changegroup: remove reordering control (BC) This logic - including the experimental bundle.reorder option - was originally added in a8e3931e3fb5 in 2011 and then later ported to changegroup.py. The intent of this option and associated logic is to control the ordering of revisions in deltagroups in changegroups. At the time it was implemented, only changegroup version 1 existed and generaldelta revlogs were just coming into the world. Changegroup version 1 requires that deltas be made against the last revision sent over the wire. Used with generaldelta, this created an impedance mismatch of sorts and resulted in changegroup producers spending a lot of time recomputing deltas. Revision reordering was introduced so outgoing revisions would be sent in "generaldelta order" and producers would be able to reuse internal deltas from storage. Later on, we introduced changegroup version 2. It supported denoting which revision a delta was against. So we no longer needed to sort outgoing revisions to ensure optimal delta generation from the producer. So, subsequent changegroup versions disabled reordering. We also later made the changelog not store deltas by default. And we also made the changelog send out deltas in storage order. Why we do this for changelog, I'm not sure. Maybe we want to preserve revision order across clones? It doesn't really matter for this commit. Fast forward to 2018. We want to abstract storage backends. And having changegroup code require knowledge about how deltas are stored internally interferes with that goal. This commit removes reordering control from changegroup generation. After this commit, the reordering behavior is: * The changelog is always sent out in storage order (no behavior change). * Non-changelog generaldelta revlogs are reordered to always be in DAG topological order (previously, generaldelta revlogs would be emitted in storage order for version 2 and 3 changegroups). * Non-changelog non-generaldelta revlogs are sent in storage order (no behavior change). * There exists no config option to override behavior. The big difference here is that generaldelta revlogs now *always* have their revisions sorted in DAG order before going out over the wire. This behavior was previously only done for changegroup version 1. Version 2 and version 3 changegroups disabled reordering because the interchange format supported encoding arbitrary delta parents, so reordering wasn't strictly necessary. I can think of a few significant implications for this change. Because changegroup receivers will now see non-changelog revisions in DAG order instead of storage order, the internal storage order of manifests and files may differ substantially between producer and consumer. I don't think this matters that much, since the storage order of manifests and files is largely hidden from users. Only the storage order of changelog matters (because `hg log` shows the changelog in storage order). I don't think there should be any controversy here. The reordering of revisions has implications for changegroup producers. Previously, generaldelta revlogs would be emitted in storage order. And in the common case, the internally-stored delta could effectively be copied from disk into the deltagroup delta. This meant that emitting delta groups for generaldelta revlogs would be mostly linear read I/O. This is desirable for performance. With us now reordering generaldelta revlog revisions in DAG order, the read operations may use more random I/O instead of sequential I/O. This could result in performance loss. But with the prevalence of SSDs and fast random I/O, I'm not too worried. (Note: the optimal emission order for revlogs is actually delta encoding order. But the changegroup code wasn't doing that before or after this change. We could potentially implement that in a later commit.) Changegroups in DAG order will have implications for receivers. Previously, receiving storage order might mean seeing a number of interleaved branches. This would mean long delta chains, sparse I/O, and possibly more fulltext revisions instead of deltas, blowing up storage storage. (This is the same set of problems that sparse revlogs aims to address.) With the producer now sending revisions in DAG order, the receiver also stores revisions in DAG order. That means revisions for the same DAG branch are all grouped together. And this should yield better storage outcomes. In other words, sending the reordered changegroup allows the receiver to have better storage order and for the producer to not propagate its (possibly sub-optimal) internal storage order. On the mozilla-unified repository, this change influences bundle generation: $ hg bundle -t none-v2 -a before: time: real 355.680 secs (user 256.790+0.000 sys 16.820+0.000) after: time: real 382.950 secs (user 281.700+0.000 sys 17.690+0.000) before: 7,150,228,967 bytes (uncompressed) after: 7,041,556,273 bytes (uncompressed) before: 1,669,063,234 bytes (zstd l=3) after: 1,628,598,830 bytes (zstd l=3) $ hg unbundle before: time: real 511.910 secs (user 466.750+0.000 sys 32.680+0.000) after: time: real 487.790 secs (user 443.940+0.000 sys 30.840+0.000) 00manifest.d size: source: 274,924,292 bytes before: 304,741,626 bytes after: 245,252,087 bytes .hg/store total file size: source: 2,649,133,490 before: 2,680,888,130 after: 2,627,875,673 We see the bundle size drop. That's probably because if a revlog internally isn't storing a delta, it will choose to delta against the last emitted revision. And on repos with interleaved branches (like mozilla-unified), the previous revision could be an unrelated branch and therefore be a large delta. But with this patch, the previous revision is likely p1 or p2 and a delta should be small. We also see the manifest size drop by ~50 MB. It's worth noting that the manifest actually *increased* in size by ~25 MB in the old strategy and decreased ~25 MB from its source in the new strategy. Again, my explanation for this is that the DAG ordering in the changegroup is resulting in better grouping of revisions in the receiver, which results in more compact delta chains and higher storage efficiency. Unbundle time also dropped. I suspect this is due to the revlog having to work less to compute deltas since the incoming deltas are more optimal. i.e. the receiver spends less time resolving fulltext revisions as incoming deltas bounce around between DAG branches and delta chains. We also see bundle generation time increase. This is not desirable. However, the regression is only significant on the original repository: if we generate a bundle from the repository created from the new, always reordered bundles, we're close to baseline (if not at it with expected noise): $ hg bundle -t none-v2 -a before (original): time: real 355.680 secs (user 256.790+0.000 sys 16.820+0.000) after (original): time: real 382.950 secs (user 281.700+0.000 sys 17.690+0.000) after (new repo): time: real 362.280 secs (user 260.300+0.000 sys 17.700+0.000) This regression is a bit worrying because it will impact serving canonical repositories (that don't have optimal internal storage unless they are reordered - possibly as part of running `hg debugupgraderepo`). However, this regression will only be noticed by very large changegroups. And I'm guessing/hoping that any repository that large is using clonebundles to mitigate server load. Again, sending DAG order isn't the optimal send order for servers: sending in storage-delta order is. But in order to enable storage-optimal send order, we'll need a storage API that handles sorting. Future commits will introduce such an API. Differential Revision: https://phab.mercurial-scm.org/D4721

filelog: drop index attribute (API) The previous commit removed the last consumer of the "index" attribute on the file storage interface. The index is an extremely low-level data structure that is revlog specific and isn't appropriate to expose as part of a generic storage API. There may be a market for an efficient data structure to obtain metadata on every revision for a file. But if there is, it should be designed using e.g. named attributes for lookup instead of a list-like of 8-tuples. Let's drop the attribute from filelog and remove the attribute from the file storage interface. Differential Revision: https://phab.mercurial-scm.org/D4720

upgrade: use rawsize() instead of revlog index The revlog index is a very low-level data structure and it shouldn't be exposed to the storage interface - at least not in its current form. upgrade.py is the only consumer of the index attribute on file storage in the repository. This commit rewrites that final consumer to use rawsize() instead of going through the index. This is actually the more proper API to use, as rawsize() will accurately report the size of revisions which have a negative size in the index. Differential Revision: https://phab.mercurial-scm.org/D4719

manifest: add rawsize() proxy (API) I'm not keen about doing this. But it unblocks efforts to remove "index" from the file storage interface. We will probably remove this once we have a better upgrade API in place. Differential Revision: https://phab.mercurial-scm.org/D4718

upgrade: report size of backing files, not internal storage size upgrade.py is the only consumer of filelog.index, which I'd like to eliminate from the file storage interface. This commit changes the upgrade code to report the storage size of files by looking at the size of the files backing its storage instead of looking at the index. I'm not convinced the approach in this patch will live very long because it is relying on low-level attributes like "opener" and "files," which may behave very differently on non-revlog storage. But the data is only used for reporting purposes and it does get us one step closer to eliminating "index." A side-effect of this change is we now report the size of the revlog index data - not just the revision data. I think this is more accurate. Differential Revision: https://phab.mercurial-scm.org/D4717

filelog: store filename directly on revlog instance This attribute is only used by LFS. It is used by one of the revlog flag processor functions, which gets an instance of the revlog - not the file storage type. So, it makes sense to store this attribute on the revlog instead of the filelog. With this change, I'm pretty confident that LFS is no longer directly accessing file storage interface members that are revlog centric. i.e. it gets us one step closer to eliminating revlog-centric APIs from the file storage interface! Differential Revision: https://phab.mercurial-scm.org/D4715

lfs: access revlog directly LFS is monkeypatching filelog.filelog and is then accessing various filelog attributes in the monkeypatched function. This is all fine. But some of the attributes being accessed by LFS are revlog centric and shouldn't be exposed on the file storage interface. This commit changes the monkeypatched functions to access proxied attributes on self._revlog instead of self. This should be safe to do because non-revlog repositories should not be using filelog instances: instead they should have a separate class to represent file storage. So it is reasonable for LFS to assume the _revlog attribute exists and points to a revlog. Differential Revision: https://phab.mercurial-scm.org/D4714

largefiles: automatically load largefiles extension when required (BC) This is very similar to what we just did for LFS but for largefiles. See recent commit messages for the rationale here. Differential Revision: https://phab.mercurial-scm.org/D4713

lfs: don't add extension to hgrc after clone or share (BC) Now that repository loading in core supports automatically loading the lfs extension when the "lfs" requirement is present, we no longer need to update the .hg/hgrc of newly-created repos to load the lfs extension! I'm marking this as BC because it is a change in behavior. But users should not notice unless they create an LFS repo with new Mercurial and then attempt to use it with an old versions that doesn't support automatic extension loading. Differential Revision: https://phab.mercurial-scm.org/D4712

localrepo: automatically load lfs extension when required (BC) If an unrecognized requirement is present (possibly due to an unloaded extension), the user will get an error message telling them to go to https://mercurial-scm.org/wiki/MissingRequirement for more info. And some requirements clearly map to known extensions shipped by Mercurial. This commit teaches repository loading to automatically map requirements to extensions. We implement support for loading the lfs extension when the "lfs" requirement is present. This behavior feels more user-friendly to me and I'm having trouble coming up with a compelling reason to not do it. The strongest argument I have against is that - strictly speaking - requirements are general repository features and there could be N providers of that feature. e.g. in the case of LFS, there could be another extension implementing LFS support. And the user would want to use this non-official extension rather than the built-in one. The way this patch implements things, the non-official extension could be missing and Mercurial would load the official lfs extension, leading to unexpected behavior. But this feels like a highly marginal use case to me and doesn't outweigh the user benefit of "it just works." If someone really wanted to e.g. use a custom LFS extension, they could prevent the built-in one from being loaded by either defining "extensions.lfs=/path/to/custom/extension" or "extensions.lfs=!", as the automatic extension loading only occurs if there is no config entry for that extension. Differential Revision: https://phab.mercurial-scm.org/D4711