Color changes for wild cards

[codeedog]

Love this tool. When I tried the following (package.json has two start- scripts):

concurrently -c cyan,yellow "npm:start-*"

the output of both subtasks were displayed with cyan rather than the first being colored cyan and the second being colored yellow. Instructions indicate that when color count is less than command count, the last color listed is used. I suspect whatever does the color logging sees only the command line argument count (a single command) and doesn't realize its a wildcard command so doesn't look for the actual number of commands running. Or, the wildcard processing code doesn't update some command count global. Or, the wildcard processing code is somehow treated as only one process.

In any event, it'd be nice to have wildcard initiated commands individually colored.

My work around is to just explicitly start the commands by name and I still get the coloring. This works:

concurrently --c yellow,cyan npm:start-ng npm:start-express"

In addition, to explore some edge cases, I wrote a simple shell script that prints to the terminal and sleeps (see below). When I run it in with the wildcard, the wildcard logging is treated as a single process. Examples:

concurrently -k -c cyan,yellow,green "bash foo.sh" "npm:start-*"
concurrently -k -c cyan,yellow,green "npm:start-*" "bash foo.sh"

The first line outputs the shell logging in cyan and the two wildcards in yellow.
The second line outputs the two wildcards in cyan and the shell logging in yellow.

foo.sh

echo "hello"
sleep 1
echo "slept 1"
sleep 5
echo "slept 5"
sleep 60
echo "slept 60"

[daniloraisi]

I was analyzing the code, and it seems that the problem is the way concurrently prefix colors.

As of this line:

concurrently/bin/concurrently.js

Line 153 in 0bc490f

lastColor = prefixColors[index] || lastColor;

It is considering only the first command to apply the color instead of applying each one when using a wildcard.

The command parser needs a way to apply prefix color when splitting npm wildcards.

@milahu, I can work on a PR to fix this.

[gustavohenke]

Fix out in v6.3.0.