Thank you for your reply. Unfortunately, my module is written in C, so
I suspect the modperl list might not be as helpful.
Here is the skeleton of my upload code. The function given is the
content handler for this particular request. There are no significant
prior input filters, and no significant posterior output filters. When
I POST an upload, this content handler saves the files in the multipart
POST, and returns a trivial response (not fully shown). In my testing,
it tends to leak memory proportional (but not equal) to the size of the
total upload.
I initialize mod_apreq in the way offered in the documentation: by
LoadModule'ing it, and expecting it to insert itself into the latest
possible input filter position (which should not be difficult, since I
have no other custom input filters).
I apologize for the length of this email. Thank you for helping out!
Eddie
__snip__
// a content handler
static int do_post(request_rec *r)
{
apreq_handle_t *handle;
const apreq_module_t *module;
apreq_param_t *filebody, *testfilebody;
apr_bucket_brigade *bb;
apr_bucket *b;
char *loopfilename, *myip, *disk, *tmpfile, *newfilename,
*extensionfilename, *lastdot;
char *localorigfilename_raw, *localorigfilename, *nameindex = NULL;
const char *buf = NULL;
apr_file_t *f = NULL;
unsigned int md_len, copies, filesize = 0, any_processed = 0;
apr_size_t bytes;
int i, lresult = 0;
handle = (apreq_handle_t *)apreq_handle_apache2(r);
apreq_read_limit_set(handle, READ_LIMIT);
module = handle->module;
localorigfilename_raw = (char *)apr_pcalloc(r->pool, 512);
loopfilename = (char *)apr_pcalloc(r->pool, 16);
for (i = 0; i < MAX_FILES_PROCESSED; i++)
{
filesize = 0;
sprintf(loopfilename, "%s%d", DEFAULT_FILE_PARAM_NAME, i);
// i do typically test whether this mulitpart part is a file -- for
the sake of this skeleton, that code is omitted
filebody = module->body_get(handle, loopfilename);
memcpy(localorigfilename_raw, (filebody->v).data,
strlen((filebody->v).data));
bb = filebody->upload;
tmpfile = create_temp_path(disk, (char *)localorigfilename);
if (apr_file_open(&f,
tmpfile,APR_WRITE|APR_CREATE|APR_BINARY|APR_TRUNCATE,APR_UREAD|APR_UWRITE|APR_UEXECUTE,
r->pool) != APR_SUCCESS)
{
return HTTP_FORBIDDEN;
}
// dump the values in this brigade:
for (b = APR_BRIGADE_FIRST(bb);
b != APR_BRIGADE_SENTINEL(bb);
b = APR_BUCKET_NEXT(b))
{ if (APR_BUCKET_IS_EOS(b))
{
break;
}
else if (apr_bucket_read(b, &buf, &bytes, APR_BLOCK_READ) ==
APR_SUCCESS)
{
// We have a bucket full of raw data
apr_file_write(f, buf, &bytes);
filesize += bytes;
}
}
apr_file_close(f);
if (bb != NULL)
{
apr_brigade_destroy(bb);
// apr_brigade_cleanup(bb);
}
} // for i in 0 to MAX_FILES_PROCESSED
if (any_processed > 0)
{
return OK;
}
else
{
return HTTP_FORBIDDEN;
}
}
__snip__
Philip M. Gollucci wrote:
Edward L. Abrams wrote:
Hello,
I apologize for my brief intrusion on this list for a question that
may or may not be within the boundaries of the list's purpose. I
have been using libapreq to do file uploads for some time. Lately,
I've noticed that I have memory leaks. I don't know whether it's my
newbie use of mostly-undocumented code, or the code itself (I favor
the former!).
This is generally for developement of libapreq itself, but sometimes
user questions are on topic.
You might try modperl (at) perl (dot) apache (dot) org
If this is the wrong place to ask what I am doing wrong in my code,
where is the right place to ask? The modules@ list was unresponsive,
and I could find no other close candidate. If this is not the wrong
place to ask, then let me know, and I'll show you something that I
believe you will be able to help me correct very easily.
Can you show some code ?
