[
https://issues.apache.org/jira/browse/THRIFT-4244?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18114209#comment-18114209
]
Jens Geyer commented on THRIFT-4244:
------------------------------------
Where the string literal syntax is documented:
* {{doc/specs/idl.md}}, which is also what [https://thrift.apache.org/docs/idl]
shows, only has rule 36. It does not mention escape sequences. Read literally,
a backslash is an ordinary character, a double quote cannot appear in a
double-quoted literal and line breaks are allowed. That is the reading in the
description above, but it is not what the compiler does:
{noformat}
[36] Literal ::= ('"' [^"]* '"') | ("'" [^']* "'")
{noformat}
* Nothing else covers it. [https://thrift.apache.org/docs/types] only says that
strings are UTF-8, and {{tutorial/tutorial.thrift}} had no example either.
The actual syntax, checked against current master:
{noformat}
- Enclosed in double or single quotes. The other quote character can be used
as is: "don't", 'say "hi"'
- Escape sequences, in both quote styles: \" \' \\ \n \r \t
- Any other character after a backslash is an error, so a literal backslash
has to be doubled: "Y-m-d\\TH:i:s.uP"
- No numeric escapes such as \x41 or \u00e4; non-ASCII characters are
written directly
- A literal cannot span lines
- The same rules apply to include paths and annotation values
{noformat}
[PR #3842|https://github.com/apache/thrift/pull/3842] documents this in
{{tutorial/tutorial.thrift}} (a plain comment, not a doc comment, so generated
code does not change) and corrects rule 36 in {{doc/specs/idl.md}}, which the
website picks up on its next build.
Related: several generators (c_glib, cl, dart, delphi, kotlin, lua, rb, rs) do
not escape string constant values correctly for their language, so a constant
containing a quote, a backslash or a line break can give invalid code or a
different value, e.g. for Rust:
{noformat}
pub const X: &str = "say "hi"";
{noformat}
Details in THRIFT-6236.
_Drafted with AI assistance (Claude Opus 5); reviewed and posted by Jens Geyer._
> PHP compiler errors out if escape character is part in string constant
> ----------------------------------------------------------------------
>
> Key: THRIFT-4244
> URL: https://issues.apache.org/jira/browse/THRIFT-4244
> Project: Thrift
> Issue Type: Bug
> Components: PHP - Compiler
> Affects Versions: 0.10.0
> Environment: official thrift docker image
> https://hub.docker.com/_/thrift/
> Reporter: Pascal von Rickenbach
> Assignee: Volodymyr Panivko
> Priority: Minor
> Fix For: 0.25.0
>
> Time Spent: 2h 10m
> Remaining Estimate: 0h
>
> The compiler throws an error when defining a string constant that contains an
> escape character. E.g a backslash like in this definition:
> const string ISO8601U = "Y-m-d\TH:i:s.uP";
> According to the Thrift IDL only double quotes are forbidden in the string.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
