Login
Timeline
Login

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

44 check-ins using file f-apps/f-open.c version a9c612558c

2021-03-13
05:22
Added --timer flag to fcli to measure runtime from fcli_setup() until post-main() shutdown. check-in: 2d8c87e253 user: stephan tags: trunk
04:10
Added total/current file counters to fsl_repo_extract_state so that clients can calculate progress percentage. check-in: 356da76d92 user: stephan tags: trunk
04:08
Added s2.Buffer.sha3(). check-in: 44f7331517 user: stephan tags: trunk
01:12
Updated f-s2sh.s2 init script for s2sh2 internals. check-in: 8d497fd57b user: stephan tags: trunk
01:09
Swapped out f-s2sh's shell.c with shell2.c (s2sh2 instead of s2sh classic). check-in: e20cc9720f user: stephan tags: trunk
2021-03-12
19:56
Commented out an assert which prohibits running f-sanity on a non-leaf build. check-in: f2f1612a0c user: stephan tags: trunk
19:32
Merged in cx-scratchpad-rework. check-in: 178f36ed54 user: stephan tags: trunk
19:31
The hashing routines were responsible for the unexpected buffer resizing, as they were explicitly resizing their output buffers to fit. Closed-Leaf check-in: 06f5974d60 user: stephan tags: cx-scratchpad-rework
17:28
Added a new file which had gotten lost in local stash saves/pops. Removed now-incorrect assert(). check-in: 4b0a66807f user: stephan tags: cx-scratchpad-rework
17:26
Completely reworked how fsl_cx manages scratch buffers but it's leading, rather inexplicably, to double the number of allocs and memory use in f-sanity. Checking in due to paranoia of losing this work. check-in: 01c1cff578 user: stephan tags: cx-scratchpad-rework
13:48
Minor API renaming, for consistency. check-in: f2fc1f3efd user: stephan tags: trunk
13:47
fsl_deck_parse() now silently skips over F-cards for which fsl_is_reserved_fn() returns true, as fossil does the same thing since late 2020. check-in: 520d31d80a user: stephan tags: trunk
13:06
Moved some code around and renamed a struct member for consistency with similar structs. check-in: 9bd1cfea4b user: stephan tags: trunk
09:15
Minor refactoring in prep for upcoming changes. check-in: 888cb5cc24 user: stephan tags: trunk
2021-03-11
23:05
Doc updates. Changed default checkin deltaPolicy to 0. Corrected a falsehood about Chuck Norris. Refactored checkout's rm-empty-dirs bits into fsl_rm_empty_dirs(). Fixed mis-reset of fsl_checkout_file_add_opt::counts. Minor malloc optimization in the post-checkout file removal bits. check-in: fd313baef4 user: stephan tags: trunk
18:56
Minor doc updates and added a couple TODOs related removing now-unmanaged files after an rm op. check-in: 419099b063 user: stephan tags: trunk
18:15
Ported f-add's dir-recursion and ignore-glob capabilities into fsl_checkout_file_add() and expanded that function's feature/option set considerably. check-in: 61d3e64e6f user: stephan tags: trunk
15:29
Added fsl_deck_clean2(). Minor iprovements in f-parseparty, most notably to run the leaf rebuild after crosslinking checkins and/or tags, to avoid having (only sometimes) incorrect is-leaf markings (the reason why that happens, and whether it's an actual mistake or a side-effect of unfortunate parse ordering) is unclear). Other trivial internal cleanups. check-in: 6bdb294442 user: stephan tags: trunk
03:58
Minor doc touchups and char-to-bool conversions. check-in: 8a393e5255 user: stephan tags: trunk
03:57
Updated s2 binding for the fsl_deck::F2==>F refactoring. check-in: c7702424bd user: stephan tags: trunk
2021-03-10
17:36
Minor fcli API usage modernization. check-in: c04ce8aaf1 user: stephan tags: trunk
17:00
Minor doc/help text cleanups. check-in: 801690e98c user: stephan tags: trunk
16:41
Refactored fsl_dircrawl_f() to take a fsl_dircrawl_state struct as an argument and added current recursion depth to the passed-on state. check-in: 4a70a0dd17 user: stephan tags: trunk
15:47
Added fsl_dircrawl() for recursively crawling directories and added recursive dir and ignore-glob support to f-add. fsl_buffer_reuse() now returns its argument. check-in: c1199e807f user: stephan tags: trunk
15:41
If fcli_end_of_main() sees that there is a non-0 result code but no message, it synthesizes a message. check-in: 86350ecd42 user: stephan tags: trunk
12:42
Ported in upstream fix [fossil:b491c196c23ec4a240f4]. check-in: 2706788d19 user: stephan tags: trunk
08:54
Corrected missing +x bit in manifests broken in the F-card refactoring. check-in: d0d4c78fb2 user: stephan tags: trunk
08:32
Merged in f-card-array branch. Not the huge performance win i was hoping for but an overall win. check-in: 0626035798 user: stephan tags: trunk
08:25
Renamed the 'F2' F-card temporary porting name back to F. Closed-Leaf check-in: 63e886fff0 user: stephan tags: f-card-array
08:06
New F-card list model is working but the savings for it are underwhelming to say the least. At most, 10-15% alloc count savings on f-parseparty runs, with a slightly higher RAM total. Disabled delta checksum self-test, as it's a big time hit and fossil also disables that by default. (Edit: but this changes cuts the time needed for f-parseparty to process the core fossil repo by a bit more than half. It's still dog slow compared to a fossil rebuild, though.) check-in: 2440c84b01 user: stephan tags: f-card-array
06:20
Lib compiles now but the f-apps, and maybe the other language bindings, are still out of date. Lol - that means i can't check this in with f-ci. check-in: e484d5f96e user: stephan tags: f-card-array
06:02
Initial work on refactoring fsl_deck::F from an array of pointers to an array of objects to speed up (potentially drastically) artifact parsing by cutting F-card (de)allocations by 99% or more. Does not yet compile. check-in: 725c13c0d0 user: stephan tags: f-card-array
2021-03-09
14:03
Fixed a memleak in f-tag. Added --tag flag to f-ci. Tightened validation of tags at output-time (as opposed to parse-time) such that it is no longer legal to add the same tag name/target combination twice in the same artifact, even if they have different types (add/cancel/propagate). check-in: 200bddbb8a user: stephan tags: trunk
06:30
Added a flag to tell fsl_checkin_commit() whether or not to scan vfile for changes, analog to the same option in fsl_repo_checkout(). A few char-to-bool conversions. check-in: 3bff6c7de8 user: stephan tags: trunk
06:06
Merged in confirmer-co branch. check-in: 28bcebca4e user: stephan tags: trunk
06:04
Re-added inadvertently-removed Makefile.in. Closed-Leaf check-in: 9261998d3b user: stephan tags: confirmer-co
06:01
Minor confirmer cleanups and docs. check-in: d58d30707f user: stephan tags: confirmer-co
05:17
It seems we just about have the confirmer bits working for f-co. Had to change the semantics of --keep, but the older semantics feel really weird. Needs a bit more testing. check-in: ed0fa2b56a user: stephan tags: confirmer-co
01:22
Fixed a couple broken error checks. f-co --quiet --verbose now lists only files which actually changed, rather than all files. check-in: 85113ee198 user: stephan tags: confirmer-co
01:07
Removed a stale historical java file. check-in: ba65bbe94e user: stephan tags: confirmer-co
00:13
Minor tweaks of the confirmer bits. check-in: 13fb58bfdd user: stephan tags: confirmer-co
2021-03-08
23:55
Minor cosmetic cleanups. check-in: ce584dc03b user: stephan tags: confirmer-co
23:42
Optimizations in fsl_repo_checkout() to reduce fsl_content_get() calls, only writing files which vfile says have changed. Cuts alloc counts by approx. 3-4x and memory cost by 5-8x in basic tests. f-co now has a blank status square for files which were not modified, to make it easier to spot the real changes. check-in: 735e9e32ab user: stephan tags: confirmer-co
19:36
Initial work on integrating fsl_cx_confirm() with the checkout process. It basically seems to work but has several notable TODOs and FIXMEs. check-in: 6dd8166303 user: stephan tags: confirmer-co