Re: Avoid erroring out when unable to remove or parse logical rewrite files to save checkpoint work

Hi,

On 2022-08-27 14:06:32 +0530, Bharath Rupireddy wrote:
> @@ -50,7 +51,7 @@ copydir(char *fromdir, char *todir, bool recurse)
>  
>  	while ((xlde = ReadDir(xldir, fromdir)) != NULL)
>  	{
> -		struct stat fst;
> +		PGFileType	xlde_type;
>  
>  		/* If we got a cancel signal during the copy of the directory, quit */
>  		CHECK_FOR_INTERRUPTS();
> @@ -62,18 +63,15 @@ copydir(char *fromdir, char *todir, bool recurse)
>  		snprintf(fromfile, sizeof(fromfile), "%s/%s", fromdir, xlde->d_name);
>  		snprintf(tofile, sizeof(tofile), "%s/%s", todir, xlde->d_name);
>  
> -		if (lstat(fromfile, &fst) < 0)
> -			ereport(ERROR,
> -					(errcode_for_file_access(),
> -					 errmsg("could not stat file \"%s\": %m", fromfile)));
> +		xlde_type = get_dirent_type(fromfile, xlde, false, ERROR);
>  
> -		if (S_ISDIR(fst.st_mode))
> +		if (xlde_type == PGFILETYPE_DIR)
>  		{
>  			/* recurse to handle subdirectories */
>  			if (recurse)
>  				copydir(fromfile, tofile, true);
>  		}
> -		else if (S_ISREG(fst.st_mode))
> +		else if (xlde_type == PGFILETYPE_REG)
>  			copy_file(fromfile, tofile);
>  	}
>  	FreeDir(xldir);

It continues to make no sense to me to add behaviour changes around
error-handling as part of a conversion to get_dirent_type(). I don't at all
understand why e.g. the above change to make copydir() silently skip over
files it can't stat is ok?

Greetings,

Andres Freund