I've found what appears to be a strange bug. I've got a script which runs:
...
The source code is on a UNC share. mpress is seeing share violations on the EXE.
I cannot reproduce this if I copy the whole directory to my local hard drive.
I cannot reproduce this if I use the release build of the compiler (i.e. swap the two EXEs in the Compiler directory).
The most that I can determine from Process Monitor is as follows:
1. mpress copies the EXE to a temp file
2. mpress hits a sharing violation on the target (EXE) file (three retries) presumably when writing out the compressed EXE
mpress.exe (2.19) does not return a status code under these circumstances, so ahk2exe is not being made aware of the problem. (I have tested this by using /mpress 0 and running mpress manually as a separate step).
By and large I'm having more luck running mpress manually after ahk2exe completes — typically this works OK. Not always, but usually.
This is on my XP computer. I can't reproduce this under Windows 7.
Not a huge issue, I just need to double-check that compression actually worked each time
There's some very weird juju with ahk2exe/mpress and some horrible timing sensitivities.
ahk2exe status code
ahk2exe.exe does not appear to return a status code if compile fails. (I'm also getting occasional "Error: Unable to copy AutoHotkeySC binary file to destination." and my script cannot pick up on this, and tries to run the compress stage, as there's no errorlevel output from this. I can't say off-hand exactly what the message is as it won't do it now.)