unable to clone from servers

[Ross Boylan]

No success cloning from repo.mercurial-scm.org or foss.heptapod.net, despite successful clones from the former a few weeks ago and a successful pull from the former just minutes earlier.

Details

======
``` 

bash-3.2$ date; hg clone https://repo.mercurial-scm.org/evolve evolve2
Fri Sep  4 00:43:59 PDT 2026
requesting all changes
abort: HTTP request error (incomplete response)
(this may be an intermittent network failure; if the error persists, consider contacting the network or server operator)
bash-3.2$ date; hg clone https://foss.heptapod.net/mercurial/evolve evolve2
Fri Sep  4 00:46:49 PDT 2026
abort: error: bad HTTP status line: ''

```

Running on MacOS, M1 chip, using the hg shipped as part of `TortoiseHg.app` v 7.0.1.

I presume it is using the Python shipped with that app, although there are a few others on the computer.

My connection is DSL, and so only moderately fast.  https://www.speedtest.net/result/196337719 shows a network test around the same time; it does seem to have very high upstream latency.

evolve and topic extension enabled with the full path to the extension, i.e., .hgrc includes

```

[extensions]
fsmonitor =
evolve = /Users/ross/src/thg-build/evolve/hgext3rd/evolve
topic = /Users/ross/src/thg-build/evolve/hgext3rd/topic

```

Background

==========

I had previously cloned `evolve` along with most repositories under `thg-build`.  I did not have the `evolve` or `topic` extensions installed at the time.  I have since enabled them, I think (in TortoiseHg settings they show as checked but greyed out and various topic and evolve facilities do appear).  I was able to update the repository, but it did not pick up the topics.  So I thought I'd try a fresh clone.  But cloning from the first repository got an `incomplete response` both times I tried it, and the clone from heptapod doesn't even seem to be using a compatible protocol.

Previously, I noticed that access to the repositories on mercurial-scm.org could be quite erratic and slow, and a recent commit (https://foss.heptapod.net/mercurial/evolve/-/commit/55285f91558b809944c94336154a830a687ac65d) mentions server responsiveness as a problem.  But it hasn't been this bad before, and given the time, 1am US West Coast,  it seems unlikely the servers are heavily loaded.

Thanks.

Ross

[Antonio Muci]

Confirmed.

I am able to pull from an already cloned evolve repo (from the same URL).

$ hg version
Mercurial Distributed SCM (version 7.2.4)

$ hg pull
pulling from https://www.mercurial-scm.org/repo/evolve/
real URL is https://repo.mercurial-scm.org/evolve/
searching for changes
adding changesets
adding manifests
adding file changes
added 20 changesets with 88 changes to 49 files (-1 heads)
new changesets 546566d96604:cf09c46e47b3
(run 'hg update' to get a working copy)

However, if I try to clone the evolve repo from scratch, I have the same "abort: HTTP request error (incomplete response)".

$ hg clone https://repo.mercurial-scm.org/evolve/
destination directory: evolve
requesting all changes
adding changesets
transaction abort!
rollback completed

abort: HTTP request error (incomplete response)
(this may be an intermittent network failure; if the error persists, consider contacting the network or server operator)

It seems a problem on repo.mercurial-scm.org.

--
You received this message because you are subscribed to the Google Groups "TortoiseHg Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to thg-dev+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/thg-dev/923fd670-2280-4dac-9aa7-af4daa257f13n%40googlegroups.com.