If you want to have this newsletter mailed to you or you want to make comments/suggestions about the format/content then send an email to acedb@sanger.ac.uk.
A new acedb developer joins the team, a number of new features for users and a number of bug fixes.
After a hiatus of some months following the departure of Fred Wobus from our team (now pursuing his hobby of scuba diving in Thailand !), we have been joined by John Sturdy (jcgs@sanger.ac.uk). John is a Computer Science Ph.D. and has worked for various software companies locally for some years. Initially he will be working on improving the acedb web site so you should see quite a few improvements there over the next few weeks.
An irritating feature of using "update" mode in the "Tree" window has been that the window is repositioned at the top when the mode is selected, thus losing the users place each time an update is performed. This has now been fixed: select a tag by clicking on it, then select "update" from the right button menu and the tag will stay visible in the window.
Searching for a particular tag in a large object in the tree window can be tedious, especially when this involves opening collapsed nodes to find the tag. The right button menu now has an additional "Find tag" item that allows you to search for a tag (it will find the tag whether the tag is within a collapsed node or not).
A horizontal scroll bar has been added to the fmap display so its no longer necessary to turn columns off to see whats displayed over to the right of the display. Headings/Titles/Column buttons are not redrawn as you scroll because this slows scrolling down too much.
Most MS-Windows users have 2-button mice which means they
have not been able to use functions that were linked to the middle button
on Xwindows/Unix machines. This has now been fixed by making shift-leftmousebutton
do the same job as the middle button for 2-button users.
The "?" command to get help in tace has changed in the way it works.
If there is only one command matching the string following "?" then the syntax
for that one
command will be printed, if there is more than one, then syntax for all matching commands will
be printed, e.g.
"?"
"? d"
"d"
"? du"
"du"
"spick" addtion to keyset manipulating commands in tace.If you use the spush, spop etc. commands for manipulating keysets in tace
then you should
note a useful addition to these of spick (courtesy of Lincoln Stein).
spick allows you to
choose one of the stacked keysets and make it the active keyset without having to spop the
stack to get to the one you want.
Printing acedb windows now works from Windows 95 in the same way as for acedb on Unix machines.
You can now use the up and down arrows to scroll through the items in the keyset window.
Oh dear, while I'd love to say this bug was fixed, its not and can't be for the foreseeable future. The cause of the bug is based in NFS/lockmanager problems which arise from time to time on the Sanger network. Acedb uses lock files to make sure that only one user can have write access to a database at a time, this works well most of the time. There can be problems in two situations however: when the network is heavily loaded so that NFS starts to time out meaning that requests to create/write lockfiles stall and when lock managers on machines remote from yours on the network begin to hang (usually because of the NFS problems) meaning that acedb will be unable to lock its lock files. In either case xace or tace will simply hang, i.e. not respond to mouse clicks etc.. When this happens, all you can do is kill the program using the Unix "kill" command, you will need to use "ps" to find out the process id of the hung program and then type "kill -9 xxxxx" where xxxxx is the process id. The acedb code cannot detect that this problem has occurred because it cannot run while waiting to create/lock the lockfile. Sadly there is nothing we can do to solve this problem.
There is a bug in the November code which may cause the "Gene finder" button in fmap to crash the xace program. The problem happens when the definition of the ?Method class in the models.wrm file in the users database does not match the definition in the models.wrm file distributed with the code. Previous versions of acedb did not detect this mismatch and continued to run, probably with some errors in display of some fmap information. Detection of this sort of parse error was tightened up recently to support new function in AcePerl, hence the reason for the code failing now. Database administrators should check that the definition of the ?Method class in their models.wrm files matches that distributed with the code level they are using.
We hope to make the code more tolerant of this problem in the future by altering the code to check the model of the database against its method own object and decide if it can continue if there is a difference. This will probably not happen until the New Year.
Tab key:There was a bug whereby if the user pressed the Tab key
while inputting text, the window would then no longer accept input. This has now been fixed
and tabs can be used as input to text boxes.
You can pick up the monthly builds from: