summaryrefslogtreecommitdiffstats
path: root/bftw.c
AgeCommit message (Collapse)Author
2017-03-16Color link targets for -lsTavian Barnes
Fixes #18.
2017-02-09bftw: Make the nameoff of "///" point to "/"Tavian Barnes
This simplifies a few things such as -name handling for ///.
2017-02-09bftw: Add the DIR* to bftw_stateTavian Barnes
Can't forget to close it that way.
2017-02-08Add support for -x?type with multiple typesTavian Barnes
This functionality is already part of GNU findutils git.
2017-02-07bftw: Add mising closedir() to error pathTavian Barnes
2017-02-06bftw: Plug a leak if dirqueue_push() failsTavian Barnes
If bftw_add() succeeds but dirqueue_push() fails, we need to clean up the just-added dircache_entry. Otherwise it will leak, and we'll also fail the cache->size == 0 assertion. Fix it by extracting the dircache-related parts of bftw_pop() into a new helper function bftw_gc(), and call it from bftw_pop() as well as the bftw_push() failure path.
2017-02-05bftw: Compute nameoff correctly for the root in BFTW_DEPTH modeTavian Barnes
2017-02-05Implement -printf/-fprintfTavian Barnes
Based on a patch by Fangrui Song <i@maskray.me>. Closes #16.
2016-12-18Implement -regex, -iregex, and -regextype/-ETavian Barnes
2016-12-17bftw: Clean up the dirqueue implementation a bitTavian Barnes
2016-12-04Move portability code into util.hTavian Barnes
2016-11-23bftw: Infer the flags in ftwbuf_stat()Tavian Barnes
2016-11-21bftw: Make a defensive copy of the ftwbufTavian Barnes
The callback may modify the ftwbuf, but we check it after the callback (for typeflag and statbuf). Have them mutate a copy instead.
2016-11-21bftw: Always initialize dircache_entry::{dev,ino}Tavian Barnes
If stat() fails, they won't get filled in otherwise. Then cycle detection would have read uninitialized values.
2016-11-21bftw: Make bftw_flags more similar to fts() options.Tavian Barnes
2016-11-14Check for readdir() errors everywhere.Tavian Barnes
2016-11-13bftw: Keep trailing slashes on the root in BFTW_DEPTH mode.Tavian Barnes
2016-11-03bftw: Don't fail just because we couldn't open/read a directory.Tavian Barnes
With BFTW_RECOVER set, we're not supposed to fail just because a single measly directory couldn't be handled. But using state.error as scratch space made us fail in this case. The end result is that #7 resurfaced, so fix it again.
2016-10-24Implement -ignore_readdir_race.Tavian Barnes
2016-10-02bftw: Add support for some exotic file types, where available.Tavian Barnes
2016-10-02bftw: Handle errors from readdir().Tavian Barnes
2016-09-10bftw: Fix do/to typo in a comment.Tavian Barnes
2016-08-24bftw: Initialize typeflag to BFTW_UNKNOWN.Tavian Barnes
It was totally broken on filesystems that spit out DT_UNKNOWN.
2016-05-22dstring: Clean up the API a bit.Tavian Barnes
2016-05-17bftw: Use realloc() to grow the dirqueue.Tavian Barnes
2016-05-17bftw: Remove some debugging counters that were left in accidentally.Tavian Barnes
2016-04-13dstring: Split out the dynamic string logic.Tavian Barnes
2016-02-23bftw: Update at_flags when not following a broken symbolic link.Tavian Barnes
2016-02-23bftw: Plug a leak when the root is not a directory.Tavian Barnes
2016-02-22bftw: Use the currently open directory as at_fd in BFTW_CHILD mode.Tavian Barnes
2016-02-21bftw: Use O_CLOEXEC.Tavian Barnes
2016-02-21bftw: Don't store the terminating '\0' in dircache_entry names.Tavian Barnes
2016-02-21bftw: Use a better cache eviction policy.Tavian Barnes
Instead of simple LRU, we now evict the open entry with the lowest refcount. This reduces the average number of components passed to openat() by a significant margin, and speeds bfs up by about ~5%.
2016-02-20bftw: Shrink the LRU before finding the parent.Tavian Barnes
Otherwise we might close the found parent.
2016-02-19bftw: Clean up dirqueue implementation a bit.Tavian Barnes
2016-02-19bftw: Don't keep DIR*'s around.Tavian Barnes
DIR*'s were being kept around so dirfd(dir) could be passed to future openat() calls. But DIR*'s are big, holding a cache of filenames etc. read by readdir(). Instead, store the raw fd and dup() it to open a DIR* with fdopendir(). This way we can call dirclose() as soon as possible, while still keeping an open fd. Ideally there would be a way to closedir() without invoking close() on the underlying fd, but this is a good approximation. Reduces memory footprint by around 64% in a large directory tree.
2016-02-17bftw: Use a circular buffer to implement the dirqueue.Tavian Barnes
2016-02-14Implement -mount/-xdev.Tavian Barnes
2016-02-13Fix -name handling when the root has trailing slashes.Tavian Barnes
2016-02-13Follow links if appropriate in predicates.Tavian Barnes
2016-02-09Implement -L/-follow.Tavian Barnes
2016-02-06bftw: Don't give up when following a broken symlink.Tavian Barnes
2016-02-04Implement -P and -H.Tavian Barnes
2016-02-04Don't use typedefs to avoid struct/enum tags.Tavian Barnes
2016-01-30bftw: Add missing #include.Tavian Barnes
2016-01-30bftw: Fix fd leak if fdopendir() fails.Tavian Barnes
2015-09-26Optimize -maxdepth in -depth mode.Tavian Barnes
2015-09-26bftw() interface improvements:Tavian Barnes
- Use enums instead of ints where it makes sense - Move the file path inside struct BFTW - Expose a fd and relative path for *at() calls
2015-09-08Add -depth support.Tavian Barnes
The resulting order is fairly weird, as files are still returned in breadth-first order, but directories are returned in a backwards order based on when their reference counts drop to zero. But it's good enough for -delete support.
2015-09-06Factor some code out of bftw().Tavian Barnes
generated by cgit v1.2.3 (git 2.25.1) at 2024-07-16 08:01:38 +0000