##// END OF EJS Templates
hgweb: encode WSGI environment using the ISO-8859-1 codec...
hgweb: encode WSGI environment using the ISO-8859-1 codec The WSGI specification (PEP 3333) specifies that on Python 3 all strings passed by the server must be of type str with code points encodable using the ISO 8859-1 codec. For some reason, I introduced a bug in 2632c1ed8f34 by applying the reverse change. Maybe I got confused because PEP 3333 says that arbitrary operating system environment variables may be contained in the WSGI environment and therefore we need to handle the WSGI environment variables like we would handle operating system environment variables. The bug mentioned in the previous paragraph and fixed by this changeset manifested e.g. in the path of the URL being encoded in the wrong way. Browsers encode non-ASCII bytes with the percent-encoding. WSGI servers will decode the percent-encoded bytes and pass them to the application as strings where each byte is mapped to the corresponding code point with the same ordinal (i.e. it is decoded using the ISO-8859-1 codec). Mercurial uses the bytes type for these strings (which makes much more sense), so we need to encode it again using the ISO-8859-1 codec. If we use another codec, it can result in nonsense.

File last commit:

r50527:048f829a default
r51713:9ed281bb stable
Show More
nodemap_docket.rs
108 lines | 3.6 KiB | application/rls-services+xml | RustLexer
use crate::errors::{HgError, HgResultExt};
use bytes_cast::{unaligned, BytesCast};
use memmap2::Mmap;
use std::path::{Path, PathBuf};
use crate::vfs::Vfs;
const ONDISK_VERSION: u8 = 1;
pub(super) struct NodeMapDocket {
pub data_length: usize,
// TODO: keep here more of the data from `parse()` when we need it
}
#[derive(BytesCast)]
#[repr(C)]
struct DocketHeader {
uid_size: u8,
_tip_rev: unaligned::U64Be,
data_length: unaligned::U64Be,
_data_unused: unaligned::U64Be,
tip_node_size: unaligned::U64Be,
}
impl NodeMapDocket {
/// Return `Ok(None)` when the caller should proceed without a persistent
/// nodemap:
///
/// * This revlog does not have a `.n` docket file (it is not generated for
/// small revlogs), or
/// * The docket has an unsupported version number (repositories created by
/// later hg, maybe that should be a requirement instead?), or
/// * The docket file points to a missing (likely deleted) data file (this
/// can happen in a rare race condition).
pub fn read_from_file(
store_vfs: &Vfs,
index_path: &Path,
) -> Result<Option<(Self, Mmap)>, HgError> {
let docket_path = index_path.with_extension("n");
let docket_bytes = if let Some(bytes) =
store_vfs.read(&docket_path).io_not_found_as_none()?
{
bytes
} else {
return Ok(None);
};
let input = if let Some((&ONDISK_VERSION, rest)) =
docket_bytes.split_first()
{
rest
} else {
return Ok(None);
};
/// Treat any error as a parse error
fn parse<T, E>(result: Result<T, E>) -> Result<T, HgError> {
result
.map_err(|_| HgError::corrupted("nodemap docket parse error"))
}
let (header, rest) = parse(DocketHeader::from_bytes(input))?;
let uid_size = header.uid_size as usize;
// TODO: do we care about overflow for 4 GB+ nodemap files on 32-bit
// systems?
let tip_node_size = header.tip_node_size.get() as usize;
let data_length = header.data_length.get() as usize;
let (uid, rest) = parse(u8::slice_from_bytes(rest, uid_size))?;
let (_tip_node, _rest) =
parse(u8::slice_from_bytes(rest, tip_node_size))?;
let uid = parse(std::str::from_utf8(uid))?;
let docket = NodeMapDocket { data_length };
let data_path = rawdata_path(&docket_path, uid);
// TODO: use `vfs.read()` here when the `persistent-nodemap.mmap`
// config is false?
if let Some(mmap) =
store_vfs.mmap_open(&data_path).io_not_found_as_none()?
{
if mmap.len() >= data_length {
Ok(Some((docket, mmap)))
} else {
Err(HgError::corrupted("persistent nodemap too short"))
}
} else {
// Even if .hg/requires opted in, some revlogs are deemed small
// enough to not need a persistent nodemap.
Ok(None)
}
}
}
fn rawdata_path(docket_path: &Path, uid: &str) -> PathBuf {
let docket_name = docket_path
.file_name()
.expect("expected a base name")
.to_str()
.expect("expected an ASCII file name in the store");
let prefix = docket_name
.strip_suffix(".n.a")
.or_else(|| docket_name.strip_suffix(".n"))
.expect("expected docket path in .n or .n.a");
let name = format!("{}-{}.nd", prefix, uid);
docket_path
.parent()
.expect("expected a non-root path")
.join(name)
}