That's great. For those of us who are doing cross-platform work, having forward-slash as a fully-usable path separator is a real benefit. (And "fully-usable" means with the completions that TCC and other great shells support.) That may seem silly to those whose muscle-memory has back-slash...

Yes. My bad. It is the use of '/' as a separator that confounds expansion of ~/Downl<tab> as now restated in post #1.

(1st post edited to reflect a reporting error. As Charles says, ~\Down<tab> can be completed. But with forward slash as a path separator, it cannot.) It's nice that ~ can stand for one's home directory. And it's great that forward slash is generally usable as a path separator. Nicer yet would...

The title says it all. I find the semi-recent addition of leading ~ in a pathname to mean "home directory" to be very handy. (Especially so since that has long been true in *Nix shells.) I also find key-requested expansion of pathnames very handy for navigating directory hierarchy. Those two...

Here is a simple demo of the incorrect behavior: > copy \\wsl$\Ubuntu-21r10\home\larrybr\.bash_aliases \Tmp\ \\wsl$\Ubuntu-21r10\home\larrybr\.bash_aliases => C:\Tmp\.bash_aliases TCC: (Sys) The request was aborted. 0 files copied 1 failed I expect a copy to occur, much like this: >...

Thanks, Charles. The trick was to get "alias DEL=*DEL /E" executed in 4START.btm whether or not it is an interactive shell. (My initial effort was doomed because I had it among the interactive shell aliases.)

I find that when a makefile uses "DEL blah-blah" to preemptively delete some file(s) that may not exist, I have to run it under cmd.exe instead of TCC. This is because, to cmd.exe, the absence of the file target is not cause for DEL to return error, whereas TCC.exe does return an error. Is...

Thanks Vince. I had overlooked or forgotten that /P option for DO. So, this works: do fn in /P (fossil extra) (del %fn) I submit that my suggested enhancement to the filelist feature is still meritorious.

My understanding is that con: refers to a device (or pseudo-device) and will not refer to the input stream when that has been redirected. I was trying to use a DSCM command which reports "extra" files, not under version control, to specify a set of files to be deleted. For example, hoping '-'...

I looked for a way to pipe a filename list into a command expecting to operate on a file set. The "@File lists" feature looked like it should do it, but does not, yet. I submit that DOSTUFF @- should treat standard input as a file list unless a file named - is found.

Consider this common output from FTYPE: batfile="%1" %* If, with administrator permissions, you enter this in cmd.exe: FTYPE batfile="%1" %* , then that first output will remain unchanged. But if you had entered it into TCC.exe, you would get: batfile="" . This is because cmd.exe...

I recently discovered yet another cmd.exe quirk which reminded me why I detest it so. It special-cases the FTYPE command, doing either limited or no % expansion on the line before processing it. This means that the same FTYPE ... %1 ... %* command entered into cmd.exe and TCC produce different...

Thanks, Vince. Now if I could just persuade the list command's search to do without a new, modal dialog window, TCC v26 would be close to perfectly usable without access to the desktop. I suppose I can use aliases to use another lister as a work-around for my unthinking (muscle memory) use of...

Sometimes, the Quick Help is all the help that can be had. For example, when using TCC through an SSH connection spanning thousands of miles. I admit that in such cases, I usually can access either web help or a more local .chm help, but that is not so quick. Sometimes I just want a reminder...

Is there some way to configure TCC so that it will not attempt any window popups and funnel its I/O through the STDIN and STDOUT streams? When I run that shell through an SSH connection, I typically have no convenient way to access the desktop or use a mouse to do anything in another window. So...