Login
Timeline
Login

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

10 check-ins using file src/fsl.c version 6b9d738381

2023-04-19
02:11
Apply upstream [fossil:7c75e47b3c130ff1]: when merging files with no common ancestor, treat it as if an imaginary common ancestor were an empty file. check-in: b67c69a3ca user: stephan tags: no-common-ancestor
2023-03-07
14:32
Change the anonymous user's default perms from 'hmncz' to 'hz' to match new fossil(1) defaults. check-in: 9a8269d737 user: stephan tags: trunk
2023-02-02
10:08
Integrate upstream patch [https://fossil-scm.org/home/info/52a66829d655ff35dc52]: when cherrypicking, do not include marked-as-private merge parents in the Q-card list. check-in: fc0ae4b07e user: stephan tags: trunk
10:04
Correct FTS index check handling and ensure the cached value of the check is invalidated on open/close of the repo. check-in: 8f60ec4138 user: stephan tags: trunk
2023-01-24
03:21
Add an assert() to the previous checkin. check-in: 65cdfccb1f user: stephan tags: trunk
03:09
Update fsl__search_doc_touch() to be able to handle both FTS4 and FTS5 search indexes, now that the port to FTS5 has started at [fossil:/timeline?r=search-fts5]. This failure was found by trying to make the first checkin on that branch. check-in: 5764b7efa6 user: stephan tags: trunk
2022-11-24
13:51
add FCLI_RC_NO_CMD result code for fcli_dispatch_commands() To help distinguish between invalid commands and propagated FSL_RC_NOT_FOUND errors returned from successfully dispatched commands. ok stephan@ check-in: 46008704a6 user: mark tags: trunk
13:49
make fcli_cmd_aliascmp() return semantics match convention Return 0 if a match is found as one would expect from a *_cmp() routine. ok stephan@ check-in: 6e0d1fbc09 user: mark tags: trunk
2022-11-21
03:30
squelch gcc-11.2.0 uninitialised variable warning If fsl_buffer_cstr2() errors out, nOrig isn't initialised. ok stephan@ check-in: ef78842aea user: mark tags: trunk
2022-11-20
12:20
ensure dstsz-1 is greater than the current length of dst Although somewhat contrived it's a potential problem nonetheless: if the caller passes a destination buffer that contains a string greater than the passed in size of the destination buffer (which is senseless at it means the buffer is already overflowed), we write the full length of src, which could overflow dst. This change makes our fsl_strlcat() implementation behave the same as OpenBSD's strlcat(3) in testing. check-in: f2676cbcc2 user: mark tags: trunk