When I run xpdf I get the message "Warning: Attempt to remove nonexistent passive grab".
The program seems to run in spite of this. I googled for the message, with a number of hits, but none had any suggestion of a solution, or a cause.
Developers who introduce warning messages like this should be sentenced to read the complete works of Kim Jong Il.
On Sun, 29 Apr 2007 23:54:06 +0100 Timothy Murphy tim@birdsnest.maths.tcd.ie wrote:
When I run xpdf I get the message "Warning: Attempt to remove nonexistent passive grab".
The program seems to run in spite of this. I googled for the message, with a number of hits, but none had any suggestion of a solution, or a cause.
Really?
This is on the first page of Google when I search for that phrase:
http://www.cs.bham.ac.uk/research/projects/poplog/bugfixes/BUGREPORTS
On that page, you find this explanation:
QUOTE: Warning: Attempt to remove non-existent passive grab
Answer: They are meaningless, and you want to ignore them. Do this (from Kee Hinckley) by installing an XtWarning handler that explicitly looks for them and discards them:
static void xtWarnCB(String message) { if (asi_strstr(message, "non-existant passive grab", TRUE)) return; ...
They come from Xt, and (W. Scott Meeks): "it's something that the designers of Xt decided the toolkit should do. Unfortunately, Motif winds up putting passive grabs all over the place for the menu system. On the one hand, we want to remove all these grabs when menus get destroyed so that they don't leak memory; on the other hand, it's almost impossible to keep track of all the grabs, so we have a conservative strategy of ungrabbing any place where a grab could have been made and we don't explicitly know that there is no grab. The unfortunate side effect is the little passive grab warning messages. We're trying to clean these up where possible, but there are some new places where the warning is generated. Until we get this completely cleaned up (1.2 maybe), your best bet is probably to use a warning handler." END OF QUOTE
It appears to be a cosmetic bug.
On Sun, 2007-04-29 at 23:54 +0100, Timothy Murphy wrote:
When I run xpdf I get the message "Warning: Attempt to remove nonexistent passive grab".
The program seems to run in spite of this. I googled for the message, with a number of hits, but none had any suggestion of a solution, or a cause.
Developers who introduce warning messages like this should be sentenced to read the complete works of Kim Jong Il.
I suspect that warning is for developers, not users. It would be nice if those kind of errors were only produced if debug package was installed - or if started with a switch asking for it.
The message may not even be part of xpdf but from a library xpdf uses (perhaps uses incorrectly or in a deprecated way that will break in future - hence a warning)
On Sun, 2007-04-29 at 23:54 +0100, Timothy Murphy wrote:
When I run xpdf I get the message "Warning: Attempt to remove nonexistent passive grab".
Developers who introduce warning messages like this should be sentenced to read the complete works of Kim Jong Il.
Hmm, I've seen ones that say "an error has occured, because an error has occurred." I think that was on a Mac.
Somewhere I have a screen grab of a useless pop-up error message on Linux. There's a box with an alert icon and an okay button. But there's no clue as to what program generated the error, nor what the error is. The only text it has is the okay button.
Tim wrote:
On Sun, 2007-04-29 at 23:54 +0100, Timothy Murphy wrote:
When I run xpdf I get the message "Warning: Attempt to remove nonexistent passive grab".
Developers who introduce warning messages like this should be sentenced to read the complete works of Kim Jong Il.
Hmm, I've seen ones that say "an error has occured, because an error has occurred." I think that was on a Mac.
Somewhere I have a screen grab of a useless pop-up error message on Linux. There's a box with an alert icon and an okay button. But there's no clue as to what program generated the error, nor what the error is. The only text it has is the okay button.
My favorite came from an old AT&T SysVr3 box. It was: "Error: should not happen".
Les Mikesell wrote:
My favorite came from an old AT&T SysVr3 box. It was: "Error: should not happen".
I believe the standard wording was
"Error: cannot happen"
That is standard wording for "I have a switch/chase/multiple selection statement, and all legal values of the selection argument are accounted for, but for completenes and so the program will do something realistic in all situations, I included a default case which displays this message, and takes some appropriate action."
Mike
On Mon, 30 Apr 2007 07:51:44 -0500 Les Mikesell lesmikesell@gmail.com wrote:
My favorite came from an old AT&T SysVr3 box. It was: "Error: should not happen".
Ah, if we are gonna dig up errors from the good old days, my favorite was always the IBM JCL error something like:
IEF020I DCB ERROPT=ABE and/or no SYNAD exit specified.
When you looked it up in the trusty messages and codes manual, you'd get the vastly improved version:
IEF020I The DCB ERROPT was set to ABE and/o no SYNAD exit was specified. Programmer action: Correct error and resubmit job.
Now that's what I call helpful! (For those dying to know, the actual error was always that the $#@! operator in the comp center mounted the wrong tape on the wrong drive :-).
On Mon, 2007-04-30 at 17:36 -0400, Tom Horsley wrote:
Ah, if we are gonna dig up errors from the good old days, my favorite was always the IBM JCL error something like:
IEF020I DCB ERROPT=ABE and/or no SYNAD exit specified.
When you looked it up in the trusty messages and codes manual, you'd get the vastly improved version:
IEF020I The DCB ERROPT was set to ABE and/o no SYNAD exit was specified. Programmer action: Correct error and resubmit job.
Many years ago, on a computer far away ;-) we had a Data General mainframe that would give you two lines worth of error report if you stuffed up. It'd give you its (very detailed) error message about what was wrong, and would throw the line back at you that was wrong, with the cursor sitting at the point in the line the error occurred. All you had to do was fix your typing and press enter.
Sodding personal computers, on the other hand, would just say SYNTAX ERROR, and leave you scratching your head. At least my old Amiga had a "why" command. If you didn't understand the previous short error message, typing in "why" would give you a more lengthy explanation of the error. It saved you digging out a manual.
On 30Apr2007 16:27, Tim ignored_mailbox@yahoo.com.au wrote: | Somewhere I have a screen grab of a useless pop-up error message on | Linux. There's a box with an alert icon and an okay button. But | there's no clue as to what program generated the error, nor what the | error is. The only text it has is the okay button.
I salute this one, but recently had one nearly as good, and far more annoying.
I've been hacking some Wiki extensions at work that load project information on-the-fly into pages from the project/user database we have, so you can say things like: <sysdb>members:foo</sysdb> and have it fill stuff in - in this way the page doesn't need maintenance over time.
The extension inserts javascript that modifies the page in place. Worked fine on Firefox but aborted the page load on IE. Under IE6 it gave a popup that said "cannot load page", and no further detail. The page would be truncated where the error occurred. Unhelpful, but at least you could see where it was failing.
During the debugging someone "upgraded" IE from 6 to 7 on the box I was testing with. The improved version still aborts the page load, still produces the dinky "cannot load page" with no further detail, but as soon as you click OK it _replaced_ the page with some inane "help" page. It suggested I check that I was still connected to the internet and other equally bogus ideas.
So under IE6 you got a useless error message. Under IE7 you get the same useless message, and then it takes away the thing you were working on!
Cheers,
On Mon, 2007-04-30 at 16:27 +0930, Tim wrote:
Somewhere I have a screen grab of a useless pop-up error message on Linux. There's a box with an alert icon and an okay button. But there's no clue as to what program generated the error, nor what the error is. The only text it has is the okay button.
And here it is (attached). It gives no clue as to what it's about. For all you know could be about to delete some important files by pressing the okay button.
I don't know how I first accountered this, but this is how I got it today: I tried to unmount a USB flash drive within Nautilus, while another terminal still had it as its current working directory. The attempt first popped up a sane error warning that it couldn't do it (cannot unmount volume, an application is preventing the volume from being unmounted). Then moments after okaying away that one, this one popped up.
I have seen this before, and not while unmounting a volume within Nautilus.
-
Cameron Simpson -
Frank Cox -
Les Mikesell -
Michael A. Peters -
Mike McCarty -
Tim -
Timothy Murphy -
Tom Horsley