IMAP Memory considerations with large ‘FETCH’ sizes.
The following comments concern classes in the
[org.apache.common.net.imap](https://commons.apache.org/proper/commons-net/apidocs/org/apache/commons/net/imap/package-summary.html)
package.
Consider the following imap ‘fetch’ exchange between a client (>) and server
(<):
> A654 FETCH 1:2 (BODY[TEXT])
< * 1 FETCH (BODY[TEXT] {80000000}\r\n
…
< * 2 FETCH …
< A654 OK FETCH completed
The first untagged response (* 1 FETCH …) contains a literal {80000000}
indicating the data to come will be a size of 80MB.
After reviewing the
[source](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l298),
it is my understanding, the entire 80MB sequence of data will be read into
Java memory even when using
‘[IMAPChunkListener](https://commons.apache.org/proper/commons-net/apidocs/org/apache/commons/net/imap/IMAP.IMAPChunkListener.html)’.
According the the documentation:
Implement this interface and register it via
[IMAP.setChunkListener(IMAPChunkListener)](https://commons.apache.org/proper/commons-net/apidocs/org/apache/commons/net/imap/IMAP.html#setChunkListener-org.apache.commons.net.imap.IMAP.IMAPChunkListener-)
in order to get access to multi-line partial command responses. Useful when
processing large FETCH responses.
It is apparent the partial fetch response is read in full (80MB) before
invoking the ‘IMAPChunkListener’ and then discarding the read lines (freeing up
memory).
Back to the example:
> A654 FETCH 1:2 (BODY[TEXT])
< * 1 FETCH (BODY[TEXT] {80000000}\r\n
…. <— read in full into memory then discarded after calling IMAPChunkListener
< * 2 FETCH (BODY[TEXT] {250}\r\n
…. <— read in full into memory then discarded after calling IMAPChunkListener
< A654 OK FETCH completed
Above, you can see the chunk listener is good for each individual partial fetch
response but does not prevent a large partial from being loaded into memory.
Let’s review the
[code](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l298):
[296](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l296)
int literalCount = IMAPReply.literalCount(line);
Above counts the size of the literal, in our case 80000000 or 80MB (for the
first partial fetch response).
[297](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l297)
final boolean isMultiLine = literalCount >= 0;
[298](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l298)
while (literalCount >= 0) {
[299](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l299)
line=_reader.readLine();
[300](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l300)
if (line == null) {
[301](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l301)
throw new EOFException("Connection closed without indication.");
[302](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l302)
}
[303](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l303)
replyLines.add(line);
[304](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l304)
literalCount -= line.length() + 2; // Allow for CRLF
[305](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l305)
}
The literal count above starts at 8000000 and is decremented until reaching
non-zero negative val where 80MB is read in full and the while loop returns.
[306](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l306)
if (isMultiLine) {
[307](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l307)
final IMAPChunkListener il = chunkListener;
[308](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l308)
if (il != null) {
[309](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l309)
final boolean clear = il.chunkReceived(this);
[310](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l310)
if (clear) {
[311](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l311)
fireReplyReceived(IMAPReply.PARTIAL, getReplyString());
[312](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l312)
replyLines.clear();
Now, after all 80MB is loaded into memory in full, invoke the IMAPChunkListener
and finally throw away the lines freeing memory. And moving onto the next.
[313](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l313)
}
[314](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l314)
}
[315](https://gitbox.apache.org/repos/asf?p=commons-net.git;a=blob;f=src/main/java/org/apache/commons/net/imap/IMAP.java;h=d97f1073d8b97545d0a063c6832fe55c116166e2;hb=HEAD#l315)
}
I’m considering modifying the getReply() method, shown above, to chunk the
partial responses breaking up the literal so that it’s not loaded into memory
in full. This is to prevent the entire 80MB literal value from being loaded
into memory.
This would be configurable as not to break the existing users of the API.
Something like .setBreakLargeLiteralSize(true), when breakUpLargeLiteralSize is
true, a maxLiteralBuffer value is used to chunk the literal preventing all 80MB
from being loaded in full, instead loading chunks of it, calling
IMAPChunkListener sooner than later. This would require implementations of
IMAPChunkListener to handle this behavior if it was turned on. The default
behavior will see this chunking disabled as to not break the existing users.
Essentially an opt-in feature reducing the risk.
What are you thoughts or concerns with this? Do you agree?
