I found the last instance in just a few minutes with this technique, as
there are only a few pplaces in the code where that exception is thrown:
I placed a breakpoint on all instances of ImageNotFoundException being
generated, and then ran in debug until a break; then stepped up through the
stack trace. Perhaps 10 minutes or so and I had found the offending code.
With luck the other instances of thiswill be as easy to find.
Also, the instance I found last night was an initialization string like:
“;moved;”, so a search could also be made for the image name with a trailing
semicolon as a second way to track down instances of this.
Using the v3.2 series, (the last available is svn7848) The
Processingg Tile Image function kills this Module : No cards
available into the Module Deck.
After loading into v3.2 : Deck is now empty Using v3.1 series ; Deck
is full and no image problem found.
According to the errorLog, the names of the image files which aren’t found
all have no extensions.
This is related to the “moved” image being reported as missing in some
modules. I wonder if the switch to not appending ‘.gif’ to the names of
images we can’t find is going to be problematic. I was hoping we’d finally
be able to stop that…
I found the last instance in just a few minutes with this technique, as
there are only a few pplaces in the code where that exception is thrown:
I placed a breakpoint on all instances of ImageNotFoundException being
generated, and then ran in debug until a break; then stepped up through the
stack trace. Perhaps 10 minutes or so and I had found the offending code.
With luck the other instances of thiswill be as easy to find.
Also, the instance I found last night was an initialization string like:
“;moved;”, so a search could also be made for the image name with a trailing
semicolon as a second way to track down instances of this.
But these are images from the module Gregory is loading, you won’t find
them referenced as strings anywhere in our code.
I made a copy of my current mod, opened it in the editor, images tiled.
I made a change to the module version info and saved it. Re-opened the
mod for editing and it tiled again. Log created in the Documents and
Settings → Vassal is listed below.
I need a bit more information in the diagnostic output than what that
build printed. I’ve uploaded a build marked “mtime0” for you to try.
Do the same, and post the log from that build.
Try the “mtime1” build instead. I had an idea of what might be causing
the problem, so included a potential fix in this build.
That would be evident quickly from the error message; if so, a regex search through the Buildfile should locate patterns like “image=[”][^.“]+[”]" for remedial action. I don’t have a command-line regex tool handy; do you?
I ran the mtime1 build and it did appear to solve the re-tiling of the map every time I open it.
However, When I saved the mod via “Save As” and give it a different file name, it does re-tile when I open it. Is that normal or should the mod maintain the tiling information regardless of how it’s saved?
But these are images from the module Gregory is loading, you won’t
find
them referenced as strings anywhere in our code.
That would be evident quickly from the error message; if so, a regex
search through the Buildfile should locate patterns like
“image=[”][^.“]+[”]" for remedial action. I don’t have a command-line
regex tool handy; do you?
Yes: sed, awk, perl, probably five or six more if I look around.
My point was that this could affect many modules (most of which we have
access to), as well as save and log files (most of which we don’t have
access to).
I ran the mtime1 build and it did appear to solve the re-tiling of the
map every time I open it.
Good.
However, When I saved the mod via “Save As” and give it a different file
name, it does re-tile when I open it. Is that normal or should the mod
maintain the tiling information regardless of how it’s saved?
Since (2) originally occurred before I found this thread and wasn’t paying close attention from a new-build-testing point of view, lets chalk it up to user error.
I was originally assuming (1) was user error as well until someone else here said they also have the problem.
I will try (3) again later today or tomorrow morning using build svn7855
(does ‘svn’ indicate you guys are using Subversion for an archive/code repository?)
I tried installing and running svn7850 (on a Windows XP machine) just
now to see what it’s like. The fonts are smaller for me, just like the
screenshots on page 5 of the thread. Upon loading my Here I Stand
module, it reported “Bad Data in Module: Image not found” for everything
that’s an SVG–and none of my toolbar buttons using SVG appear. I’m
guessing support for SVG is being left out for the sake of simplicity
right now?
No, it’s not intentional, SVG is just broken at the moment. (I
discovered this myself last night.)
This is fixed in svn7857. (BTW, Here I Stand looks like a rather wild
game.) The problem was that that SVG loader was improperly adding
“images/” to all of the SVG image paths when they already began with
“images/”.
Does anyone know of a module which uses SVG maps? We should test that,
as well.
I tried installing and running svn7850 (on a Windows XP machine) just
now to see what it’s like. The fonts are smaller for me, just like the
screenshots on page 5 of the thread. Upon loading my Here I Stand
module, it reported “Bad Data in Module: Image not found” for everything
that’s an SVG–and none of my toolbar buttons using SVG appear. I’m
guessing support for SVG is being left out for the sake of simplicity
right now?
No, it’s not intentional, SVG is just broken at the moment. (I
discovered this myself last night.)
This is fixed in svn7857. (BTW, Here I Stand looks like a rather wild
game.) The problem was that that SVG loader was improperly adding
“images/” to all of the SVG image paths when they already began with
“images/”.
Does anyone know of a module which uses SVG maps? We should test that,
as well.
I mean that it looks like a lot of disparate things are going on at
once, as well as having one player which is mainly non-military and
one player which is both.
The tiling/caching pre processing of module images only works on base
module. It does not pick up any images found in module extensions that
are loading. This causes a lot of …
011-09-12 11:07:03,389 [3-AWT-EventQueue-0] WARN
VASSAL.tools.ErrorDialog - Image not found:
24dd6351be204abaf61bb8abcbf61b3080511eeb
2011-09-12 11:07:03,389 [3-AWT-EventQueue-0] WARN
VASSAL.tools.ErrorDialog - Image not found:
becbd125990949fdc340d79413c792b030f02b4f
…and so on
No map board shows if the map was in the extension naturally (or any
other image from extension of course). That is going to be a problem
Try svn7874. As it is now, each extension will end up with its own
slicing process and progress dialog, which is kind of stupid. I’m mainly
interested in whether this works at all at present.
So… the server won’t allow beta 3.2 builds to connect? Is that correct? I thought we’d test it out with a live connection but couldn’t connect; figured if all the players had the same version of the client the server would allow it.
