Re: No error checking when reading from file using zstd in pg_dump

From: Daniel Gustafsson <daniel(at)yesql(dot)se>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: Evgeniy Gorbanev <gorbanyoves(at)basealt(dot)ru>, pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: Re: No error checking when reading from file using zstd in pg_dump
Date: 2025-06-16 14:11:58
Message-ID: 3815B6CD-924B-40FA-823A-55E2F8F24602@yesql.se
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

> On 16 Jun 2025, at 15:56, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> I've not checked to see what the other users of this API do, but
> if they're all like this then we need to fix that comment.

AFAICT all other callers of this API are throwing an error with pg_fatal, and
so does the function in question for ZStd decompression errors. If we handle
the case of fread() returning 0 to indicate an error like the below *untested
sketch* (with a better error message) this function is fully API compliant as
well.

/* If we have no more input to consume, we're done */
if (cnt == 0)
+ {
+ if (ferror(unconstify(void *, input->src)))
+ pg_fatal("could not read data to decompress: %m");
+
break;
+ }

If this seems like a good approach then Zstd_getc can be simplified as well as
it no longer needs to call ferror, it still needs to check feof though.

--
Daniel Gustafsson

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2025-06-16 14:20:07 Re: No error checking when reading from file using zstd in pg_dump
Previous Message Tom Lane 2025-06-16 14:09:28 Re: [PATCH] Remove unused #include's in src/backend/utils/adt/*