hg-stable: Changelog

annotate: do not construct attr.s object per line while computing history Unfortunately, good abstraction has a cost. It's way slower to construct an annotateline() object than creating a plain tuple or a list. This patch changes the internal data structure from row-based to columnar, so the decorate() function can be instant (i.e. no Python in hot loop.) For code readability, the outermost tuple is switched to an attr.s object instead. (original, row-based attr.s) $ hg annot mercurial/commands.py --time > /dev/null time: real 11.470 secs (user 11.400+0.000 sys 0.070+0.000) $ hg annot mercurial/commands.py --time --line-number > /dev/null time: real 39.590 secs (user 39.500+0.000 sys 0.080+0.000) (this patch, columnar) $ hg annot mercurial/commands.py --time > /dev/null time: real 11.780 secs (user 11.710+0.000 sys 0.070+0.000) $ hg annot mercurial/commands.py --time --line-number > /dev/null time: real 12.240 secs (user 12.170+0.000 sys 0.090+0.000) (cf. 4.3.3, row-based tuple) $ hg annot mercurial/commands.py --time --line-number > /dev/null time: real 19.540 secs (user 19.460+0.000 sys 0.080+0.000)

wireproto: explicitly track which requests are active We previously only tracked which requests are receiving. A misbehaving client could accidentally have multiple requests with the same ID in flight. We now explicitly track which request IDs are currently active. We make it illegal to receive a frame associated with a request ID that has already been dispatched. Differential Revision: https://phab.mercurial-scm.org/D2901

wireproto: use named arguments when passing around frame data Named arguments is easier to reason about compared to positional arguments. Especially when you have many positional arguments. Differential Revision: https://phab.mercurial-scm.org/D2900

wireproto: define attr-based classes for representing frames When frames only had 3 attributes, it was reasonable to represent them as a tuple. With them growing more attributes, it will be easier to pass them around as a more formal type. So let's define attr-based classes to represent frame headers and full frames. Differential Revision: https://phab.mercurial-scm.org/D2899

wireproto: define human output side channel frame Currently, the SSH protocol delivers output tailored for people over the stderr file descriptor. The HTTP protocol doesn't have this file descriptor (because it only has an input and output pipe). So it encodes textual output intended for humans within the protocol responses. So response types have a facility for capturing output to be printed to users. Some don't. And sometimes the implementation of how that output is conveyed is super hacky. On top of that, bundle2 has an "output" part that is used to store output that should be printed when this part is encountered. bundle2 also has the concept of "interrupt" chunks, which can be used to signal that the regular bundle2 stream is to be preempted by an out-of-band part that should be processed immediately. This "interrupt" part can be an "output" part and can be used to print data on the receiver. The status quo is inconsistent and insane. We can do better. This commit introduces a dedicated frame type on the frame-based protocol for denoting textual data that should be printed on the receiver. This frame type effectively constitutes a side-channel by which textual data can be printed on the receiver without interfering with other in-progress transmissions, such as the transmission of command responses. But wait - there's more! Previous implementations that transferred textual data basically instructed the client to "print these bytes." This suffered from a few problems. First, the text data that was transmitted and eventually printed originated from a server with a specic i18n configuration. This meant that clients would see text using whatever the i18n settings were on the server. Someone in France could connect to a server in Japan and see unlegible Japanese glyphs - or maybe even mojibake. Second, the normalization of all text data originated on servers resulted in the loss of the ability to apply formatting to that data. Local Mercurial clients can apply specific formatting settings to individual atoms of text. For example, a revision can be colored differently from a commit message. With data over the wire, the potential for this rich formatting was lost. The best you could do (without parsing the text to be printed), was apply a universal label to it and e.g. color it specially. The new mechanism for instructing the peer to print data does not have these limitations. Frames instructing the peer to print text are composed of a formatting string plus arguments. In other words, receivers can plug the formatting string into the i18n database to see if a local translation is available. In addition, each atom being instructed to print has a series of "labels" associated with it. These labels can be mapped to the Mercurial UI's labels so locally configured coloring, styling, etc settings can be applied. What this all means is that textual messages originating on servers can be localized on the client and richly formatted, all while respecting the client's settings. This is slightly more complicated than "print these bytes." But it is vastly more user friendly. FWIW, I'm not aware of other protocols that attempt to encode i18n and textual styling in this manner. You could lobby the claim that this feature is over-engineered. However, if I were to sit in the shoes of a non-English speaker learning how to use version control, I think I would *love* this feature because it would enable me to see richly formatted text in my chosen locale. Anyway, we only implement support for encoding frames of this type and basic tests for that encoding. We'll still need to hook up the server and its ui instance to emit these frames. I recognize this feature may be a bit more controversial than other aspects of the wire protocol because it is a bit "radical." So I'd figured I'd start small to test the waters and see if others feel this feature is worthwhile. Differential Revision: https://phab.mercurial-scm.org/D2872

wireproto: service multiple command requests per HTTP request Now that our new frame-based protocol server can understand how to ingest multiple, possibly interleaved, command requests, let's hook it up to the HTTP server. The code on the HTTP side of things is still a bit hacky. We need a bit of work around error handling, content types, etc. But it's a start. Among the added tests, we demonstrate that a client can send frames for multiple commands iterleaved with each other and that a later issued command can respond before the first one has finished sending. This makes our multi-request model technically superior to the previous "batch" command. Differential Revision: https://phab.mercurial-scm.org/D2871

wireproto: support for receiving multiple requests Now that we have request IDs on each frame and a specification that allows multiple requests to be issued simultaneously, possibly interleaved, let's teach the server to deal with that. Instead of tracking the state for *the* active command request, we instead track the state of each receiving command by its request ID. The multiple states in our state machine for processing each command's state has been collapsed into a single state for "receiving commands." Tests have been added so our branch coverage covers all meaningful branches. However, we did lose some logical coverage. The implementation of this new feature opens up the door to a server having partial command requests when end of input is reached. We will probably want a mechanism to deal with partial requests. For now, I've tracked that as a known issue in the class docstring. I've also noted an abuse vector that becomes a little bit easier to exploit with this feature. Differential Revision: https://phab.mercurial-scm.org/D2870

wireproto: add request IDs to frames One of my primary goals with the new wire protocol is to make operations faster and enable both client and server-side operations to scale to multiple CPU cores. One of the ways we make server interactions faster is by reducing the number of round trips to that server. With the existing wire protocol, the "batch" command facilitates executing multiple commands from a single request payload. The way it works is the requests for multiple commands are serialized. The server executes those commands sequentially then serializes all their results. As an optimization for reducing round trips, this is very effective. The technical implementation, however, is pretty bad and suffers from a number of deficiencies. For example, it creates a new place where authorization to run a command must be checked. (The lack of this checking in older Mercurial releases was CVE-2018-1000132.) The principles behind the "batch" command are sound. However, the execution is not. Therefore, I want to ditch "batch" in the new wire protocol and have protocol level support for issuing multiple requests in a single round trip. This commit introduces support in the frame-based wire protocol to facilitate this. We do this by adding a "request ID" to each frame. If a server sees frames associated with different "request IDs," it handles them as separate requests. All of this happening possibly as part of the same message from client to server (the same request body in the case of HTTP). We /could/ model the exchange the way pipelined HTTP requests do, where the server processes requests in order they are issued and received. But this artifically constrains scalability. A better model is to allow multi-requests to be executed concurrently and for responses to be sent and handled concurrently. So the specification explicitly allows this. There is some work to be done around specifying dependencies between multi-requests. We take the easy road for now and punt on this problem, declaring that if order is important, clients must not issue the request until responses to dependent requests have been received. This commit focuses on the boilerplate of implementing the request ID. The server reactor still can't manage multiple, in-flight request IDs. This will be addressed in a subsequent commit. Because the wire semantics have changed, we bump the version of the media type. Differential Revision: https://phab.mercurial-scm.org/D2869

wireproto: buffer output frames when in half duplex mode Previously, when told that a response was ready, the server reactor would instruct the caller to send frames immediately. This was OK as an initial implementation. But it would not work for half-duplex connections where the sender can't receive until all data has been transmitted - such as httplib based clients. In this commit, we teach the reactor that output frames should be buffered until end of input is seen. This required a new event to inform the reactor of end of input. The result from that event will instruct the consumer to send all buffered frames. The HTTP server is buffered by default. This change effectively hides the complexity of buffering within the reactor so that transports need not be concerned about it. This helps keep the transports "dumb" and will make implementing multiple requests-responses per atomic exchange (like an HTTP request) much simpler. Differential Revision: https://phab.mercurial-scm.org/D2860

wireproto: define and implement responses in framing protocol Previously, we only had client-side frame types defined. This commit defines and implements basic support for server-side frame types. We introduce two frame types - one for representing the raw bytes result of a command and another for representing error results. The types are quite primitive and behavior will expand over time. But you have to start somewhere. Our server reactor gains methods to react to an intent to send a response. Again, following the "sans I/O" pattern, the reactor doesn't actually send the data. Instead, it gives the caller a generator to frames that it can send out over the wire. Differential Revision: https://phab.mercurial-scm.org/D2858

wireproto: implement basic command dispatching for HTTPv2 Now that we can ingest frames and decode them to requests to run commands, we are able to actually run those commands. So this commit starts to implement that. There are numerous shortcomings. We can't operate on commands with "*" arguments. We can only emit bytesresponse results. We don't yet issue a response in the unified framing protocol. But it's a start. Differential Revision: https://phab.mercurial-scm.org/D2857

wireproto: nominally don't expose "batch" to version 2 wire transports The unified frame-based protocol will (eventually) support multiple requests per client transmission. This means that the [very hacky] "batch" command has no purpose existing in this protocol. This commit marks the command as applying to v1 transports only. But because SSHv2 == SSHv1 currently, we had to hack it back in for the SSHv2 transport. Bleh. Tests changed because the capabilities string changed. The order of tokens in the string is not important. Differential Revision: https://phab.mercurial-scm.org/D2856

wireproto: implement basic frame reading and processing We just implemented support for writing frames. Now let's implement support for reading them. The bulk of the new code is for a class that maintains the state of a server. Essentially, you construct an instance, feed frames to it, and it tells you what you should do next. The design is inspired by the "sans I/O" movement and the reactor pattern. We don't want to perform I/O or any major blocking event during frame ingestion because this arbitrarily limits ways that server pieces can be implemented. For example, it makes it much harder to swap in an alternate implementation based on asyncio or do crazy things like have requests dispatch to other processes. We do still implement readframe() which does I/O. But it is decoupled from the server reactor. And important parsing of frame headers is a standalone function. So I/O is only needed to obtain frame data. Because testing server-side ingest is useful and difficult on running servers, we create a new "debugreflect" endpoint that will echo back to the client what was received and how it was interpreted. This could be useful for a server admin, someone implementing a client. But immediately, it is useful for testing: we're able to demonstrate that frames are parsed correctly and turned into requests to run commands without having to implement command dispatch on the server! In addition, we implement Python level unit tests for the reactor. This is vastly more efficient than sending requests to the "debugreflect" endpoint and vastly more powerful for advanced testing. Differential Revision: https://phab.mercurial-scm.org/D2852