Web lists-archives.com

Setup-x86(_64) (v2.889) ... just a question


Just a question, meaning it may have been discussed in cygwin-apps,
but I have been reading that list for quite a while now.

Recently I reported a problem:

    (Another setup v2.889 regression? additional information)

Jon T. applied a big hammer to solve the problem ... (in short, in
the entry for cscope ,the version number in the "prev" subentry was
replaced as follows:

    version: 15.8a-2
(i.e. version number was replaced by a "version number"
 that is NOT part of the filename)

Today David Lee reported a related problem:

    (Setup problem (2.889 32-bit) )

I replayed what David described and confirmed the issue. Then I
downloaded and installed the most recent version of liblz4_1.

When setup (install mode) is invoked again, it wants to install the
previous version of liblz4_1 (of course!).

... so I applied Jon's big hammer to setup.ini and replaced

    131-1 by 1.7.5-0 (version:) in the prev subentry (pkg liblz4_1)

That removed the issue (again, of course!).

However to my surprise /etc/setup/installed.db now lists:

64-@@ grep -e cscope -e lz4 /drv/e/Cygwin64-test/etc/setup/installed.db
cscope cscope-15.8a-2.tar.bz2 1 <==== artifical version number
liblz4_1 liblz4_1-1.7.5-0.tar.bz2 0 <==== ditto

These files do NOT exist. The file names were "artificially crafted".

The files that do exist, are:

64-@@ ls -l release/cscope release/lz4/liblz4_1
total 700
-rw-r--r-- 1 Henri None 381409 Mar  6 13:18 cscope-
-rw-r--r-- 1 Henri None 330264 Mar  6 13:16 cscope-15.8b-1.tar.xz

total 52
-rw-r--r-- 1 Henri None 30408 Mar  6 12:37 liblz4_1-1.7.5-1.tar.xz
-rw-r--r-- 1 Henri None 17924 Mar  6 11:15 liblz4_1-131-1.tar.xz

Is this the future of setup? Meaning, is it the intention that it may
be possible that installed.db refers to files that do not exist?


Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple