Paul Burba wrote:

> On Thu, Sep 6, 2012 at 7:32 PM,  <[email protected]> wrote:
>>  URL: http://svn.apache.org/viewvc?rev=1381800&view=rev
>>  Log:
>>  Re-implement svn_io_read_length_line as this is one of the
>>  most-called functions in SVN. Instead of reading data a byte
>>  at a time, read 128 byte chunks and scan those.
[...]
> ==============================================================================
>>  --- subversion/trunk/subversion/libsvn_subr/io.c (original)
>>  +++ subversion/trunk/subversion/libsvn_subr/io.c Thu Sep  6 23:32:11 2012
>>  @@ -3427,30 +3427,60 @@ svn_error_t *
>>   svn_io_read_length_line(apr_file_t *file, char *buf, apr_size_t *limit,
>>                           apr_pool_t *pool)
>>   {
>>  +  /* variables */
>>  +  apr_size_t total_read = 0;
>>  +  svn_boolean_t eof = FALSE;
>>     const char *name;
>>     svn_error_t *err;
>>  -  apr_size_t i;
>>  -  char c;
>>  +  apr_size_t buf_size = *limit;
>> 
>>  -  for (i = 0; i < *limit; i++)
>>  +  while (buf_size > 0)
>>       {
>>  -      SVN_ERR(svn_io_file_getc(&c, file, pool));
>>  -      /* Note: this error could be APR_EOF, which
>>  -         is totally fine.  The caller should be aware of
>>  -         this. */
>>  -
>>  -      if (c == '\n')
>>  +      /* read a fair chunk of data at once. But don't get too ambitious
>>  +       * as that would result in too much waste. Also make sure we can
>>  +       * put a NUL after the last byte read.
>>  +       */
>>  +      apr_size_t to_read = buf_size < 129 ? buf_size - 1 : 128;
>>  +      apr_size_t bytes_read = 0;
>>  +      char *eol;
>>  +
>>  +      /* read data block (or just a part of it) */
>>  +      SVN_ERR(svn_io_file_read_full2(file, buf, to_read,
>>  +                                     &bytes_read, &eof, pool));
>>  +
>>  +      /* look or a newline char */
>>  +      buf[bytes_read] = 0;
>>  +      eol = strchr(buf, '\n');
>>  +      if (eol)
>>           {
>>  -          buf[i] = '\0';
>>  -          *limit = i;
>>  +          apr_off_t offset = (eol + 1 - buf) - bytes_read;
>>  +
>>  +          *eol = 0;
>>  +          *limit = total_read + (eol - buf);
>>  +
>>  +          /* correct the file pointer:
>>  +           * appear as though we just had read the newline char
>>  +           */
>>  +          SVN_ERR(svn_io_file_seek(file, APR_CUR, &offset, pool));
>>  +
>>             return SVN_NO_ERROR;
>>           }
>>  -      else
>>  +      else if (eof)
>>           {
>>  -          buf[i] = c;
>>  +          /* no EOL found but we hit the end of the file.
>>  +           * Generate a nice EOF error object and return it.
>>  +           */
>>  +          char dummy;
>>  +          SVN_ERR(svn_io_file_getc(&dummy, file, pool));
>>           }
>>  +
>>  +      /* next data chunk */
>>  +      buf_size -= bytes_read;
>>  +      buf += bytes_read;
>>  +      total_read += bytes_read;
>>       }
>> 
>>  +  /* buffer limit has been exceeded without finding the EOL */
>>     err = svn_io_file_name_get(&name, file, pool);
>>     if (err)
>>       name = NULL;
>> 
>> 
> 
> Anybody else seeing this?

Yes, I'm seeing a failure with the same error message, on my Ubuntu Linux 
system.

- Julian


> I haven't figured out why yet, but r1381800 is causing failures on my
> Windows box, all similar to this:
[...]
> I: svn: E070014: Can't read file
> 'C:\SVN\src-trunk\Debug\subversion\tests\cmdline\svn-test-work\local_tmp\repos\db\revs\0\0':
> End of file found
[...]

Reply via email to