[SOLVED] tree giving false matches when using pattern flag

Arch900 · 01-25-2022, 05:37 AM

Hi all,

This question is maybe more suited for tree's github page, but I want to make sure I'm not being an idiot.

When using more than one wildcard as a pattern with tree I get what I consider to be false matches. For example

Code:

tree -P "*.gpg" --prune .

works as expected and shows me all my .gpg files, however

Code:

tree -P "*.html*.gpg" --prune .

matches both files ending in .html.gpg AND files ending in just .html (but not files ending in just .gpg). In my opinon it should only match files that have both .html and .gpg in the title.

Am I wrong, or am I making a mistake in the command?

shruggy · 01-25-2022, 05:56 AM

Definitely looks like a bug. I can reproduce this with tree 1.7.0 and 1.8.0. The latest version is 2.0.1 though, and the changelog indicates there were changes in -P processing recently. I'd try the latest version before filing a bug report.

Update. Still present in 2.0.1. Report the bug. The homepage has an option of sending a nasty e-mail to the maintainer with the subject "You Bastard!"

Arch900 · 01-25-2022, 07:02 AM

Quote:

Originally Posted by shruggy

Definitely looks like a bug. I can reproduce this with tree 1.7.0 and 1.8.0. The latest version is 2.0.1 though, and the changelog indicates there were changes in -P processing recently. I'd try the latest version before filing a bug report.

Update. Still present in 2.0.1. Report the bug.

Thank you for confirming. I've reported it now (by email - I think the github I was looking at is a fork as it's a bit out of date).

Quote:

Originally Posted by shruggy

The homepage has an option of sending a nasty e-mail to the maintainer with the subject "You Bastard!"

That's funny. I like maintainers with a sense of humour!

Arch900 · 01-26-2022, 01:39 PM

I've had a reply to my report!

Quote:

Looks like a bug. I think this will fix it, for 2.0.1 insert the
line:

match = 0;

after line 1145 of tree.c, i.e.:

if(!*pat) return 1;

in the patmatch function. It seems that match should default to no match
unless the remainder of the pattern matches something. The line at 1159
might also need to be moved just before the 'return match' line (outside
of the else). That was likely placed in the else clause alone in error.

Thanks for the report. I'd like to add a few things before another update,
so it might be a bit (just hopefully not another 3 years.)