OK, I'm back with more information. I've set debug to true, and looked
at the progress messages. I noticed two differences in the messages
between my XP machine and the Win7 test machine:
1. On the Win7 machine, I get a series of messages about copying files
into the _MEI folder; these don't happen on the XP machine. Also,
there's a few messages after this:"Setting up to run child",
"Creating child process", "Waiting for child process to finish",
"_MEI... is <path>", "Already in the child - running", "manifestpath:".
2. Next on the Win7 machine, I get "Error activating the context".
This message comes from utils.c, function CreateActContext.
Apparently, CreateActCtx is failing.
The extra messages make it look as though it's getting a two pass
execution on Win7 (although a _MEI folder is created in both cases).
I should mention that the Win7 machine is 64 bit (my development machine
is 32 bit). I've noticed that the installer (using NSIS) installs into
the (x86) folder by default.
By the way, I think I've solved the mixed-case "PyWinTypes25.dll" issue:
in hook-pywintypes.py, I changed the format string on line 20 to all
lower case. With that done, I no longer get the "exists, but shouldn't"
warning message.
Any pointers appreciated. This is beginning to impact a high-priority
project.
On 3/5/2012 1:23 PM, Don Dwiggins wrote:
I've tried the branch, and gotten the same result. I looked at the
change in build.py, and it seems like a reasonable thing to do.
However, there's something else going on here.
As I mention below, we're dealing with two different files, which have
the same name except for case, but _different_ sizes. I can see where
the smaller, all lower case file comes from: it's definitely part of
my Python 2.5 installation (it lives in Windows/system32), and PyI
picks it up and includes it as it should. The strange part comes in
the process of setting up the _MEI folder at the start of execution of
the main application. At some point, the other, larger, mixed case
file has been put in the folder.
(On further investigation: I ran ArchiveViewer on the main .exe file,
and among the the output were the two following lines (not contiguous):
(1731776, 0, 0, 0, 'd', '..\\dummy\\dummy:PyWinTypes25.dll'),
(1731776, 0, 0, 0, 'd', '..\\dummy\\dummy:pywintypes25.dll'),
That led me to look more closely at the voluminous output from the FyI
run. I found the following line in the analysis of several of the
applications:
W: pywintypes is changing it's name to PyWinTypes25
I hope this will help figure out what's going on and how to stop it.
Another hint: a bit of grepping found the warning message in
source/common/launch.c. Maybe the "normcasing" should be done there
as well.
-----
Separate issue: in the merged setup, the MSVCR71.dll file only exists
in the dummy directory. When I start the main application, after the
warning message, I get the system error message "The program can't
start because MSVCR71.dll is missing from your computer." If I copy
the file from dummy to the main application folder, I don't get this
message.
Again, thanks for the help,
--
Don Dwiggins
Advanced Publishing Technology
--
You received this message because you are subscribed to the Google Groups
"PyInstaller" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/pyinstaller?hl=en.
