Summary: 582 instances, 513 unique Text Count /// TODO: Fix error handling so it returns an error when the input is 1 # TODO: Handle the error if os.listdir() fails 1 // HgtagsFnodes, // TODO Do we want to support this? 1 // TODO: It may make sense to maintain an existenceChanged bit to distinguish 1 // TODO: Perhaps we should do some short-term negative caching? 1 // TODO: optimize "other_nodes" construction 1 // TODO: Stop using this one-off for Mononoke server, and instead use the one from cmdlib. 1 // TODO: pretty printing of the request details 1 // TODO: ensure none of the params is empty 1 // TODO: Do this for all blobstores that can support unlink, not just SQLBlob 2 // TODO: avoid copies here by switching this to Bytes 1 // TODO: replace with HgParents? 1 // Pushkey, // TODO Do we want to support this? 1 // TODO: Make things more configurable. 1 // TODO: Implement build version lookup, e.g. 1903 1 // TODO: directly operate on the RepoPath components. 1 # TODO: support windows 1 // TODO: One possible optimization here is to set a bit (reference count 1 // TODO: Add method for batch modification, takes iterator of added, removed, changed, or 1 // TODO: Reconsider re-assign master cases. Currently they are ignored. 1 // TODO: newVector.reserve 1 # TODO: swap should not know about histeditrule's internals 1 # TODO: update me? 1 // TODO: 1 # TODO add support for limiting number of keys / values in collections 1 // TODO (T30456231): Avoid this copy 1 // TODO: Would want Scuba and such here. 1 # TODO: unload 1 # TODO add support for limiting size of bytestrings 1 tokio::time::pause(); // TODO: pause only works with the `current_thread` Runtime. 1 // TODO: We need to add an extra layer to absorb all the exceptions generated in 1 // TODO: Expose a constructor for this so that it cannot be constructed wiht seed: None. 1 // TODO: Better yet, this shouldn't use atime at all and instead keep an 1 // TODO: the only state valid when the stream terminates is 1 // TODO: Introduce Metadata bytes? 1 path.clone(), // TODO: get rid of clone (it is borrowed) 1 * TODO: It's perhaps worth doing some small-string optimization here. 1 # TODO implement support for rendering sections, like 1 // TODO: should we maintain a 64-bit SpookyHashV2 checksum to ignore 1 // TODO: We could attempt to restart the process here. 1 // TODO: maybe trigger a background msync 1 # TODO: Consider following renames. 1 # TODO: We do not currently invoke match.traversedir 1 // CheckBookmarks, // TODO Do we want to support this? 1 // TODO: Try to make this into a chain instead of a manual state machine 1 // TODO: Use get_or_insert_owned once stabilized (https://github.com/rust-lang/rust/issues/60896) 1 # TODO: Consider properly templatize the output. So users can choose fields 1 // TODO (T30456231): Avoid this copy 1 // TODO: CHUNK_BY_PUBLIC_POSSIBLE_VALUES 1 # TODO: Move these to async analysis reading from blackbox. 1 # TODO: We should eventually read the default config_dir path from the config 1 # TODO: check if inode_number is actually in the correct subdirectory. 1 // TODO: make this streaming 1 // TODO: If the inode is not currently in cache, we could avoid calling fsync. 1 # TODO: Index this? 3 TODO: If we want to keep the behavior described above for exact matches, we 1 * TODO: If inode is set, this field generally should not be used, and the 1 // TODO: For better throughput on large tree collections, it might make 1 # TODO: avoid this string concatenation if possible 1 // TODO: For minitop, we may want to allow querying for < 1s, but this 1 // TODO: If it were safe to access Subscription::unsubscribe when the lock 1 // TODO: Verify signature on upload token 1 // TODO: deprecate versions 3-5 after capability matching is stable 1 // TODO: On Python 3 the Unicode APIs are smarter: we might be able to use 1 # TODO: Consider migrate to repo.pull to avoid the race 1 // TODO: Can we verify members here (git_path_isvalid) 1 // TODO: Stop using the legacy addStatValue() call that checks to see 1 // TODO: Much of this function could be moved into the .cpp file. 1 # FIXME: Do something here? 1 // TODO: This is exported for testing - is this the right place for it? 1 // TODO: bail on (some forms of) errors? 1 // TODO: Should we avoid promoting if interest is UnlikelyNeededAgain? 1 # TODO: Implement a custom WindowsOutput class that provides color support in 1 // TODO: replace with freeInodes - it's much more efficient to free a bunch 1 # TODO: find a way to block on ssl flushing its buffer 1 # Pick random action. TODO: Instead, prompt user when resolving 1 // TODO: filename_len < 0 == error 1 // FIXME: multibyte? 1 #![allow(deprecated)] // TODO: T29077977 convert from put_X:: -> put_X_be 1 // TODO: if the underlying stores gain the ability to do hash-based comparisons, 1 # TODO: All ignore logic should be encapsulated in the matcher and 1 // TODO: Convert this along with other store calls to impl Storable for 1 // TODO: Count the number of thrift requests 1 // TODO: actually implement this 1 // TODO: contents have changed: we probably should propagate 1 // TODO: Rename, move more functionality to wire types? 1 # avoid cycle (TODO: should be removed in default branch) 1 // TODO: we currently do not have these information stored in Mercurial. 1 # TODO: replace this with Path.readlink once Python 3.9+ 1 // TODO add a checkout decRef counter 1 # TODO: Look at the manifest for the repo 1 // TODO: RocksDB can return a "TryAgain" error. 1 // TODO: load file contents directly from context->store if gitIgnoreEntry is 1 # TODO: Fix for multiple parents 1 # FIXME: This appears to be called by `hg reset`, so we provide a dummy 1 // TODO: We should relax the bounds on cachelib's set_cached. We don't need all of this: 3 // TODO: Uncomment when tracing >= 0.1.18. 1 # TODO we probably need to do env scrubbing? 1 // TODO: Update Metaconfig so we just have this in config: 1 // TODO: consider moving the event to another thread and populating the 1 # TODO: Reword these messages prior to public release 2 // TODO: would be nice to deduplicate these function calls. We need the 1 # TODO: add support for HGPLAIN=+feature,-feature syntax 1 # TODO: These tests do not report the file as modified after truncation 1 # system packages are installed. TODO: For users not defined in the 1 * TODO: we should make this a bit more channel type agnostic. 1 // TODO: It'd be nice to generate those too. 1 // TODO: Add error handling 1 // TODO: add FileMetadata to File and "DirectoryMetadata" to Directory 1 # TODO: Opening a file with O_TRUNC inside an EdenFS mount fails on Windows 1 // TODO: EDGE_TYPE_POSSIBLE_VALUES 1 # TODO: generate a main executable that will invoke Python3 1 // ErrorPushRaced, // TODO Do we want to support this? 1 # TODO: daemon memory/inode stats on line 2 1 // TODO: use runtimeStyle and coordsize 2 // CheckHeads, // TODO Do we want to support this? 1 /// TODO: This does not match the Python implementation. Namely: 1 // TODO: It would be nice to also check for mismatch between 1 // TODO: Branch current matrix so that save/restore has no effect 1 // TODO: root can't be a Leaf 1 self.srcencoding = "utf-8" # TODO: Read from git source projects 1 // TODO: LFS support 1 // TODO: -1 means this part has been interrupted. Handle that 1 // TODO: Consider migrating to async, or get rid of strip in tests. 1 // TODO this needs to check if the unique_ptr is valid 1 // TODO: clean up `is_allowed_user` to avoid this clone. 1 date = datetime.datetime.utcnow() # TODO: date should never be empty 1 // TODO: Also apply permissions changes to the entry. 1 // TODO: figure out what we should do to notify the user, if anything. 1 // TODO: It might be worth maintaining a memory-mapped set to rapidly 1 // TODO: placeholder types are generally `()` 1 // TODO: The "state config" only has one configuration knob now. When 1 // CheckPhases, // TODO Do we want to support this? 1 // FIXME: this cloned! will go away once HgRepo is asyncified 1 # TODO: Break this import cycle somehow. (filectx -> ctx -> fileset -> 1 // TODO: remove clone 1 // TODO: report cs_ids and actual error where possible 1 // TODO: We are discarding expiry and headers here. We probably shouldn't. 1 // TODO: just enforce ASCII-only and make this simpler. 1 // TODO: something reasonable here 1 # TODO: We should probably cache getmeta as well 1 // TODO: Releasing this ObjectPtr here can run arbitrary deleters which 1 // TODO: (jsgf) T17932873 distinguish between not existing vs some other error 2 // TODO: Instead of refreshing tunables every loop iteration, 1 // TODO: attributes in this file aren't renamed to 0, 1, ... 1 // TODO: remove these dynamic casts in favor of a QueryInterface method 1 # TODO: on Windows we replace backslash \ with / since 1 // TODO: We should consider checking if we have in flight BackingStore 1 # TODO: Introduce some sort of defaults config that can select by 1 // TODO: gauges: rx_bytes and tx_bytes; histograms: request_time_ms, response_delay_ms 1 // TODO (T26959816) -- allow Thrift to represent binary as Bytes 1 // TODO: This check should be removed and instead a 1 # TODO remove those users and make this a ProgrammingError. 1 // TODO (T30456231): Avoid this copy 1 # TODO: Just invoke self.normal() here for now. 1 // TODO: implement readFile with pread instead of lseek. 1 # TODO: Our caller still won't be able to distinguish between an empty 1 // TODO: Eventually, we should only serialize the child entry's dtype into 1 # TODO: move to absentfilectx 1 # TODO: Consider GC hidden commits as well and only keep objects used by 1 // TODO: Finish refactor and remove pub 1 // TODO: should these be more structured? e.g. base = p1 some of the time 1 // TODO: Take ownership of key 1 // FIXME: Try to "infer repo" here. 1 * TODO: This field is not updated when an inode's mode bits are changed. 1 // TODO: This function should operate with ObjectId instead of RootId. 1 // TODO: Should we perhaps hold the parentInfo read-lock for the duration 1 // TODO fill this in with data to support takeover on windows 1 # TODO: intelligently stop nuclide-server associated with eden 1 // TODO: NODE_TYPE_POSSIBLE_VALUES, default = DEFAULT_VALUE_ARG, also hide 1 # TODO: perhaps this should be a table of relation functions 1 // TODO: since we call INSTRUMENT_THRIFT_CALL in _globFiles, the time 1 // TODO: add vector support if needed. 1 // TODO: this will unload by atime, atime is not updated by stat -- fix it 1 // TODO: applyToInodes currently forces allocation of inodes for all specified 2 // TODO: it would be nicer to actually check the exception to 1 * TODO: operator->() is not implemented 3 # TODO: Retrieve this data via Thrift. 1 // TODO: log the column names in the structured event 1 # TODO: check whether the hook item is trusted or not 1 # TODO: Make this a dedicated request 1 # TODO: These symlink permissions should ideally be 0o777 1 # TODO: It shouldn't be too hard to turn RocksDB and sqlite3 into optional 1 // FIXME: Investigate why some SQL queries kicked off by the walker take 30s or more. 2 // TODO: be smarter around how failures are handled when reading from the store 1 // TODO: remove cast once we don't use Thrift to represent overlay entry 1 # TODO: clean up 1 # TODO (liubovd): remove the filer once the issue is fixed. 1 # TODO: Finding all EdenFS processes is not yet implemented on Windows 1 // TODO: Once the whole config is hot reloadable, move redaction 1 # TODO HTTP < 1.0 support 1 # TODO: this algorithm has the potential to traverse parts of 1 // TODO: we should consider using a standard memory framework across 2 # TODO: get desired bundlecaps from command line. 1 # TODO: Changing contents of store outside transaction 1 // TODO: Use a separate thread pool to do the git I/O 3 // TODO: (torozco) T43766262 We should embed attempts in retry()'s Error type and use it 1 # flags (TODO: wrapping ?) 1 # TODO: set poroper phase boundaries (affects secret 1 // TODO: It would be nice to add a smarter API to the BackingStore so 1 // TODO: separate the panic handler out from logging 1 # TODO: We potentially should support handling this scenario ourself in 2 // TODO: move `Blobrepo::store_file_changes` logic in here 1 TODO: Remove this once remotenames is default on everywhere. 1 None, # TODO: Use time zone info from commit 2 TODO audit/clean consumers and more clearly define the contents of this 1 # TODO: It would be nice to automatically look up the current 1 // TODO: remove this field, it is no longer used 1 // TODO: It might be worth eliminating the end_ pointer and always adding 1 # TODO: In case not date is given, we should take the old commit date 1 // TODO: possibly enable compression support once this is fixed. 2 # TODO: make a single query to the bundlestore.index 1 // TODO: (rain1) T30973227 parse contents and print out in a better fashion 1 // TODO: It would be nice if we can differentiate not found error with 2 // TODO: Use Vec instead of BTreeMap 1 // TODO: implement secure signed tokens 1 // TODO: We should compress if this is too big. 1 // FIXME: stop cloning when this fn is async 1 # TODO: We should probably rename this argument to --state-dir. 1 // TODO: We are discarding verify actions. 1 # TODO: This won't be reliable for filenames with newlines. We need to parse a 1 File(file::File), // TODO stream 1 // TODO: remove these dynamic casts in favor of a QueryInterface method 1 // TODO: More appropriate status codes here 1 # TODO: Figure out how to remove refs. 1 // TODO: We should at least generate a warning here. We cannot because we 1 Executable(file::File), // TODO stream 1 // TODO: EdenFS for Windows does not support symlinks yet, the only 1 // TODO: remove these dynamic casts in favor of a QueryInterface method 1 # TODO: Check how 'remoteheads' is used at upper layers, and if we 1 # TODO: Maybe we could use some better suggestions here, depending on 1 # TODO: ra.get_file transmits the whole file instead of diffs. 1 // TODO: add support for file wirepacks once that's a thing 2 # TODO: https://github.com/python/typeshed/issues/1240 1 // TODO: This should be done in parallel. Maybe with the new vfs async batch APIs? 1 # TODO: `mkscratch` doesn't currently know how best to place things on 1 # TODO: move this to filectxorabsent 1 # TODO: Add a workingfilectx.write(otherfilectx) path so we can use 1 // TODO: Currently we end up writing out overlay data for TreeInodes 1 // TODO: split the inode into 2 separate copies 1 // TODO: Load the inode in parallel with loading the source control data 1 //! resulting in `HgNodeHash` (TODO: rename HgNodeHash -> NodeId?). This means manifests and files 1 // TODO: we need `or ignore` otherwise we hit primary key violations 1 // TODO: Perhaps we should do some short-term negative 1 // TODO: do we really want a uniform distribution over component chars 1 # TODO add support for limiting size of buffered partial values 1 // TODO (liubovd): improve sqlite query to make it semantically the same 1 // TODO: FUSE_SPLICE_READ allows using splice(2) here if we enable it. 1 # TODO Call this from somewhere 1 // TODO: Should we have further extras? 1 # TODO: clear kernel caches 1 # TODO: register map store offset with 'tr' 1 # TODO: figure out whether "copied" needs to be preserved here. 1 pub p1: Option, // TODO: How hard is it to udpate those? 1 // TODO: There are a few possible optimizations here, but get() is so 1 # TODO we may want to turn this into an abort when this functionality 1 // TODO: We currently don't handle GIT_FILEMODE_COMMIT 1 // TODO: We could store this more efficiently by deduplicating filenode IDs. 1 # TODO print CA cert info 1 // TODO in future, let caller handle receiving on the channel. 1 // TODO: Set a flag in handler_ to reject future subscription requests. 1 // TODO: trace requesting pid 1 // TODO: replace canonicalize() with std::path::absolute() when 36 # TODO: arg checking including path 1 // TODO: move to mercurial_types 1 /// TODO: There is more validation that could be done here. Windows has a broad list of illegal 1 # TODO: get bundlecaps from remote 1 * TODO: It would be easy to extend this function to allocate a range of 1 // TODO: would be nice to log some useful context here! 1 // TODO: follow up and look at the new flags; particularly 1 # TODO: Remove when we can rely on Python 3.7 everywhere. 1 // TODO: audit the path for symlinks and weirdness 1 // TODO: Implement this. 2 # FIXME: Find a way to avoid calling expandargs twice. 1 * TODO: When EdenFS no longer uses hg import helper subprocesses and when 1 // TODO: Might be interesting to add an option to see queuing events. 1 // FIXME: Had to turn it to a vector to avoid "One type is more general than the other" 1 // TODO: When we can rely on C++17, it might be cheaper to move the node 1 // TODO: come up with a better way to maintain these dependencies T77090285 1 * TODO: The mount point rename lock does not exist yet. We need to add it 1 // TODO: Report the pid of the client that requested the first checkout 1 # TODO: Solve the problem instead of working around it. 1 // TODO: This check could occur after the loadedInodes_ assertion below to 1 // TODO: make this more rigorous, e.g. by using params such that p95 = size. 1 // TODO: At the cost of reimplementing parts of Try, we could save a byte or 1 # TODO: check whether path is "trusted" or not 1 # TODO: figure out tests 1 // TODO: we need `or ignore` otherwise we hit primary key violations 1 // TODO: Add unode manifest, fsnode and skmf support 1 # TODO: make sure the sharedpath file points to a valid .hg directory that 1 // TODO: Update this to support stepparents 1 // TODO: If we are doing a force update we should probably short circuit in 1 # TODO: suggest user to archive the old workspace if they want to 1 // TODO: batch request during GC 1 # TODO: Log to some persistent place 1 // TODO: In the future we should probably only allow callers to invoke 2 // FIXME: it's used not only in pushrebase, so it worth moving 1 // TODO: Add requestMetrics for NFS. 1 // TODO: In the future it may be nice to add a sourceValue field that 1 // TODO: test write() when duplicate files and duplicate nodes passed 1 # TODO populate attributes on outgoing instance instead of setting 1 # TODO: Consider changing 'self' to 'reporef()'. 1 // TODO: handle negative values for all the above 1 // TODO: A possible future optimization here is, if a SHA-1 is known when 1 // TODO: will need to encode tree manifests here as well 1 # TODO: fire this inside filesystem. fixup is a list of files that 1 # TODO: Figure out what's the other cases here. 1 // ErrorAbort, // TODO Do we want to support this? 1 # TODO: Consider encapsulating these things in a class so there's only 1 # TODO: There may be some way to approximate an availability 1 # TODO: Add hints about possible matches. 1 // TODO: Use (self.1 + n).min(self.2) once ord_max_min is available (Rust 1.22) 1 // TODO: This should probably check the LocalStore for the blob first, 1 timestamps: Arc>, // TODO: it'd be nice the hooks could be a lifetime. 1 # FIXME: I'd rather just show the help here automatically 1 // TODO (harveyhunt): Pass identities rather than Some(identities) once LFS server has 1 # TODO: perhaps 'ctx' should be dropped from mapping because manifest 1 # TODO: add support for filectx. probably each template keyword or 1 // TODO: check for permissions changes 1 // TODO: NODE_HASH_VALIDATION_POSSIBLE_VALUES 1 // ErrorUnsupportedContent, // TODO Do we want to support this? 1 // TODO: Check using ctx.identities, and deny if neither are provided. 1 // TODO: Handle full paths to the types, such as 1 # TODO: expose debugger-enabled on the UI object 1 TODO: I'd like to either extend this or add a variant that 1 # TODO: Consider moving visibleheads out? 1 // TODO: Review this decision. 1 // TODO: remove clone 1 # TODO: https://github.com/python/mypy/issues/2427 1 None => remap_parents(ctx, &source_cs, self, CandidateSelectionHint::Only).await?, // TODO: check if only is ok 1 // TODO: Do we want to validate the client's Accept & Content-Type headers here? 1 // FIXME: Had to turn it to a vector to avoid "One type is more general than the other" 1 # TODO: perhaps this should be evalfuncarg(), but it can't because hgweb 1 // TODO: FacebookKV for this one? 1 // TODO: Replace this with the 'exitcode' crate once it's available. 1 # TODO: Break up complex function 1 // TODO: Deprecate ScribeLogger and rename this class ScribeLogger. 1 // TODO: directly return stream of streams 1 // TODO: There is probably another thundering herd problem here. If we change a file twice, 1 # TODO: is there a way to just walk the files in the commit? 1 // TODO: produce an error part for (some kinds of?) errors 1 // TODO: This function needs to be extended 1 _ => panic!("TODO: make this an error"), 1 // TODO: This is a bit fragile as a bug in commit_pending not returning a path could lead 1 # TODO: get rid of (potential) inconsistency 1 * TODO: Replace rootID with objectId once objectId is widened. 1 // TODO: It may be worth using a different hash function. The default 1 # TODO: decide use of config section for this extension 1 // TODO: Fix this after updating Hyper: https://github.com/hyperium/hyper/pull/2187 1 // TODO: (rain1) T30970792 parse contents and print out in a better fashion 1 // TODO: NODE_TYPE_POSSIBLE_VALUES 3 * TODO: 32 bytes for an optional, immutable argument string is excessive. 1 # TODO: On Windows we require the Python libraries as well. 1 # TODO this response parsing is probably suboptimal for large 1 // TODO: Once we have EdenAPI production ready, remove this. 1 /* TODO @cuev: This is prob where we'd use args.maxcount: 1 # TODO: We aren't extremely strict about error checking for one-value 1 // TODO: We need to clean up the EdenMount / InodeMap destruction process a 1 // TODO: in the future we should build a better mechanism for controlling this 1 // TODO: check `user_version` and migrate schema if necessary 1 // TODO: maybe use structured error? 1 # TODO: Improve clone protocol so streamclone transfers remote names. 1 # TODO restructure sending of requests to avoid 1 * TODO: This data structure could be implemented more efficiently. But since 1 // TODO: remove when no old clients exists 1 // TODO: move to a bookmarks crate 1 # TODO: if the author and message are the same, don't create a 1 // TODO: Hopefully at some point we will track file sizes in the 1 // TODO: export this from streampager 1 * TODO: It might be beneficial to add a parameter to request more than one 1 // TODO: How to tell whether it is downloading or uploading? 1 rdv_stats.inflight.increment_value(fb, -1); // TODO: This should use a scopeguard... 1 # TODO: We should enforce that services set HG_CALLER_ID to their oncall. We 1 // TODO: 1 // TODO: This code is slightly racy, since threads could receive the signal 1 # TODO: Revisit this after segmented changelog, which makes it 1 // TODO: This might be an error case. 1 // TODO: Can we optimize to reuse the existing PathWithHash we got? 1 // TODO: only convert to ephemeral if a removal took place 1 # TODO: Allow project-specific remote 1 # TODO: deltas 1 // TODO: Do we need this? Why? 1 # TODO: implement something better if necessary and feasible 1 // ReplyPushkey, // TODO Do we want to support this? 1 // TODO: Expose git_path_cmp from libgit2 and use it here 1 // TODO: Missing types, e.g. RedactionKeyList, DeletedManifest, 1 // TODO: rotate/truncate the startup log 1 // TODO: (rain1) T30974137 add a better way to print out file contents 1 # TODO: Once we fully convert to Python 3 we can change this to the much nicer Python 3 1 // ErrorPushkey, // TODO Do we want to support this? 1 // TODO: unload inodes 1 * TODO: To support large files, they should be split into a series of blobs, 1 // TODO: Move hgid to a new DirectoryMetadata struct. 1 # TODO compression options should be derived from bundlespec parsing. 1 # TODO: Move this into a background task that renders from 1 // TODO: clear kernel caches here 1 // TODO: Once we build a new backing store and can replace our 1 * TODO: It never makes sense to rlock an LRU cache, since cache hits mutate 1 // TODO: Update this monitoring code to use FileChangeMonitor. 1 // TODO: NODE_HASH_VALIDATION_POSSIBLE_VALUES, also hide 1 # TODO: crash EdenFS with TreeInode.cpp:3035] Check failed: !newScmEntry->isTree() 1 // TODO: O(N^2) 1 // TODO: Run hooks here (this is where repo_client would run them). 1 // TODO: Optimize this so that the string does not get copied. 1 # TODO: process the subdirectories of data_dir_name 1 # TODO: Validate input repo, e.g. all branches present 1 // TODO: EDGE_TYPE_POSSIBLE_VALUES, default = DEEP_VALUE_ARG, also hide 1 // FIXME: it would also be great to do a manifest diff for old_id 1 # TODO: arg checking 1 // TODO: How should we handle normal non-keyed errors? 2 // TODO: we probably run hooks even if no changesets are pushed? 1 // TODO: Pass through an executor? 1 // TODO: remove entry.getInodeNumber() from both the overlay and the 1 # TODO: 's/svnrev/globalrev' after turning off Subversion servers. We will 1 # TODO: similar to the query string, we should not unescape the 1 // ReplyObsmarkers, // TODO Do we want to support this? 1 // TODO: Support path normalization on case insensitive file systems 1 # TODO: We could potentially try automatically unmounting and remounting. 1 # TODO remove call to nodesbetween. 1 // TODO: benchmark taking a buffer as a parameter 1 # TODO: ancestors should probably be (name, node) -> (value) 1 // TODO: INTERNED_TYPE_POSSIBLE_VALUES 2 // TODO: This needs to be implemented to support umount without the 1 # TODO: replace is needed to workaround a bug in toml 1 // TODO: Move the file into the repair directory, with some unique name 2 * TODO: add maxSize parameter to cause the command to fail if the file is 1 // TODO: Following is broken for curves due to 1 // TODO: We may want to add the Args type defaults are set for and 1 // TODO: A batch API allows us to move to having all nodes have a computed hash without losing 1 // TODO: T45411456 Fix this by teaching the client to expect extra commits to correspond to the bookmarks. 1 # TODO: remove the default once all callers that pass branchmerge=False 1 # TODO: support range, rewrite tests, and drop startdepth argument 1 // TODO: Do not let a negated character class pattern match a "." at the start 1 // TODO: We should customize it by reading the value from the config. This is 1 // TODO: make this configurable on checkout start 1 // TODO: T20282158 clean up access of child inode information. 1 // TODO parse out NodeType from string key prefix if we can. Or better, make blobstore keys typed? 1 // TODO: can the node and flag that are passed in be None? 1 // TODO: Make this an Arc> so we can persist the vfs pathauditor cache 1 // TODO: chunk_len < 0 = error 1 # TODO: Move this into a background task that renders from 1 // TODO: Verify signature on upload token 1 // TODO: Future improvements 1 // TODO: Should we perhaps fall through anyway to report conflicts for 1 # TODO: We eventually want p1, p2, and linkrev exposed on this class, 1 // TODO: generalize this to other listkey types 1 # TODO: get bundlecaps from remote 1 // TODO: This code is also used on durable nodes for the purpose of generating 1 # TODO: change this? 1 // TODO: Use shell_escape once in tp2. 1 // CheckUpdatedHeads, // TODO Do we want to support this? 1 // TODO: add other data (like expiration time). 1 // TODO: add default implementation 4 // TODO: don't send too large or too small chunks to clients 1 // TODO for now this is expected to be the default flag val, but later a Value 1 // TODO make sure that _t is different for each thread 1 # TODO: if we know which new heads may result from this rollback, pass 1 // TODO: Implement 2 // TODO: switch mononoke server to use the same flags as everybody. 1 path.clone(), // TODO: get rid of clone (it is borrowed later) 1 // TODO: (rain1) T30368905 Instead of using this struct directly, have a wrapper around it that 1 // TODO: could print some specific inode values here 1 # TODO: Validate output somehow 1 // TODO (liuz): Rather than issuing one call per filesystem interface, it 1 // TODO: Is it worth implementing relatime-like logic? 1 // TODO: Try to produce copy-info? 1 // FIXME: this code has an assumption that parent with a smaller generation number is a 1 // TODO: Take into account log timings to yield a more faithful 1 // TODO: It'd be nice to have some eviction here. 1 // TODO: Maybe we should walk the inodes and see if we can dematerialize 1 // Bookmarks, // TODO Do we want to support this? 1 # TODO: audit that both source paths have the same inode number 1 // TODO: Review this decision. 1 # TODO: Remove after updating blobimport 2 // TODO: It would be somewhat nicer to look in the ObjectStore and see what 1 // TODO: This could return semi_.isReady() if we also changed thenTry to 1 let in_parent = flags.intersects(StateFlags::EXIST_P1); // TODO: Also check against P2? 1 # TODO: write a test after Python 3 migration 1 # TODO: In the future it would be nice to support pre-compiling the source 1 // TODO: Need to use the same CacheKey here. 1 // TODO: Make this configurable? Perhaps as a global, since it's something that only makes sense at 1 /// FIXME: max_legth parameter is not necessary. We can use .take() method on the stream 1 // TODO: Avoid this clone 1 // TODO: Unit tests for this. We could use a client that lets us do stub things. 1 "run fuse in debug mode"); // TODO: remove; no longer needed 1 // TODO: It might be worth skipping this during ProcessNameCache shutdown, 1 # TODO: only put in cache if pulling from main server 1 // TODO: do not require clone here 1 // TODO: directly return stream of heads 1 // TODO: after finishing these comparisons, update the cached mtimes of files so we 1 # TODO: Does no seem like the best way to do this 1 # TODO: we could just go ahead and build it here, but I 1 // Output, // TODO Do we want to support this? 1 # TODO there are still redundant operations within basefilectx.parents() 1 # TODO: Path normalization 1 // TODO: Need to run hooks on Push, not just Pushrebase 1 // TODO: in the future we can have this look at some configuration that 1 // TODO: It'd be nice to chain those. 1 # TODO: It would be nice to try and get the contents of the 2 # TODO: get rid of this when EdenFS migrates to TreeState. 1 // TODO: this can only be implemented once we have a way 1 // TODO (T104724681): use a nicer representation for a bit set to combine these 1 /// TODO: implement command-based filtering logic 1 // TODO: We probably should add a Cursor::managedClone() function that does 1 // TODO: add an API to DiffCallback to report user errors like this 1 # TODO: treeonly bundlerepo's won't work here since the manifest 1 // TODO: RocksDB can return a "TryAgain" error. 2 # TODO: A \r\n is used 1 TODO consider adding limits as to the maximum amount of data that can 1 // TODO: It would be nice if the .eden inode was created before 1 # TODO: Ideally, this function would inspect the current cgroup for any 1 # TODO: Consider other simple actions such as mode changes 1 # TODO: 's/svnrev/globalrev' after turning off Subversion 1