Hang infinitely while reading html



I am using Aspose to read a html document. It seems used up 100% CPU usage and hung there infinitely while reading the attached document. I understand that Aspose does not support all HTML tags currently and it shouldn’t hang there just like this. My live server that running it just hang there, and all services that hosted there are down today just because of that. I understand I have no proper error handling for this. But it is just totally unacceptable.

Can you please have a look on this html?


Shu Yih



Thanks for your report. We seem to have discovered the bug causing falling into infinite loop when processing this document. We will keep you posted.


After further investigation, it appeared that the problem lies in the document, though. It seems to be properly formed but it is just too complex for the importer because it contains lots of nested tables. This causes very long processing which might be treated as infinite loop (the document seems to be imported eventually but the time it takes is surely inacceptable). Word opens your document okay on one of our machines but it takes about 12 sec to open it on another one!

So basically please try to simplify this particular document. We are going to refine table importing in the future.


We get Word documents from very varied sources and therefore have no control over how they are formatted. In this case12 secs would actually be an acceptable time to wait for us.

Can we know when your deadline is for refining the importing of tables? We are already using this in a production environment and although it happens rarely but everytime we get a similarly formatted document, it hangs our servers. We need a solution fast.



We have fixed the hanging issue. However, processing this particular document still takes around 20 secs on a P2.6 machine and the output is not correct. We'll try to fix that until the next hotfix is out if it is possible.

We are working on the table layout right now since many customers request that; however, taking into account we are developing lots of highly requested (and not the simplest) features simultaneously, I can't promise table layout will be reworked quickly.