##// END OF EJS Templates
tagcache: distinguish between invalid and missing entries...
tagcache: distinguish between invalid and missing entries The TortoiseHg repo has typically not had a newly applied tag accessible by name for recent releases, for unknown reasons. Deleting and rebuilding the tag cache doesn't fix it, though deleting the cache and running `hg log -r $new_tag` does. Eventually the situation does sort itself out for new clones from the server. In an effort to figure out what the issue is, Pierre-Yves David suggested listing these entries in the debug output more specifically. This isn't complete yet- the second test change that says "missing" is more like "invalid", since it was truncated. The problem there is the code that reads the raw array truncates any partial records and then fills it with 0xFF, which signifies that it is missing. As a side note, that means the check for the length when validating an existing entry never fails. Differential Revision: https://phab.mercurial-scm.org/D9811

File last commit:

r46195:426294d0 default
r47245:5aac1a1a default
Show More
runcommand.rs
66 lines | 2.4 KiB | application/rls-services+xml | RustLexer
// Copyright 2018 Yuya Nishihara <yuya@tcha.org>
//
// This software may be used and distributed according to the terms of the
// GNU General Public License version 2 or any later version.
//! Functions to run Mercurial command in cHg-aware command server.
use bytes::Bytes;
use std::io;
use std::os::unix::io::AsRawFd;
use tokio_hglib::codec::ChannelMessage;
use tokio_hglib::{Connection, Protocol};
use crate::attachio;
use crate::message::{self, CommandType};
use crate::uihandler::SystemHandler;
/// Runs the given Mercurial command in cHg-aware command server, and
/// fetches the result code.
///
/// This is a subset of tokio-hglib's `run_command()` with the additional
/// SystemRequest support.
pub async fn run_command(
proto: &mut Protocol<impl Connection + AsRawFd>,
handler: &mut impl SystemHandler,
packed_args: impl Into<Bytes>,
) -> io::Result<i32> {
proto
.send_command_with_args("runcommand", packed_args)
.await?;
loop {
match proto.fetch_response().await? {
ChannelMessage::Data(b'r', data) => {
return message::parse_result_code(data);
}
ChannelMessage::Data(..) => {
// just ignores data sent to optional channel
}
ChannelMessage::InputRequest(..)
| ChannelMessage::LineRequest(..) => {
return Err(io::Error::new(
io::ErrorKind::InvalidData,
"unsupported request",
));
}
ChannelMessage::SystemRequest(data) => {
let (cmd_type, cmd_spec) = message::parse_command_spec(data)?;
match cmd_type {
CommandType::Pager => {
// server spins new command loop while pager request is
// in progress, which can be terminated by "" command.
let pin = handler.spawn_pager(&cmd_spec).await?;
attachio::attach_io(proto, &io::stdin(), &pin, &pin)
.await?;
proto.send_command("").await?; // terminator
}
CommandType::System => {
let code = handler.run_system(&cmd_spec).await?;
let data = message::pack_result_code(code);
proto.send_data(data).await?;
}
}
}
}
}
}