![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Entry tags:
software annoyances
At work I've been running an ancient and venerable version of cygwin (Unix shell for Windows machines), designated "B20". Once too often I got annoyed by not having a useful utility available, so I finally decided to upgrade. Not being stupid, I did not allow it to install over my current copy.
The new version (no obvious version number; current as of Friday) breaks for me in fundamental ways. I don't understand why. Yes, I have read TFM.
First off, it doesn't automatically source my .bashrc the way the old version does. Yes, HOME is set correctly (and the old version gets its). Peculiar, but an obvious workaround exists. Perhaps relatedly, it ignores the "start in [directory]" setting in the shortcut definition. (I want it to start in a directory that is not HOME but it won't.)
Second, by default it emits some text in dark green on my black background, bypassing the color settings I've given it. This problem goes away after I source my .bashrc manually. I guess it's doing something strange with the default prompt, which my .bashrc changes to show the current working directory. Wacky, but if it doesn't recur I can ignore that.
However, there are two real problems. First, even after processing my .bashrc, which includes all the usual definitions of environment variables like ANT_HOME and JAVA_HOME and PATH, it can't find ant. When I run ant via an explicit path, it can't find java. I fired up new and old cygwin shells side by side and applied printenv to all the relevant variables; the definitions match. (The new cygwin has added some stuff to PATH and casts everything as "/cygdrive/d/blah_blah_blah", but it doesn't seem to be missing anything.) This is fatal; it does me no good to have the latest Unix tools if I can't run ant or (apparently) java.
And second, noticed in passing, the new version is sloooooowwwww. I issued a printenv command (for one variable, not the entire environment) in the new shell, switched to the old shell, typed the same command, and switched back, and the old one completed several seconds before the new one did. I have no idea what's going on.
I thought cygwin pretty much worked out of the box, but it's not doing it for me. Frustrating, especially when the old version did. I wonder what's wrong.
The new version (no obvious version number; current as of Friday) breaks for me in fundamental ways. I don't understand why. Yes, I have read TFM.
First off, it doesn't automatically source my .bashrc the way the old version does. Yes, HOME is set correctly (and the old version gets its). Peculiar, but an obvious workaround exists. Perhaps relatedly, it ignores the "start in [directory]" setting in the shortcut definition. (I want it to start in a directory that is not HOME but it won't.)
Second, by default it emits some text in dark green on my black background, bypassing the color settings I've given it. This problem goes away after I source my .bashrc manually. I guess it's doing something strange with the default prompt, which my .bashrc changes to show the current working directory. Wacky, but if it doesn't recur I can ignore that.
However, there are two real problems. First, even after processing my .bashrc, which includes all the usual definitions of environment variables like ANT_HOME and JAVA_HOME and PATH, it can't find ant. When I run ant via an explicit path, it can't find java. I fired up new and old cygwin shells side by side and applied printenv to all the relevant variables; the definitions match. (The new cygwin has added some stuff to PATH and casts everything as "/cygdrive/d/blah_blah_blah", but it doesn't seem to be missing anything.) This is fatal; it does me no good to have the latest Unix tools if I can't run ant or (apparently) java.
And second, noticed in passing, the new version is sloooooowwwww. I issued a printenv command (for one variable, not the entire environment) in the new shell, switched to the old shell, typed the same command, and switched back, and the old one completed several seconds before the new one did. I have no idea what's going on.
I thought cygwin pretty much worked out of the box, but it's not doing it for me. Frustrating, especially when the old version did. I wonder what's wrong.
random nostalgia
no subject
S
no subject
I also just noticed that it's not slow out of the box; after I source my .bashrc it becomes really slow. I have no idea why; all my .bashrc does is set some environment variables and redefine the prompt. The file was of course developed for the old version (and, mostly, by people who were already here when I got here), but I don't see anything in there that should matter. Very strange.
no subject
S