Minor bug fixes for Felix, Memory Serves

16 views
Skip to first unread message

S. Patrick Eaton

unread,
Jan 18, 2012, 8:04:30 PM1/18/12
to felix...@googlegroups.com
Hello, everyone.

There are two problems that I'm hoping Ryan can look into when time
allows. I think both are things that may have been fixed in previous
versions, but now are occurring again in the current versions.

1.) In Felix 1.6.7.2, when a translation is updated (this might apply
only to remote memories; I haven't tested it on local ones), both the
"Date created" and "Last modified" attributes are changed to the time of
the update. In some (all?) cases, the "Created by" and "Modified by"
attributes also seem to change to the user who made the update.

This isn't a very big deal because it doesn't prevent us from getting
work done, but one of our translators has been bringing this up
repeatedly for quite some time now. I'm not sure exactly why, but
apparently this unexpected behavior drives her nuts.

2.) In Memory Serves 1.9 (yes, I've noticed that 1.9.5 was released
recently, but haven't updated to the latest version yet), search results
appear as expected for the first page, but attempting to click through
to subsequent pages results in the following error:

---

500 Internal Server Error

The server encountered an unexpected condition which prevented it from
fulfilling the request.

Powered by CherryPy 3.2.0

---

This happens both when searching within individual memories/glossaries
and when using the Global Search.

This is a bit more of a problem than the first one, but at the moment it
hasn't been preventing us from getting work done.

Is anyone else experiencing similar problems?

Best regards,

Sako Eaton


Ginstrom IT Solutions

unread,
Jan 18, 2012, 10:14:02 PM1/18/12
to felix...@googlegroups.com
> On Behalf Of S. Patrick Eaton

> 1.) In Felix 1.6.7.2, when a translation is updated (this might apply only
> to remote memories; I haven't tested it on local ones), both the "Date
> created" and "Last modified" attributes are changed to the time of the
> update. In some (all?) cases, the "Created by" and "Modified by"
> attributes also seem to change to the user who made the update.

> 2.) In Memory Serves 1.9 (yes, I've noticed that 1.9.5 was released
recently,
> but haven't updated to the latest version yet), search results appear as
> expected for the first page, but attempting to click through to subsequent
> pages results in the following error:

Fortunately, I was able to find the causes of these bugs and fix them
quickly. I've released version 1.9.5.1 with the fixes.

http://felix-cat.com/tools/memory-serves/

Unfortunately, this uncovered some holes in my testing regime for Memory
Serves, which I will be working to eliminate.


Regards,
Ryan

=================================
Ryan Ginstrom
Felix Translation Memory Software
sup...@felix-cat.com
http://felix-cat.com/
+1-425-402-418
=================================


S. Patrick Eaton

unread,
Jan 19, 2012, 7:41:00 PM1/19/12
to felix...@googlegroups.com
Thanks for the quick response, Ryan!

> Fortunately, I was able to find the causes of these bugs and fix them
> quickly. I've released version 1.9.5.1 with the fixes.
>
> http://felix-cat.com/tools/memory-serves/

I'm afraid these fixes seem to have exacerbated, not ameliorated, the
problems on our end. In our environment, with both versions 1.9.5.1 and
1.9.5, the 500 Internal Server Error occurs on all of the following pages:

* [host]/memories/
* [host]/glossaries/
* [host]/globalsearch/

I have gone back to Memory Serves 1.9 for the time being, and now this
error only occurs on search results pages after the first one, as before.

Our server PC is running Windows 7, if that helps at all. If there is
any other information I can provide that would help pinpoint the source
of these problems, please let me know.

Thanks again for your blazingly fast support!

Best regards,

Sako

Ginstrom IT Solutions

unread,
Jan 19, 2012, 9:31:47 PM1/19/12
to felix...@googlegroups.com
> Behalf Of S. Patrick Eaton
> on our end. In our environment, with both versions 1.9.5.1 and 1.9.5, the
500
> Internal Server Error occurs on all of the following pages:
>
> * [host]/memories/
> * [host]/glossaries/
> * [host]/globalsearch/

Thanks for this report. When you installed Memory Serves, did you select the
installation for all users, or "just for me?"

Also, if you are logged in as an admin, when you go to the Home page, you
will get a link to "Admin," then from there to "View Log." Could you try
copying and pasting the Server Error Log and Server Access Log for me?

> Thanks again for your blazingly fast support!

Alas, if it had but fixed your issue... :/

Ginstrom IT Solutions

unread,
Jan 24, 2012, 2:49:02 AM1/24/12
to felix...@googlegroups.com
> Behalf Of S. Patrick Eaton
> I'm afraid these fixes seem to have exacerbated, not ameliorated, the
problems
> on our end. In our environment, with both versions 1.9.5.1 and 1.9.5, the
500
> Internal Server Error occurs on all of the following pages:

Thanks to some extra help from Sako, I have isolated the cause of this
issue, and released version 1.9.5.2 with a fix.

http://felix-cat.com/tools/memory-serves/

The bug was caused by a failure to apply bug fixes to the part of the code
that supports earlier installs of Memory Serves.

S. Patrick Eaton

unread,
Jan 24, 2012, 5:42:48 AM1/24/12
to felix...@googlegroups.com

> Thanks to some extra help from Sako, I have isolated the cause of this
> issue, and released version 1.9.5.2 with a fix.
>
> http://felix-cat.com/tools/memory-serves/

Thank you for the quick work on this fix, Ryan!

I wish I could say that this has now solved the problem, but I'm afraid
it has not.

Upon installation of 1.9.5.2, Memory Serves started up as it normally
does, and the server error when accessing the pages mentioned in my
previous message was indeed fixed, but when I accessed the pages for
both the Memories and the Glossaries, I immediately noticed that both
had been cut down to a fraction of their original size. Our main memory,
for example, has around 70,000 entries in it, but after this upgrade, it
had closer to 7,000. Other memories and glossaries were in a similarly
truncated state. Yikes!

And this was weird: One of the glossaries listed was one that we do not
use. It was the Edict (Japanese to English) glossary, which I think we
may have registered once a really long time ago, but removed from Memory
Serves years ago to take some of the load off the server.

Fortunately, we had backups, so I've now downgraded back to Memory
Serves 1.9 and the previous version of the database.

Also, when attempting to connect to Memory Serves 1.9.5.2 from Felix, I
get errors (I'll send you the logs off-list), so even if we could get
1.9.5.2 to work with the backed up version of the database, I'm not sure
it would work with Felix.

I hope this information, along with the logs, will prove useful.

Best regards,

Sako

Reply all
Reply to author
Forward
0 new messages