Changed fossil-scm.org/fossil to fossil-scm.org/home, per recent change on that end. check-in: a54f70e148 user: stephan tags: trunk | |
| 18:23 | Stephan's discussion document on the Fossil data model check-in: 8876efccd3 user: dan tags: trunk | |
| 15:46 | Minor doc updates and improved a case of error propagation. check-in: 88f850cd66 user: stephan tags: trunk | |
| 15:45 | Added fcli_fax() (free-at-exit). check-in: e8b7fd94e6 user: stephan tags: trunk | |
| 13:51 | Amend make config files to generate compilation database. Add MAKE_COMPILATION_DB variable to enable the '-MJ' flag on supported compilers to generate a compile_commands.json compilation database. Closed-Leaf check-in: b9be211162 user: mark tags: make-compdb | |
|
2021-02-09
| ||
| 19:42 | Typo fix. check-in: 7c8f09ac3d user: stephan tags: trunk | |
| 18:21 | Re-added autosetup-test-tclsh. Is the +x bit retained now? check-in: b242c6bef3 user: stephan tags: trunk | |
| 18:20 | Temporarily removing autosetup-test-tclsh to try to work around the ability to +x it. check-in: c6d06e4728 user: stephan tags: trunk | |
| 18:18 | Wha??? Added +x bit again - apparently libf's checkin process does not catch attribute-only changes properly? Edit: this checkin was made with fossil(1), but same thing. check-in: dab3ac63d1 user: stephan tags: trunk | |
| 18:17 | Added +x bit again - apparently libf's checkin process does not catch attribute-only changes properly? check-in: cb3d4e371d user: stephan tags: trunk | |
| 18:15 | Added +x bit (wasn't this already done?). check-in: 7dab52fd0e user: stephan tags: trunk | |
| 18:08 | Merged inadvertent fork. check-in: dbfab53142 user: stephan tags: trunk | |
| 18:06 | Fixed creation of new repository (fsl_repo_new() and f-new.c). Added checkin message mimetype to repo creation options. check-in: 162d4ba4c8 user: stephan tags: trunk | |
| 11:40 | Added some protections against generating empty checkin. Removed some stale docs. check-in: 9bff657291 user: stephan tags: trunk | |
| 11:35 | finalise licenses dir / check-in: 6347de6f38 user: dan tags: trunk | |
| 11:22 | add license directory for files to point to check-in: 2aa8c19588 user: dan tags: trunk | |
| 11:10 | Added +x bit. check-in: 27e1f8ee2b user: stephan tags: trunk | |
| 09:52 | Removed some stale docs. Add new checkins to the UNSENT table (though whether or not they're in there appears to make no difference?). check-in: a52c8a0690 user: stephan tags: trunk | |
| 09:49 | Added a fsl_cx flag to silently skip, instead of error out, when crosslinking encounters a type it cannot yet handle (currently forum posts and tickets). check-in: a2c0c7820d user: stephan tags: trunk | |
| 08:54 | Re-added sqlite3ext.h (needed by the vtable bits). Added if-not-defined guards around the generated config header vars. Minor touchups to the amalgamation build. check-in: 62099c7d53 user: stephan tags: trunk | |
| 08:53 | Fixed a null pointer deref in a routine i thought never got passed NULLs. check-in: d826ff335f user: stephan tags: trunk | |
| 05:07 | Explicitly insert vfile.mhash with a NULL value. check-in: ac52d23368 user: stephan tags: trunk | |
| 05:06 | Implemented FSL_HPOLICY_SHUN_SHA1. check-in: 9418c5e60f user: stephan tags: trunk | |
| 05:03 | Removed an unnecessary var-guard. check-in: 1d9b5aa3c0 user: stephan tags: trunk | |
| 04:25 | Minor doc updates. check-in: 0d0a30f979 user: stephan tags: trunk | |
| 04:23 | Versioned file access no longer follows symlinks. check-in: d8a6788d85 user: stephan tags: trunk | |
| 00:18 | Updated license text in almost all C/H to claim copyright while attributing all of the files as heavily derived from DRH's work. check-in: 631fb3f69a user: stephan tags: trunk | |
|
2021-02-08
| ||
| 23:48 | Added forum type to f-timeline. check-in: c2d78cbb4b user: stephan tags: trunk | |
| 23:47 | Renamed the hash length constants. check-in: 0c9a6d3644 user: stephan tags: trunk | |
| 23:42 | Updated s2 bindings for new forum-deck cards. check-in: 05542c5251 user: stephan tags: trunk | |
| 23:35 | Merged in [sha3-port] branch. There's certainly still bits to test, but it now more or less supports both SHA1 and SHA3-256 hashes. check-in: ff13fc9b52 user: stephan tags: trunk | |
| 23:32 | Added deck support for forum posts, except that crosslinking them is currently unimplemented (requires more infrastructure). Closed-Leaf check-in: b0ac77411e user: stephan tags: sha3-port | |
| 20:45 | MILESTONE: f-ls just read a file listing from the main fossil repo (v2 hashes). Set f-mfparse back to hard-coded sha1 for the time being, as i'm not sure i can automatically figure out one or the other hash (need to add a flag to specify which to use). check-in: 8367e0ce6e user: stephan tags: sha3-port | |
| 20:35 | Ported in more of the fossil(1) hashing-related infrastructure. check-in: 033d78f03f user: stephan tags: sha3-port | |
| 18:23 | f-apps compile and f-sanity runs. MILESTONE. Still need to test the other f-apps, though. check-in: 76c290eb05 user: stephan tags: sha3-port | |
| 16:49 | Reverted autosetup to 0.6.9 to [https://github.com/msteveb/autosetup/issues/59]. check-in: 9538db33d6 user: stephan tags: sha3-port | |
| 16:42 | th1 was removed by stephan check-in: cc0bac7250 user: dan tags: trunk | |
| 16:03 | Baby steps towards dual sha1/sha3 support. This does not yet compile but i'm nervous about losing work done so far. It's been painless but tedious, which is not bad, all things considered. check-in: 240701a6ec user: stephan tags: sha3-port | |
| 13:34 | Apply the sqlite3.c CPPFLAGS currently used by fossil(1). check-in: f16a3dd322 user: stephan tags: trunk | |
| 07:44 | Brought core SQL schemas up to date wrt fossil(1). check-in: 490dadcad9 user: stephan tags: trunk | |
| 07:43 | Updated autosetup. check-in: 48120a033a user: stephan tags: trunk | |
| 06:53 | Resolved the s2 unit test breakage: config.h settings led to it not stat'ing a dir entry the same way the core s2 tree was, which revealed a slight bug in s2.import's is-a-dir check (namely, not recognizing a dir as a dir). check-in: 17ac91740c user: stephan tags: trunk | |
| 06:15 | Updated s2 and brought old fossil bindings up to date wrt cwal/s2 changes made since then. Core s2 unit tests are inexplicably failing with a bogus(?) OOM report with this shell build, but they work in the core s2 tree and libf unit tests are passing. check-in: f063fd4f0f user: stephan tags: trunk | |
| 06:11 | Fixed a swapped value check in fsl_cx_db_name_for_role(). check-in: a6a15065e1 user: stephan tags: trunk | |
| 04:49 | Removed long-since-obsoleted th1ish. check-in: 36e2410ca1 user: stephan tags: trunk | |
| 04:48 | Undefined all sha1-hardened-related defines at the end of that code, for the sake of an eventual amalgamation build. check-in: 6ce8ec4fef user: stephan tags: trunk | |
| 04:31 | Ported in hardened SHA1 hash and set default build config to use that one instead of standard SHA1. check-in: 0f3f1d2bf2 user: stephan tags: trunk | |
|
2021-02-07
| ||
| 14:39 | Back-ported diff numeric overflow fix from [https://fossil-scm.org/fossil/info/e2b7dca948da84b7]. check-in: f11bfb3bcf user: stephan tags: trunk | |
| 14:31 | Globally replaced http://... URLs with https://.... Most of them probably still work, too. check-in: 6af7b3b7c8 user: stephan tags: trunk | |
| 14:27 | Updated Doxygen.in and remote doc path. check-in: b78d59d054 user: stephan tags: trunk | |
| 13:39 | Updated sqlite3 for C99-ism fix. Removed sqlite3ext.h. check-in: 601041d827 user: stephan tags: trunk | |
| 10:45 | Ported in core SHA3 hash routines from fossil(1). check-in: edf64345bb user: stephan tags: trunk | |
| 10:32 | Updated build for newer gcc and sqlite3.c, in prep for a potential upcoming effort of getting the hashing bits updated for fossil v2. check-in: 451d0b8a9b user: stephan tags: trunk | |
|
2018-12-01
| ||
| 15:59 | i know this isn't maintained anymore, but found a broken assert() while porting the db wrapper into another tree. check-in: 5c9c84f81d user: stephan tags: trunk | |
|
2016-08-26
| ||
| 15:15 | latest s2 for ARM platform fixes. check-in: e961eb1201 user: stephan tags: trunk | |
|
2016-08-20
| ||
| 12:37 | fixed a mis-think in the previous commit which caused unnecessary extra allocations. check-in: 679e38294a user: stephan tags: trunk | |
| 12:13 | fixed: when parsing F-cards, defossilize their names (in-place) before insertion into the deck. Was broken with space-containing filenames. check-in: 9f317f8fa8 user: stephan tags: trunk | |
| 11:41 | s2 update. check-in: b7da6be617 user: stephan tags: trunk | |
|
2016-03-14
| ||
| 17:21 | minor s2 update. check-in: e9556c1886 user: stephan tags: trunk | |
| 17:08 | latest s2 for proc() using(...) {...}. check-in: daa02e9344 user: stephan tags: trunk | |
|
2016-03-07
| ||
| 15:57 | added FSL_OPEN_F_TRACE_SQL for use with fsl_db_open(), which tells fsl_db_open() to enable SQL tracing. Previously this feature was restricted to Fossil-open()'d db handles. check-in: a43aece996 user: stephan tags: trunk | |
|
2016-03-06
| ||
| 10:52 | latest s2. check-in: 633707c54a user: stephan tags: trunk | |
|
2016-03-05
| ||
| 21:42 | Internal cleanups in the Fossil.Db.open() binding. And this time passed the -v flag to f-tag. check-in: f15fa6ba7c user: stephan tags: trunk | |
| 13:55 | fixed broken return vals from fsl_db_existsv(). check-in: f6f9628a30 user: stephan tags: trunk | |
|
2016-02-29
| ||
| 17:41 | Rewrote the "can call dlopen() without -ldl" auto.def test so that it will fail on Linux (thereby continuing on to the next clause, which handles the Linux case) while still doing the right thing on FreeBSD. check-in: 94468be8ca user: tangent tags: trunk | |
| 14:04 | removed cgimod bits (moved them to the s2 module repo). Some CGI-related scripts kept around for later tinkering. check-in: d4b02ee824 user: stephan tags: trunk | |
| 13:16 | more -ldl flag tweaking - this is likely to break s2 on BSD again (where -ldl is not needed), but the current test for "dlopen is built-in" incorrect reports success on my linuxes then does not work at link-time. check-in: 008a3a3208 user: stephan tags: trunk | |
| 13:01 | fixed a syntax error not reported on my odroid box b/c that block was not hit. check-in: c90823b297 user: stephan tags: trunk | |
| 12:43 | straightened out the module loading flags. fixed breakage in f-s2sh.s2 caused by s2 making "import" a reserved word. Removed th1ish leftovers from cleanup rules. check-in: 2e4dd625ab user: stephan tags: trunk | |
| 11:58 | latest s2, including %d patch from Warren. Disabled loadable modules temporarily due to a local system hiccup. check-in: efd0f500fe user: stephan tags: trunk | |
| 11:41 | merged in os-x-fixes branch from Warren. One of those changes will be overwritten by a pending s2 update, though. check-in: 61d8fbeee2 user: stephan tags: trunk | |
| 11:29 | merged in freebsd-10-fixes bundle from Warren. check-in: a6b8958efc user: stephan tags: trunk | |
|
2016-02-28
| ||
| 19:34 | Added s2_amalgamation.o to the link step for cgimod.so, without which you get a bunch of undefined symbols, at least on OS X. Closed-Leaf check-in: 781346a2e4 user: tangent tags: os-x-fixes | |
| 19:30 | Removed GCC-specific -export-dynamic flag in f-s2sh's Makefile, per mailing list discussion. check-in: df42433396 user: tangent tags: os-x-fixes | |
| 19:29 | Squished printf() format specifier warning in s2 amalgamation, per mailing list discussion. (It really needs to be fixed in s2, but I need it fixed in *this* tree in order to continue work, so...) check-in: 4d5d7ace65 user: tangent tags: os-x-fixes | |
| 19:24 | autoconfigure now checks for iconv(3) in libc and libiconv; this is needed on OS X at least, and may also be needed on older SUS type Unixes. (libiconv comes to SUS via XPG4, and thence from HP/UX.) We check for it in libc first because it's built into the standard C library in glibc and FreeBSD's libc, at least. Also, added $(LDLIBS) to the link flags in shakenmake, without which the above fix has no effect. check-in: cd6b27805b user: tangent tags: os-x-fixes | |
| 17:39 | Removed -static-libgcc flag from shakenmake.make, since that's GCC-specific, breaking the build under Clang, and presumably other C compilers. It only allows fully-static binary builds, which is not a generally-useful thing anyway. (Consider the recent glibc DNS bug. (CVE-2015-7547) Do you really want to rebuild every binary on the system to fix it, or just upgrade glibc?) check-in: 00f219f225 user: tangent tags: os-x-fixes | |
| 17:15 | Fixed a fossil_strdup() -> fsl_strdup() hiding in an ifdef that's true on OS X only. check-in: a571a60261 user: tangent tags: os-x-fixes | |
|
2016-02-23
| ||
| 19:58 | Added -fPIC to CPPFLAGS in s2/Makefile, without which it won't link on FreeBSD Closed-Leaf check-in: 164d5552e8 user: tangent tags: freebsd-10-fixes | |
| 19:47 | Swapped the two dlfcn.h detection cases in auto.def. Previously, if it found that header it would then go blindly chasing after libdl. But on FreeBSD, the header exists and libdl does not because those functions are built into libc. The second case checked for this, but because dlfcn.h was found, it was never even tried. The new way tries to build against the dl functions without libraries *first* and only then if that fails does it go chasing after libdl. check-in: 870a29ea5d user: tangent tags: freebsd-10-fixes | |
| 19:15 | Updated _XOPEN_SOURCE definition from 500 to 600, as that's needed to allow lstat(2) to be visible under FreeBSD 10. check-in: 2755eb8768 user: tangent tags: freebsd-10-fixes | |
|
2016-02-18
| ||
| 16:00 | latest s2, fixed an error handling case in Fossil.Context.loadManifest() (caught by an assertion). check-in: a4b5749230 user: stephan tags: trunk | |
| 11:24 | fixed an assertion, improved an error message (just now triggered by files with spaces in their names :/). check-in: a5d8233f6d user: stephan tags: trunk | |
|
2016-02-17
| ||
| 21:27 | Latest s2 and a Happy 8th Birthday to Baako. check-in: 56901c5b0a user: stephan tags: trunk | |
|
2016-02-12
| ||
| 15:38 | more tinkering with q-ancestry.sql. check-in: c5dc22c1d2 user: stephan tags: trunk | |
| 10:32 | typo fix. check-in: 59fe7b70ce user: stephan tags: trunk | |
|
2016-02-10
| ||
| 16:33 | Removed extended s2.Buffer unit test script because this feature is now supplied by the s2 core and its unit tests. (Make sure to rm s2/UNIT*.s2 or 'make unit' may still try to import it.) check-in: 10da5cf074 user: stephan tags: trunk | |
| 16:30 | Updated s2 and related build bits, removed the libfossil-injected s2.Buffer.compress() and friends because those have been ported over to the s2 core. check-in: a25251e6b0 user: stephan tags: trunk | |
|
2016-01-31
| ||
| 17:54 | latest s2, re-enabled test combinations disabled when string interning was recently turned off. check-in: 6085b8689d user: stephan tags: trunk | |
|
2016-01-28
| ||
| 12:40 | minor script updates. check-in: 1cffc35efe user: stephan tags: trunk | |
| 12:29 | sqlite3 version 3.10.0. check-in: 99efd91396 user: stephan tags: trunk | |
|
2016-01-27
| ||
| 09:31 | s2 update to disable string interning because how s2 manages values is (occassionally) incompatible with interning, leading to cwal-level memory misuse and assertions getting triggered. check-in: f9eba44f38 user: stephan tags: trunk | |
|
2016-01-24
| ||
| 18:43 | okay, last s2 for a while. check-in: eb54d84520 user: stephan tags: trunk | |
| 02:05 | latest s2, lots of minor tweaks. check-in: 8a4665bffa user: stephan tags: trunk | |
| 02:04 | an admittedly untested changed to etPATH handling ('%/' specifier) tweak in fsl_appendf(), to avoid allocating for 0-length strings. check-in: 965a326ce2 user: stephan tags: trunk | |
|
2016-01-22
| ||
| 15:31 | added a missing return on error. check-in: c95de406e5 user: stephan tags: trunk | |
|
2016-01-13
| ||
| 16:32 | minor doc correction. check-in: 754f74c4b9 user: stephan tags: trunk | |
|
2016-01-09
| ||
| 18:17 | MOAR latest s2. check-in: efb9c69d75 user: stephan tags: trunk | |
|
2016-01-08
| ||
| 17:10 | latest s2, accommodated API change. check-in: 2827317748 user: stephan tags: trunk | |
|
2016-01-07
| ||
| 21:11 | replaced impl_fsl_buffer_swap with fsl_buffer_swap. check-in: 51bafea46e user: stephan tags: trunk | |
| 20:57 | merged in dave's old branch. The related strangely failing tests seem to work now :). Also the latest s2. check-in: feb9f32d06 user: stephan tags: trunk | |
|
2014-11-21
| ||
| 14:57 | perverse over-use of comma operator in conditional of while statement boggled valgrind's mind; converted to equivalent steps in body with a break; to please it. It's more understandable this way by humans as well. Closed-Leaf check-in: 3104155e44 user: dave tags: feech-VerSetVtblXtnCbk | |
|
2014-11-20
| ||
| 15:02 | altered 'change notification' printf parameters from %x to %p in hopes of getting some 32/64 bit compatibility for cheap. We'll see if this works and if it's xplat.... check-in: 1660df6669 user: dave tags: feech-VerSetVtblXtnCbk | |
|
2014-11-10
| ||
| 19:31 | (gcc complaint) check-in: 42a790e72e user: dave tags: feech-VerSetVtblXtnCbk | |
| 19:30 | fix parsing bug for hex address; change notification order to be after creates/updates, and before deletes. inhibit 'dtor' notification for the time being; it seems to get invoked twice for some unknown reason (like sqlite has two copies of the vtable object, or something). check-in: e80ed55cdf user: dave tags: feech-VerSetVtblXtnCbk | |
| 18:58 | changed data separator for CHNGCBK to ':', because ',' would interfere with processing. added stub callback implementation in fsl_cx check-in: a022b2905b user: dave tags: feech-VerSetVtblXtnCbk | |
|
2014-11-08
| ||
| 19:02 | one more try, sigh, to please gcc... check-in: 860f151edd user: dave tags: feech-VerSetVtblXtnCbk | |
| 18:41 | ...and gcc didn't like it of course, so let's pleasure it check-in: a1fdbeebad user: dave tags: feech-VerSetVtblXtnCbk | |
| 18:01 | interim checkin; let's see what gcc thinks of all of this... check-in: 9352ddfbad user: dave tags: feech-VerSetVtblXtnCbk | |
|
2014-11-07
| ||
| 17:10 | Create new branch named "feech-VerSetVtblXtnCbk" check-in: 086c180778 user: dave tags: feech-VerSetVtblXtnCbk | |
| 17:08 | transaction support of versioned settings passed all tests, so integrating into 'dave'. Will make new feature branch for 'optional change callback' support. Closed-Leaf check-in: 616500affb user: dave tags: dave | |
|
2014-11-05
| ||
| 00:11 | bug in 'rollback' processing; when encountering an unmodified setting, would not skip it, and would infinite loop Closed-Leaf check-in: 35d3266b60 user: dave tags: feech-settingsvirtualtabletransactionsupport | |
|
2014-11-04
| ||
| 21:41 | more exciting mundane changes for gcc check-in: 6c4afc0b2d user: dave tags: feech-settingsvirtualtabletransactionsupport | |
| 21:38 | mundane changes for gcc check-in: 57c852f9cb user: dave tags: feech-settingsvirtualtabletransactionsupport | |
| 21:29 | implementation of transaction support complete; now it's testing time... check-in: 7aecc393a7 user: dave tags: feech-settingsvirtualtabletransactionsupport | |
| 18:33 | algorithm for commit and rollback should be code complete; now need to modify xUpdate to use it. check-in: fcbede55ed user: dave tags: feech-settingsvirtualtabletransactionsupport | |
| 15:00 | interim; fleshed out algorithm impl, and made marks where surgery will occur check-in: 2864314d5f user: dave tags: feech-settingsvirtualtabletransactionsupport | |
| 14:55 | Create new branch named "feech-settingsvirtualtabletransactionsupport" check-in: e257ccc652 user: dave tags: feech-settingsvirtualtabletransactionsupport | |
|
2014-11-03
| ||
| 23:53 | interim; added members to internal data structures to facilitate transaction support on versioned settings v-table check-in: 120add9960 user: dave tags: dave | |
| 23:21 | dev studio project updates for moved headers check-in: ef81d5cb7c user: dave tags: dave | |