tests: skip filelog damage tests when not using revlogs Differential Revision: https://phab.mercurial-scm.org/D3042

tests: skip some tests when using simple store generaldelta is a revlog implementation detail. We don't need to test it with the simple store. union repos are heavily revlog based. It should be possible to run them with alternate stores. But it's not worth the trouble at this juncture. Differential Revision: https://phab.mercurial-scm.org/D3040

tests: conditionalize tests based on presence of custom extensions The test harness supports injecting extensions via --extra-config-opt. However, if you do this, various tests that print state about loaded extensions fail. This commit teaches the test harness to recognize when custom extensions are loaded so that tests can use feature sniffing to conditionalize tests based on that. Differential Revision: https://phab.mercurial-scm.org/D3039

simplestorerepo: avoid shadowing dict in list comprehension over dict Caught by pyflakes. Differential Revision: https://phab.mercurial-scm.org/D3077

narrow: pass node into revlog.revision() This is one of the few (possibly only) places where we pass an int in for filelogs. Other revlogs (notably changelog) are very heavy on int usage. But filelogs are surprisingly node centric. I'd like to formalize the interface around the use of nodes (at least for filelogs). So let's switch to a node. We can't inline revlog.node() because of a check-code rule. I think that rule is suspect. But it may be for performance reasons with changelog code. I'd rather not touch it at this time. Differential Revision: https://phab.mercurial-scm.org/D3043

changegroup: remove "revlog" from error message The previous message leaked an implementation detail. Differential Revision: https://phab.mercurial-scm.org/D3041

tests: conditionalize tests based on presence of revlogs for files ~85 tests don't like our non-revlog file store for various reasons. This commit introduces hghave functionality for declaring and querying repository features. By default, we assume repositories have revlog-based file storage. But if the HGREPOFEATURES environment variable is set, we can override the default set of repository features. If you run the test harness with our simplestorerepo extension and an environment variable set to the proper value, you can override the hghave defaults to agree with simplestorerepo's version of reality. Various tests have been modified so behavior dependent on revlog-based file storage is marked as such. This fixes a handful of test failures with our custom file storage extension. But dozens remain. The point of this commit is to demonstrate how tests will need to be modified to account for custom storage implementations. TBH, I'm not convinced hghave is the proper layer for repository feature detection. I /think/ we'll eventually want something in run-tests.py itself. But that would require inventing a new primitive in the test harness. This is all very alpha at the moment. So I think hghave is an acceptable place to hang this feature detection. I think the right time to be thinking about integrating this into run-tests.py is *after* we have a stable alternate storage implementation in core. For now, let's try to make progress towards the idea of an alternate storage backend. Differential Revision: https://phab.mercurial-scm.org/D3030

tests: add test extension implementing custom filelog storage In order to better support partial clones, we'll need alternate repository storage mechanisms that aren't based on revlogs. Today, the interface for repository storage isn't very well defined. And there are various layering violations and assumptions made throughout the code that storage is backed by revlogs. In order to support alternate storage mechanisms, we'll need to formally declare and adhere to interfaces for storage. This will be a long, arduous process. This commit creates an extension that implements non-revlog storage for files. It defines a custom type that quacks like the existing revlog/filelog API but isn't backed by a revlog. The backing storage is - for simplicity reasons - a CBOR index and per-node files representing fulltext data. The localrepository class is modified so file(f) returns instances of this class instead of filelog instances. The purpose of this extension is to tease out what the actual filelog interface is - based on running the test harness - so we can formalize that interface and then implement a *real* alternate storage backend. Using `run-tests.py --extra-config-opt` to run the test harness with this extension enabled yields 83 failures out of 634 ran tests. The most common test failures are due to: * Issues with `hg verify` * LFS and largefiles (probably flags processing related) * Narrow. * Any test touching or inspecting individual filelog paths. * help and error output that is confused by the presence of an extension. * `hg debug*` commands doing low-level, revlog-y things. An 88% pass rate is pretty good for an initial implementation if you ask me! There is a bit of duplicate code in the new extension. That's by design: a point of this code is to tease out dependencies on revlog. That being said, there is opportunity to consolidate code by moving things out of the revlog API. For example, DAG traversal operations don't necessarily need to be implemented at the storage level. (Although for performance reasons they probably do.) Once we have a more well-defined interface, we could probably define the default implementations in terms of the base interface, pull those in via class inheritance, and have implementations override with faster versions if they so choose. (Or something like that.) But for now, the duplicate code should be acceptable. Differential Revision: https://phab.mercurial-scm.org/D3029

tests: use `hg unbundle` instead of `hg pull` in some tests `hg pull <bundle>` uses the special "bundlerepo" repository. The bundlerepo code makes many assumptions about the storage of repositories. It will be difficult to teach bundlerepo to use non-revlog storage before a better storage interface is established. Many test failures using our "simple store" are related to bundlerepo: the simple store just isn't compatible with bundlerepo because of storage assumptions in bundlerepo. In order to mitigate the impact of bundlerepo on our code base, this commit changes various tests to use `hg unbundle` instead of `hg pull`. This bypasses the bundlerepo code. Tests exercising exchange functionality have not been altered, as they should be using `hg pull` and going through the bundlerepo code paths. Differential Revision: https://phab.mercurial-scm.org/D3059

hgweb: use revsymbol() for creating context from changeid These seem to be for looking up a revision that can come from the user, so revsymbol() is the right method to call (0194dac7 has more information about my plans for repo[x]). Differential Revision: https://phab.mercurial-scm.org/D3075

hgweb: inline changeidctx() Differential Revision: https://phab.mercurial-scm.org/D3074

hgweb: drop support for "manifest" parameter AFAICT (but note that this is the first time I look at hgweb code), the "mercurial" query parameter was removed from rendered pages in 36fa5db79dd5 (hgweb: convert gitweb to NWI, 2006-10-05). Search for "manifest=" in that diff to see why I think it was removed. It's about time we stop looking for the parameter in requests. Differential Revision: https://phab.mercurial-scm.org/D3073

rebase: use single transaction when running in memory rebase.singletransaction make rebase noticeably faster (~20% in a test I just ran). It is not enabled by default because it risks losing information if it aborts (see `hg help rebase`). When running rebase with the experimental in-memory option on, rebase is first attempted in memory, and if any conflicts occur, it restarts, this time writing to disk. Thus, it should be safe to turn on single-transaction mode for the in-memory phase. Differential Revision: https://phab.mercurial-scm.org/D3076

bundlerepo: use super() when calling file() We should be calling the default method, not reimplementing it. Differential Revision: https://phab.mercurial-scm.org/D3058

tests: remove superfluous config setting format.usegeneraldelta defaults to true. Differential Revision: https://phab.mercurial-scm.org/D3057

templater: mark .joinfmt as a private attribute

obsutil: make obsfateprinter() less dependent on templater joinfmt() is defined as 'lambda x: scmutil.formatchangeid(repo[x])' in showsuccsandmarkers(). Function arguments are reordered so they look more normal.

templater: abstract away from joinfmt Future patches will add a wrapper for a list of template mappings, which will implement a custom join() something like {join(mappings % template)}. The original join() function is broken down as follows: if hasattr(joinset, 'joinfmt'): # hybrid.join() where values must be a list or a dict joinitems((joinfmt(x) for x in values), sep) elif isinstance(joinset, templateutil.wrapped): # mappable.join() show() else: # a plain list, a generator, or a byte string; joinfmt was identity() joinset = templateutil.unwrapvalue(context, joinset) joinitems(pycompat.maybebytestr(joinset), joiner)

templater: micro-optimize join() with empty separator

templater: factor out generator of join()-ed items Prepares for defining join() behavior per wrapped types and getting rid of the public joinfmt attribute.

templater: pass context to itermaps() for future extension Unlike show() and tovalue(), a base mapping isn't passed to itermaps() since it is the function to generate a partial mapping.

templater: define interface for objects which act as iterator of mappings

stringutil: drop escapedata() in favor of escapestr() They are quite similar. Let's choose one that uses standard Python escape.

peer: make ui an attribute With abc interfaces, instance attributes could not satisfy @abc.abstractproperty requirements because interface conformance was tested at type creation time. When we created the abc peer interfaces, we had to make "ui" a @property to satisfy abc. Now that peer interfaces are using zope.interface and there is no import time validation (but there are tests validating instances conform to the interface), we can go back to using regular object attributes. Differential Revision: https://phab.mercurial-scm.org/D3069

repository: port peer interfaces to zope.interface zope.interface is superior. Let's switch to it. Unlike abc, which defines interfaces through a base class, zope.interface uses different types for interfaces and for implementations. So, we had to invent some new types to hold the interfaces in order to separate the interface from its default implementation. The names here could probably be better. I've been wanting to overhaul the peer interface for a while. And wire protocol version 2 will force that work. So anticipate a refactoring of these interfaces in later commits. With this commit, we no longer test abc interfaces in test-check-interfaces.py, so code for that has been removed. Differential Revision: https://phab.mercurial-scm.org/D3068 # no-check-commit because of stream_out()

wireproto: convert human output frames to CBOR This is easier than rolling our own encoding format. As a bonus, some of our artificial limits around lengths of things went away because we are no longer using fixed length fields to hold sizes. Differential Revision: https://phab.mercurial-scm.org/D3067

py3: use pycompat.bytestr() intsead of str Differential Revision: https://phab.mercurial-scm.org/D3071

py3: use print as a function in tests/test-walk.t Differential Revision: https://phab.mercurial-scm.org/D3070

repo: remove now-unused changectx() method (API) repo.changectx(x) was just a synonym for repo[x], so any extensions that fail due to this commit should switch over to that form. Differential Revision: https://phab.mercurial-scm.org/D3037

localrepo: use revsymbol() in lookup() lookup() seems to be about looking up a revision based on a symbol that may come from the user (via the wire protocol), so revsymbol() is appropriate here. Differential Revision: https://phab.mercurial-scm.org/D3055

histedit: avoid repo.lookup() for converting revnum to nodeid Differential Revision: https://phab.mercurial-scm.org/D3054

outgoing: avoid repo.lookup() for converting revnum to nodeid Differential Revision: https://phab.mercurial-scm.org/D3053

bisect: avoid repo.lookup() for converting revnum to nodeid Differential Revision: https://phab.mercurial-scm.org/D3052

transplant: avoid repo.lookup() for converting revnum to nodeid Differential Revision: https://phab.mercurial-scm.org/D3051

tests: avoid repo.lookup() for converting revnum to nodeid Differential Revision: https://phab.mercurial-scm.org/D3050

bundle: consistently keep a list of stringified revisions in "revs" Before this patch, "revs", in the "not base" branch, would be a list of mixed integral revnums, hex nodeids, and branch names. After this patch, they're all strings. They can still be a mix of hex nodeids and branch names, but the important thing for my future patches is that they're consistently in string form. Differential Revision: https://phab.mercurial-scm.org/D3049

bundle: avoid repo.lookup() for converting revnum to nodeid Differential Revision: https://phab.mercurial-scm.org/D3048

push: avoid using repo.lookup() for converting to nodeid repo.lookup(x) currently simply does repo[x].node(), which supports various types of inputs. As I explained in 0194dac77c93 (scmutil: add method for looking up a context given a revision symbol, 2018-04-02), I'd like to split that up so we use the new scmutil.revsymbol() for string inputs repo[x] for integer revnums and binary nodeids. Since repo.lookup() seems to exist in order to serve peer.lookup(), I think it should be calling revsymbol. However, we have several callers that use repo.lookup() with something that's not a string, so we need to remove those first. This patch starts doing that. Many more will follow. Differential Revision: https://phab.mercurial-scm.org/D3047

addremove: pass command-level similarity value down to scmutil.addremove() Since we've changed to carry a similarity value by opts dict, it makes sense to leave a string '0'-'100' value unmodified.

setup: add overlooked hgext.infinitepush package declaration Will fix infinitepush tests that have been failing when run without --local. Differential Revision: https://phab.mercurial-scm.org/D3038

tests: remove dependence on repo.changectx() This was one of few remaining uses of repo.changectx() in core. Differential Revision: https://phab.mercurial-scm.org/D3036

log: remove dependence on repo.changectx() This was one of few remaining uses of repo.changectx() in core. Differential Revision: https://phab.mercurial-scm.org/D3035

verify: remove dependence on repo.changectx() This was one of few remaining uses of repo.changectx() in core. Differential Revision: https://phab.mercurial-scm.org/D3034

bookmarks: switch from repo.changectx('.') to repo['.'] The two forms are synonymous and the new form is by far the more common form. Differential Revision: https://phab.mercurial-scm.org/D3033

stringutil: add function to pretty print an object This is inspired by the pprint() module/function (which we can't use because the output is different on Python 2 and 3 - namely the use of b'' literals). We hook it up to `hg debugwireproto` for printing the response to a wire protocol command. This foreshadows future peer work, which will support decoding CBOR responses into rich data structures. Differential Revision: https://phab.mercurial-scm.org/D2987

wireproto: add frame flag to denote payloads as CBOR We may eventually want a separate frame type for this. But for now this is the easiest to implement. Differential Revision: https://phab.mercurial-scm.org/D2986

wireproto: implement custom __repr__ for frame This version won't print the full payload (which could be large). It also prints human friendly values for types and flags. Differential Revision: https://phab.mercurial-scm.org/D2985

keepalive: implement readinto() This is part of the standard I/O interface. It is used by the framing protocol. So we need to implement it so frames can be decoded. Differential Revision: https://phab.mercurial-scm.org/D2984

wireproto: port protocol handler to zope.interface zope.interface is superior to the abc module. Let's port to it. As part of this, we add tests for interface conformance for classes implementing the interface. Differential Revision: https://phab.mercurial-scm.org/D2983

wireproto: separate commands tables for version 1 and 2 commands We can't easily reuse existing command handlers for version 2 commands because the response types will be different. e.g. many commands return nodes encoded as hex. Our new wire protocol is binary safe, so we'll wish to encode nodes as binary. We /could/ teach each command handler to look at the protocol handler and change behavior based on the version in use. However, this would make logic a bit unwieldy over time and would make it harder to design a unified protocol handler interface. I think it's better to create a clean break between version 1 and version 2 of commands on the server. What I imagine happening is we will have separate @wireprotocommand functions for each protocol generation. Those functions will parse the request, dispatch to a common function to process it, then generate the response in its own, transport-specific manner. This commit establishes a separate table for tracking version 1 commands from version 2 commands. The HTTP server pieces have been updated to use this new table. Most commands are marked as both version 1 and version 2, so there is little practical impact to this change. A side-effect of this change is we now rely on transport registration in wireprototypes.TRANSPORTS and certain properties of the protocol interface. So a test had to be updated to conform. Differential Revision: https://phab.mercurial-scm.org/D2982

wireproto: mark SSHv2 as a version 1 transport The version component is used for filtering/routing wire protocol commands to their proper handler. The actual version 2 of the wire protocol commands will use a different encoding of responses. We already have tests using the version 2 SSH transport and version 2 of the wire protocol commands won't be implemented atomically. This commit marks the SSHv2 transport as version 1 so it will still invoke the version 1 commands. Once the commands are all implemented in version 2, we can restore its proper behavior. Some tests had to be disabled as a result of this change. Differential Revision: https://phab.mercurial-scm.org/D2981

wireproto: stop aliasing wire protocol types (API) We generally shy away from aliasing module symbols. I think I was keeping this around for API compatibility. We've already made tons of other API breaks in the wire protocol code this release. What's one more? .. api:: ``wireproto`` module no longer re-exports various types used to define responses to wire protocol commands. Access these types from the ``wireprototypes`` module. Differential Revision: https://phab.mercurial-scm.org/D2979

wireproto: use CBOR for command requests Now that we're using CBOR in the new wire protocol, let's convert command requests to it. Before I wrote this patch and was even thinking about CBOR, I was thinking about how commands should be issued and came to the conclusion that we didn't need separate frames to represent the command name from its arguments. I already had a partially completed patch prepared to merge the frames. But with CBOR, it makes the implementation a bit simpler because we don't need to roll our own serialization. The changes here are a bit invasive. I tried to split this into multiple commits to make it easier to review. But it was just too hard. * "command name" and "command argument" frames have been collapsed into a "command request" frame. * The flags for this new frame are totally different. * Frame processing has been overhauled to reflect the new order of things. * Test fallout was significant. A handful of tests were removed. Altogether, I think the new code is simpler. We don't have complicated state around receiving commands. We're either receiving command request frames or command data frames. We /could/ potentially collapse command data frames into command request frames. Although I'd have to think a bit more about this before I do it. Differential Revision: https://phab.mercurial-scm.org/D2951

wireproto: define frame to represent progress updates Today, a long-running operation on a server may run without any sign of progress on the client. This can lead to the conclusion that the server has hung or the connection has dropped. In fact, connections can and do time out due to inactivity. And a long-running server operation can result in the connection dropping prematurely because no data is being sent! While we're inventing the new wire protocol, let's provide a mechanism for communicating progress on potentially expensive server-side events. We introduce a new frame type that conveys "progress" updates. This frame type essentially holds the data required to formulate a ``ui.progress()`` call. We only define the frame right now. Implementing it will be a bit of work since there is no analog to progress frames in the existing wire protocol. We'll need to teach the ui object to write to the wire protocol, etc. The use of a CBOR map may seem wasteful, as this will encode key names in every frame. This *is* wasteful. However, maps are extensible. And the intent is to always use compression via streams. Compression will make the overhead negligible since repeated strings will be mostly eliminated over the wire. Differential Revision: https://phab.mercurial-scm.org/D2902

wireproto: syntax for encoding CBOR into frames We just vendored a library for encoding and decoding the CBOR data format. While the intent of that vendor was to support state files, CBOR is really a nice data format. It is extensible and compact. I've been feeling dirty inventing my own data formats for frame payloads. While custom formats can always beat out a generic format, there is a cost to be paid in terms of implementation, comprehension, etc. CBOR is compact enough that I'm not too worried about efficiency loss. I think the benefits of using a standardized format outweigh rolling our own formats. So I plan to make heavy use of CBOR in the wire protocol going forward. This commit introduces support for encoding CBOR data in frame payloads to our function to make a frame from a human string. We do need to employ some low-level Python code in order to evaluate a string as a Python expression. But other than that, this should hopefully be pretty straightforward. Unit tests for this function have been added. Differential Revision: https://phab.mercurial-scm.org/D2948

wireproto: explicit API to create outgoing streams It is better to create outgoing streams through the reactor so the reactor knows about what streams are active and can track them accordingly. Test output changes slightly because frames from subsequent responses no longer have the "stream begin" stream flag set because the stream is now used across all responses. Differential Revision: https://phab.mercurial-scm.org/D2947

wireproto: add streams to frame-based protocol Previously, the frame-based protocol was just a series of frames, with each frame associated with a request ID. In order to scale the protocol, we'll want to enable the use of compression. While it is possible to enable compression at the socket/pipe level, this has its disadvantages. The big one is it undermines the point of frames being standalone, atomic units that can be read and written: if you add compression above the framing protocol, you are back to having a stream-based protocol as opposed to something frame-based. So in order to preserve frames, compression needs to occur at the frame payload level. Compressing each frame's payload individually will limit compression ratios because the window size of the compressor will be limited by the max frame size, which is 32-64kb as currently defined. It will also add CPU overhead, as it is more efficient for compressors to operate on fewer, larger blocks of data than more, smaller blocks. So compressing each frame independently is out. This means we need to compress each frame's payload as if it is part of a larger stream. The simplest approach is to have 1 stream per connection. This could certainly work. However, it has disadvantages (documented below). We could also have 1 stream per RPC/command invocation. (This is the model HTTP/2 goes with.) This also has disadvantages. The main disadvantage to one global stream is that it has the very real potential to create CPU bottlenecks doing compression. Networks are only getting faster and the performance of single CPU cores has been relatively flat. Newer compression formats like zstandard offer better CPU cycle efficiency than predecessors like zlib. But it still all too common to saturate your CPU with compression overhead long before you saturate the network pipe. The main disadvantage with streams per request is that you can't reap the benefits of the compression context for multiple requests. For example, if you send 1000 RPC requests (or HTTP/2 requests for that matter), the response to each would have its own compression context. The overall size of the raw responses would be larger because compression contexts wouldn't be able to reference data from another request or response. The approach for streams as implemented in this commit is to support N streams per connection and for streams to potentially span requests and responses. As explained by the added internals docs, this facilitates servers and clients delegating independent streams and compression to independent threads / CPU cores. This helps alleviate the CPU bottleneck of compression. This design also allows compression contexts to be reused across requests/responses. This can result in improved compression ratios and less overhead for compressors and decompressors having to build new contexts. Another feature that was defined was the ability for individual frames within a stream to declare whether that individual frame's payload uses the content encoding (read: compression) defined by the stream. The idea here is that some servers may serve data from a combination of caches and dynamic resolution. Data coming from caches may be pre-compressed. We want to facilitate servers being able to essentially stream bytes from caches to the wire with minimal overhead. Being able to mix and match with frames are compressed within a stream enables these types of advanced server functionality. This commit defines the new streams mechanism. Basic code for supporting streams in frames has been added. But that code is seriously lacking and doesn't fully conform to the defined protocol. For example, we don't close any streams. And support for content encoding within streams is not yet implemented. The change was rather invasive and I didn't think it would be reasonable to implement the entire feature in a single commit. For the record, I would have loved to reuse an existing multiplexing protocol to build the new wire protocol on top of. However, I couldn't find a protocol that offers the performance and scaling characteristics that I desired. Namely, it should support multiple compression contexts to facilitate scaling out to multiple CPU cores and compression contexts should be able to live longer than single RPC requests. HTTP/2 *almost* fits the bill. But the semantics of HTTP message exchange state that streams can only live for a single request-response. We /could/ tunnel on top of HTTP/2 streams and frames with HEADER and DATA frames. But there's no guarantee that HTTP/2 libraries and proxies would allow us to use HTTP/2 streams and frames without the HTTP message exchange semantics defined in RFC 7540 Section 8. Other RPC protocols like gRPC tunnel are built on top of HTTP/2 and thus preserve its semantics of stream per RPC invocation. Even QUIC does this. We could attempt to invent a higher-level stream that spans HTTP/2 streams. But this would be violating HTTP/2 because there is no guarantee that HTTP/2 streams are routed to the same server. The best we can do - which is what this protocol does - is shoehorn all request and response data into a single HTTP message and create streams within. At that point, we've defined a Content-Type in HTTP parlance. It just so happens our media type can also work as a standalone, stream-based protocol, without leaning on HTTP or similar protocol. Differential Revision: https://phab.mercurial-scm.org/D2907

wireproto: start to associate frame generation with a stream An upcoming commit will introduce "streams" into the frame-based wire protocol. In preparation for this invasive change, we introduce a basic "stream" class and have all operations that create frames also operate alongside a stream instance. Differential Revision: https://phab.mercurial-scm.org/D2906

tests: fix duplicate and failing test There were two "testconflictingrequestid" methods. Naturally this isn't an error in Python. And by our luck, the test was failing. So we rename the test and fix it to pass. As part of this, _sendsingleframe() now takes a frame, not a string describing the frame. This is better because action at a distance can be confusing. Differential Revision: https://phab.mercurial-scm.org/D2950

debugcommands: drop offset and length from debugindex by default These fields are an implementation detail of revlog storage. As such, they are not part of the generic storage "index" interface and shouldn't be displayed by default. Because we don't have another way to display these fields, we've retained support for printing these fields via --verbose. Yes, I know we should probably be doing all this formatting using modern formatting/templater APIs. I didn't feel like scope bloating this patch. Differential Revision: https://phab.mercurial-scm.org/D3028